Amazon Advertising Bible: ins and outs from a PPC evangelist
- Alberto Carniel
- Dec 1, 2020
- 61 min read
Updated: Aug 6
In my digital marketing career, I have heard many opinions on how to advertise on Amazon.
The truth is, they can all be wrong and right at the same time.
Why?
Because the only correct advertising strategy is the one that works for your business.
Exactly.
Your business model, product, brand identity and more are unique. What works for you in terms of advertising performance might not work for others, and vice versa.
If testing different approaches until you find the best performer is the core of marketing, is there a common basic approach to Amazon advertising that could represent a foundation for future testing and expansion?
In other words, an advertising structure that can be applied to most businesses and then customized and improved for each unique situation.
Well, what you are about to learn is my Amazon Advertising Bible: a series of standardized steps to set up the initial campaigns that work for virtually every business.
Buckle up and get ready for a comprehensive guide.
PS: To get the most from this guide, you should already know the basics of Amazon Advertising in terms of terminology, features, and available options.
Table of contents:
1. Key Metrics: ACoS vs. ROAS – Understanding Profitability
To put this guide into practice, you must first understand how to evaluate and improve a few key metrics on Amazon Advertising.
The core metrics that advertisers live by are ACoS and ROAS, and knowing your break-even point for these metrics is fundamental to running profitable campaigns.
What are ACoS and ROAS?
Advertising Cost of Sales (ACoS) measures the percentage of your ad spend relative to the sales generated. It’s calculated as:
ACoS = (Ad Spend ÷ Ad Revenue) × 100%For example, spending $50 on ads to generate $100 in sales yields an ACoS of 50%.
An ACoS of 25% means you spent $0.25 in advertising for every dollar in sales (source: advertising.camazon.com).
There is no universal “good” ACoS; whether an ACoS is good or bad depends on your profit margins and business goals.
Return on Ad Spend (ROAS) is essentially the inverse of ACoS. It measures how much revenue you earn for each dollar spent on ads. The formula is:
ROAS = Ad Revenue ÷ Ad SpendUsing the previous example, $100 in sales on $50 ad spend gives a ROAS of 2 (or 200%).
A 25% ACoS corresponds to a ROAS of 4x (400%).
Higher ROAS means more revenue for each ad dollar (more efficient spend), whereas higher ACoS means a larger portion of revenue spent on ads (less efficient).
In general, a high ROAS indicates strong ad performance, while a high ACoS indicates weaker performance – but both must be judged against your profit margins.
Key difference: ACoS is typically expressed as a percentage (the lower, the better for profitability), while ROAS is expressed as a ratio or multiple (the higher, the better). They are two sides of the same coin, so you can use whichever you prefer for analysis. (Many Amazon reports now include ROAS by default, which I personally find more intuitive.)
Calculating Break-Even ACoS and ROAS
Before you set any targets, determine your break-even ACoS – the ACoS at which you neither make nor lose money on ad spend.
Break-even ACoS is directly tied to your product’s profit margin.
To calculate it, first compute your product’s unit profit margin (before advertising). This is simply:
Profit per Unit = Selling Price – Cost of Goods – Amazon Fees – Other costs (shipping, etc.)
Then,
Break-Even ACoS = (Profit per Unit ÷ Selling Price) × 100%Practical Example:
Consider a product (say, a shampoo) with:
Retail price: $20
Manufacturing cost: $5
Amazon fees: $3
Shipping cost: $5
The profit before advertising is: $20 – $5 – $3 – $5 = $7.
That’s your per-unit margin.
Divide this by the price ($7 ÷ $20) and you get 35%, which is the break-even ACoS.
In this example, if your advertising cost of sales stays below 35%, you’re making money; if it goes above 35%, you’re actually losing money on each sale after ad costs.
The inverse concept is break-even ROAS – how much revenue you need per $1 of ad spend to break even.
You can calculate break-even ROAS as:
Break-Even ROAS = 1 ÷ (Break-Even ACoS)Practical Example:
In the above example, break-even ACoS is 0.35, so break-even ROAS is ~2.86 (or 286%).
This means every $1 spent on ads needs to generate $2.86 in sales just to cover costs.
Another way to get this: divide the product price by the profit (e.g. $20 / $7 ≈ 2.86).
Any ROAS above 2.86 (or ACoS below 35%) is profit, and anything worse means a loss on ad spend.
Why is knowing your break-even so important?
Because it’s the baseline for your targets.
Many new advertisers think there is a universal “good ACoS” (often they’ll say something like “ACoS under 25% is good”), but in reality a good ACoS or ROAS is entirely product-specific.
A 25% ACoS might be fantastically profitable for a product with a 60% margin, but disastrous for a product with a 20% margin.
Always compare your campaign ACoS/ROAS to your break-even threshold for that product.
Tip: Use a spreadsheet like Google Sheets to list all costs per product – it will help you quickly compute profit margins and break-even ACoS for each item.
TACoS vs. Total ACoS, and TROAS vs. Total ROAS
You’ll often hear the term TACoS in Amazon advertising.
The “T” can refer to Target ACoS or Total ACoS, depending on context:
Target ACoS (TACoS) – the ACoS goal you set based on your profit objectives (essentially, your ideal ACoS). For example, if your break-even ACoS is 35% and you want a 10% profit margin after ads, you might set a target ACoS of 25%. This TACoS becomes a performance benchmark for campaigns.
Total ACoS (also TACoS in some usage) – this measures total ad spend as a percentage of total revenue (ads + organic). Total ACoS gives a “big picture” view of advertising efficiency for your entire account. It accounts for the fact that ads often drive additional organic sales. For instance, if your ads drive up organic rank, your ad-attributed ACoS might look unprofitable, but total ACoS (considering the organic sales lift) could still improve. Many sellers track Total ACoS over time to see if ad initiatives are helping overall business performance.
Likewise, TROAS can mean Target ROAS (the ROAS goal you aim for) or Total ROAS (total revenue divided by total ad spend, including organic sales).
They’re just the inverse expressions of the above.
For clarity in this guide, I will use TACoS/TROAS to mean your target ad cost or return goals, and specify “Total” when referring to the blended organic + ad metric.
Why look at total performance?
Because sometimes running ads at a loss on purpose makes sense if they boost your organic sales enough.
For example, launching a new product often involves spending aggressively (high ACoS / low ROAS) to gain ranking, which later pays off in organic sales.
To know if this strategy is working, you compare pre- and post-campaign total metrics.
If your Total ACoS goes down (or Total ROAS goes up) after launching ads, it means your ads are lifting organic sales significantly.
If Total ACoS stays the same or worsens, the ads might not be justified.
Pro tip: Set a Target profit margin for your product (e.g. “I want 15% net profit after ads”). Subtract that from 100% to get your Target ACoS (in this case, 85% if you were okay with very low margin, but more realistically you might say 15% profit + 20% Amazon fees + 30% COGS, etc.). Conversely, you can compute the Target ROAS: if you want 15% profit and break-even ROAS is 2.86x, your target might be ~3.3x or higher. These targets will guide optimization—more on that next.
Using Target ROAS/ACoS for Bid Calculations
Understanding your ideal ACoS/ROAS isn’t just for post-mortem analysis—it can directly inform your bidding strategy.
There’s a handy formula popularized by Michael Erickson Facchin (CEO of Ad Badger) to compute a target CPC bid for a given target ACoS (TACoS):
Target CPC = Average Order Value × Conversion Rate × Target ACoSIn other words, if you know the average order value (AOV) for a click on a certain keyword, and that keyword’s conversion rate (CVR) on your listing, you can back-calculate how much you should bid to hit your ACoS goal.
Practical Example
For example, suppose your product’s AOV is $50 and its conversion rate for a specific keyword is 10% (0.10, meaning one in ten clicks converts to a sale).
If your target ACoS is 30%, then:
Target CPC = $50 × 0.10 × 0.30 = $1.50A $1.50 bid should, on average, result in about a 30% ACoS (roughly $5 revenue per $1.50 spend, which is ~3.33x ROAS).
If you prefer thinking in ROAS terms, you can invert the formula:
Target CPC = (Average Order Value × Conversion Rate) ÷ Target ROASIt’s the same math since Target ROAS = 1/Target ACoS.
Practical Example
Using the above numbers, target ROAS corresponding to 30% ACoS is 3.33, and indeed $50 × 0.10 / 3.33 ≈ $1.50.
This formula helps ensure you never overbid or underbid relative to your profitability goal.
If you bid too high for a keyword’s conversion rate, you’ll exceed your target ACoS; if you bid too low, you might miss out on profitable traffic.
Many sophisticated advertisers use software or spreadsheets to constantly adjust bids toward this ideal CPC.
If you don’t have an automated tool that does this in real-time, grab my free Amazon Advertising Bid Calculator – a spreadsheet that calculates these bids for your targets it’s a great way to apply data-driven bidding without guesswork.
Remember, this formula requires having a sense of your conversion rate for each keyword or ASIN target.
Early on, you might use an estimated conversion rate (e.g. if your listing converts 10% overall, use that until you have per-keyword data).
Over time, as you gather real campaign data, you can refine these inputs.
Lastly, keep in mind that your target ACoS or ROAS is a business decision.
Some campaigns (like launches or branding campaigns) intentionally have a higher ACoS (lower ROAS) for strategic reasons, whereas others (like long-term profitable campaigns) you’ll want to run under the break-even ACoS (or above the break-even ROAS).
As we go through each campaign type below, we’ll discuss how to optimize with these targets in mind.
2. Sponsored Products Campaigns: Automatic Targeting
Sponsored Products are the bread-and-butter of Amazon PPC.
An automatic targeting Sponsored Products campaign is where Amazon’s algorithm decides when to show your ad, based on your product listing content and buyer search behavior.
Auto campaigns are easy to set up and great for discovery, but they require careful control and optimization.
When to use Auto campaigns: Use automatic Sponsored Products campaigns only after your listing is fully optimized (images, title, bullets, etc.) and you’re already getting some relevant traffic and sales. The reason is that auto campaigns let Amazon decide what your product is relevant for. If Amazon doesn’t have enough data (e.g. a brand new product with no sales history) or your listing keywords are a mess, an auto campaign can waste a lot of money showing your ad to the wrong audience. That’s why you typically avoid auto campaigns for brand new product launches until you’ve done some initial keyword seeding or gotten a few sales organically or via manual campaigns.
Objective and Main KPI
The main objective of an automatic Sponsored Products campaign is simple: generate sales.
Amazon will attribute sales to an ad click for up to 7 days after a shopper clicks (and up to 14 days if you’re a Vendor Central user).
So these campaigns are primarily measured on the direct sales they generate within that attribution window.
Primary KPI: Sales volume (and secondarily ACoS/ROAS, which tells you efficiency). Because auto campaigns cast a wide net, they’re often used to explore and find new profitable search terms or ASIN targets. So, you’ll look at what sales and orders the auto campaign drives, and then mine that data for targets to potentially move into manual campaigns.
Secondary KPIs: ACoS/ROAS and conversion metrics. You want the auto campaign to run at least near your break-even efficiency if possible, but early on you might tolerate a higher ACoS (lower ROAS) as you gather data. Think of auto campaigns as both a sales driver and a research tool.
How to Structure an Automatic Targeting Campaign
When you create an Auto campaign, Amazon actually gives you four sub-targeting groups by default:
Close Match: Targets shopper searches closely related to your product. (Your ad may show on search results for very relevant keywords.)
Loose Match: Targets shopper searches loosely related to your product.
Substitutes: Targets shoppers viewing detail pages of products similar to yours (competitors).
Complements: Targets shoppers viewing detail pages of products that complement yours (frequently bought together items, etc).
Each of these four is essentially a targeting group within the auto campaign, and you can set separate bids for each.
However, budget is only set at the campaign level – and that’s crucial for how we structure our campaigns.
Best practice: Create one auto campaign for each targeting group (Close, Loose, Substitutes, Complements).
This means you will have four campaigns (per product or product group), each one focusing on one type of auto targeting.
Why do this?
Because if you run all four groups under one campaign budget, a poorly performing group can eat up budget that you’d rather allocate to a better group, and you have no way to control it.
For example, if “Close Match” and “Complements” are performing great but “Loose Match” is performing poorly, there’s no mechanism in one campaign to funnel budget only to Close/Complements.
By splitting them, you can pause or lower budget on the underperformers without affecting the others.
Structuring it this way also makes optimization easier, since each campaign’s metrics will correspond to a specific targeting type.
Structuring it this way also makes optimization easier, since each campaign’s metrics will correspond to a specific targeting type.
When setting up each auto campaign, here are the steps and tips:
Campaign Name: Use a clear naming convention for easy management. For example:“[Product Name or Identifier] | [ASIN] | SPA | Close Match” (for the Close Match campaign), and similarly “... | SPA | Loose Match”, “…| SPA | Substitutes”, “…| SPA | Complements.”Here “SPA” indicates Sponsored Products Automatic. Including the product ASIN or a descriptor ensures you know what product is being advertised. Including the targeting type lets you see at a glance what each campaign is doing.
Tip: Consistent naming is very helpful, especially if you use bulk sheets or third-party tools later. I also like to tag campaign names with indicators like “NB” or “BR” for non-brand vs. brand campaigns, but for auto campaigns this is usually all non-brand traffic by nature. More on naming conventions in the manual campaigns section.
Bidding Strategy: When creating the campaign, you’ll be prompted to choose a bidding strategy: Dynamic bids – down only, Dynamic bids – up and down, or Fixed bids. I strongly recommend starting with Dynamic bids – down only for auto campaigns (and really for most campaigns initially). This setting lets Amazon lower your bids in real time for auctions where your ad is less likely to convert, which helps control costs. It will not increase your bids above your default. This gives you some protection from overspending on irrelevant traffic. In contrast, “up and down” lets Amazon raise your bid (up to +100%) for auctions it predicts are likely to convert, and lower for unlikely ones. While “up and down” can increase volume, it also removes some control and can lead to higher CPCs/ACoS if Amazon’s predictions are wrong.
My approach: start with Down Only.
If after a while you find you’re not getting enough impressions or sales, you can try raising your base bids, and if needed, switch to “up and down” on a specific campaign to see if it boosts volume.
But beginning with down-only ensures efficiency is prioritized.
Fixed bids are rarely used except in very specific cases or experiments; they turn off all Amazon adjustments.
Default Bid: Use Amazon’s suggested bid or something close to it for each targeting group initially. The auto campaign creation will show a suggested bid range for Close/Loose/Substitutes/Complements. It’s okay to start around the suggested value to get data. You can always adjust bids later (using our trusty bid calculator or based on performance data).
Ad Group and ASINs: Each auto campaign can contain one ad group (by default Amazon makes one). I usually name the ad group the same as the targeting group (e.g., “Close Match” ad group) for clarity. Add the product ASIN(s) you want to advertise.
Important: Don’t mix very different products in one auto campaign.
If you have variations that are close substitutes (say just color variations of the same item), it’s okay to include them together.
But disparate products shouldn’t share an auto campaign because Amazon will have a hard time targeting properly, and your search term reports will be mixed together.
Different categories belong in different campaigns.
Activate only the intended target: Here’s a quirk – during the campaign creation, Amazon doesn’t let you toggle off certain auto targets. All four will be enabled by default. What I do is create the campaign, then immediately go into the ad group and pause the three target groups I don’t want active. Essentially, if this campaign is for “Close Match,” go in and set Loose, Substitutes, and Complements to inactive (you can lower their bids to the minimum or pause them). This ensures that campaign only spends on Close Match. Repeat for the others accordingly. It’s an extra step, but necessary to isolate them.
Budget: Allocate a budget that makes sense for the product and the expected CPC. As a rule of thumb, I don’t like to run any Sponsored Products campaign with less than ~$10/day budget. If you spread yourself too thin (say $5/day) and the clicks cost $1 each, you may run out of budget after just a few clicks, which is not enough to gauge performance in a day. For auto campaigns, if each of the 4 targeting types is in separate campaigns, consider how much you want to spend on auto discovery overall and divide it, or more weight to certain ones if you expect them to perform better. For instance, Close and Substitutes often drive more sales, so you might give them a bit more budget than Loose/Complements.
Initial bid strategy tip:
Start with Dynamic Down Only and moderate bids.
Over the first week, gauge impressions.
If a campaign (say the Close Match one) isn’t getting any traffic, you can slowly increase the bid.
It’s all about finding the right balance of volume and efficiency.
Adjusting placements:
For auto campaigns, Amazon allows bid multipliers for placements (Top of Search and Product Pages).
At launch, do not tweak placement multipliers.
Leave them at 0% (no increase).
Adjusting placements should come later only if you have data indicating, say, Top of Search has a much better conversion rate and you are willing to bid more for it.
Changing these too early is a common mistake.
To summarize setup:
One targeting type per campaign, clear naming including ASIN and target type, start with dynamic down bidding, and use suggested bids.
This structure sets a strong foundation for control and optimization.
Common Setup Mistakes (Auto Campaigns)
Avoid these pitfalls when running automatic targeting campaigns:
Mixing Targeting Groups in One Campaign: As explained, don’t lump Close/Loose/Subs/Comp together under one budget. You’ll lose control and visibility. Separate them for granular management.
Mixing Different Products: Don’t put fundamentally different products in the same auto campaign. If you sell eyeliners and also a pencil sharpener, those will trigger very different search terms – separate them so Amazon can focus and you can negative-match appropriately by product.
Using Auto for Unoptimized/Launch Listing: If your product is brand new with zero reviews and no keyword context, launching an auto campaign on Day 1 can lead to Amazon showing your ad in random places (because it’s testing what works). This can burn money. It’s better to start with a more directed approach (like manual broad with careful keywords – see below) for a launch, then layer in auto once Amazon has some idea of your category and you’ve refined your listing content.
Adjusting Placement Bids Too Early: As mentioned, don’t rush to boost “Top of Search” placement in auto campaigns without data. Many advertisers prematurely add, say, +50% for top-of-search thinking it will help, but if your auto campaign isn’t profitable yet, boosting bids for prime placement could just amplify the losses. Collect data first, then if you see, for example, that your conversion rate at top-of-search is double that of rest-of-search, then you might justify a placement multiplier.
Not allocating enough budget: If you set a very low daily budget (like $5) and your clicks are $0.50-$1.00, you might only get a handful of clicks per day, which makes optimization painfully slow and inconsistent. Under-budgeting can also cause your ads to go off and on throughout the day (budget pacing), missing potential sales. It’s usually better to slightly over-budget and then trim back if needed, rather than severely under-budget a campaign.
No negative keywords at launch: While the real negatives come from analyzing search term reports (see optimization section), you might already know some irrelevant terms. For example, if you sell “running shoes” and your brand name is common, maybe you never want to show for “dress shoes.” You can add those obvious negatives at the start to prevent wasted spend. Auto campaigns allow negative keywords and negative ASINs at the ad group level.
Now that your auto campaigns are up and running, let them gather data for a few days.
The more aggressive your budget, the faster you’ll accrue data to act on.
Once you have some search term data and performance metrics, it’s time to optimize.
How to Optimize an Automatic Targeting Campaign
Optimization of auto campaigns generally happens on three fronts:
Search Term Harvesting (and Negative Matching) – Controlling which customer searches or product views trigger your ads, by mining the search term report and adding negatives or transferring terms to other campaigns.
Bidding – Adjusting the bids on the four targeting groups to hit your target ACoS/ROAS.
Budget Allocation – Scaling budget up or down based on performance.
Let’s break down the process:
Step 1: Analyze Search Term Report
After a few days (or longer, depending on volume), pull the Search Term report for your auto campaign, or simply go to the campaign in Amazon’s UI and look under the “Search terms” tab for the ad group.
This will show you exactly what user search queries or ASIN detail pages triggered your ad and what results (clicks, spend, sales) came from them.
Sort or filter the search terms by Spend (highest spend first).
This surfaces the terms that are costing you the most.
Look at each high-spend term/ASIN and check if it has conversions (orders) and what the ACoS is.
Add Negatives for Irrelevant Terms:
For any search term (or auto-targeted ASIN) that is clearly irrelevant to your product and has spent money (especially if it got clicks with no sales), add it as a negative.
Use the “Negative Targeting” feature at the ad group level to exclude it.
If it’s a search term (keyword), add as a negative keyword; if it’s an ASIN or a product category that showed up, add it as a negative product target.
For example, if your auto campaign for a shampoo showed your ad for “dog shampoo” and you don’t sell dog shampoo, add “dog” as a negative keyword.
Or if it showed on an unrelated product’s page, add that product ASIN as a negative.
Use Negative Phrase match for keywords when you want to block any search containing that word.
Use Negative Exact if it’s just that specific phrase that’s bad.
If in doubt, I prefer Negative Phrase for very broad irrelevant concepts, and Negative Exact for weird one-off queries that aren’t generically bad except that exact combo.
Amazon’s match types: phrase means it can be part of a longer query, exact means that exact query only.
For instance, if “travel size shampoo” isn’t relevant to you because you only sell full size, you might negative phrase “travel” so any query with “travel” won’t trigger your ad.
But if the query was “shampoo for bald head” and you think maybe your shampoo could be bought by bald people, you might not negative it unless you’re sure it never converts.
Competitor Brand Names:
In auto campaigns, a common tactic is to negative-match your competitors’ brand names.
Why?
If someone searches specifically for “Pantene shampoo” and you’re not Pantene, chances are low they’ll click your generic shampoo ad, and if they do, they might still prefer Pantene (hurting your conversion rate).
These clicks often become wasted spend.
In my experience, adding major competitor brand names as negative phrase matches in auto campaigns saves a lot of wasted budget (and also avoids showing your ad to shoppers who aren’t likely to buy your product) – essentially preventing ad placements where you’re unlikely to win the sale.
Identify Promising Search Terms:
Now look at search terms that did generate sales.
Sort by Orders or Sales.
Highlight those that have a good number of orders or an ACoS below your target (i.e. profitable terms).
These are your hero search terms discovered by the auto campaign.
For any such term that is both relevant and promising, consider “graduating” it: i.e., take that search term and create a new manual campaign targeting it explicitly (as an exact match, or phrase, etc. in a manual Sponsored Products campaign).
By doing so, you can allocate a dedicated budget and bid to that term and potentially get even more sales from it.
When you do graduate a search term to a manual campaign, you would usually add it as a negative exact in the auto campaign to avoid bidding against yourself.
This technique is often called search term isolation or keyword harvesting – letting the auto find the terms, then moving them to manual campaigns for scaling.
Example:
Your auto campaign for shampoo got 3 sales on the search “sulfate free shampoo for curly hair” at an ACoS of 20%, which is great.
You decide to create a manual exact match campaign for “sulfate free shampoo for curly hair” to really focus on it.
Once that manual campaign is up, you add “sulfate free shampoo for curly hair” as a negative exact in the auto campaign, so Amazon stops showing the auto ad for it.
This way, your manual campaign can take over that term entirely (giving you more control over budget and bids for it).
Clicks with No Sales:
Pay special attention to search terms that have a lot of clicks but zero orders.
These are budget-eaters.
For auto campaigns, a general rule of thumb: if a search term has, say, 10+ clicks and no sales, strongly consider negating it (or at least testing it separately if you suspect it could convert with a different approach).
The threshold of “how many clicks with no sale = bad” depends on your product price and typical conversion rate.
For a low-cost item that usually converts 1 in 8 clicks, 10 clicks no sale is a red flag.
For a high-ticket $500 item, 10 clicks no sale might be normal.
Use common sense and knowledge of your category.
If unsure, you can pause that search by negating it in auto, but perhaps try it in a manual campaign to see if giving it its own budget/bid yields a sale.
The advantage of that is you isolate it – if it still won’t sell, you can turn off that separate campaign without it bleeding your main auto budget.
After culling irrelevant terms and harvesting good ones, your auto campaign becomes “tighter” over time.
Early on it might have shown for 100 different searches; after negative matching, maybe it’s down to 50 relevant ones.
That means more of your budget goes to the terms that matter.
Step 2: Adjust Bids on Targeting Groups
Remember, in an auto campaign you have one bid for Close Match, one for Loose, etc.
Over time, you’ll notice patterns – e.g. maybe “Close Match” is yielding a 20% ACoS (or 5x ROAS) which is great, while “Loose Match” is at 80% ACoS (unprofitable).
You have a few options:
If a targeting group is consistently unprofitable even after you added negatives (it’s just inherently poor), you can lower its bid significantly to reduce spend, or even pause that group entirely if it’s hopeless.
For example, if Loose Match spent $20 and made 1 sale of $20 (ACoS 100%), try cutting the Loose bid in half.
If it still performs poorly, you might pause “Loose” for that product and focus on the others.
For a group that is meeting or beating your target (e.g. Close Match doing well), you might actually raise the bid to capture more volume, as long as it stays within profitable range.
If target ACoS is 30% and Close Match is running at 15% ACoS, you have room to scale up – increase the bid a bit to see if you can get more clicks and sales until ACoS approaches target.
Keep these bid changes incremental: a good practice is ±10-20% changes and then observe. Drastic changes (like doubling a bid or cutting to near zero) can reset how Amazon treats your campaign and might cause instability. Make small changes, watch for a few days, then adjust more if needed.
Also, consider the budget at this stage.
If your auto campaign is doing great (say hitting your TACoS/TROAS goals), don’t let budget limit it – you can increase the daily budget to allow more sales.
Conversely, if it’s performing poorly and you’re in the process of fixing it, you might temporarily cap the budget to limit damage until you optimize it.
Budget is a blunt instrument, though, so ideally you fix via negatives and bids rather than just starving the campaign.
Step 3: Rinse and Repeat
Optimization is not a one-time task.
For auto campaigns, you’ll want to review search term reports regularly – at least weekly if you spend a lot, or bi-weekly for lower spend – and keep adding negatives and harvesting winners.
Over time, your auto campaigns become a source of fresh search queries.
Amazon might start showing your ad for new related terms as trends change or as your product gains traction, so keep an eye on it.
Eventually, an optimized auto campaign will reach a relatively steady state: most of the obvious negatives have been added, and it’s consistently finding a few new keywords here and there.
At that point, your bid adjustments become the main lever to maintain target ACoS/ROAS.
Using auto for ongoing research:
Even when you have robust manual campaigns (which we’ll set up next), I recommend keeping auto campaigns running (at a lower spend) as a continual research tool.
They can catch unexpected keywords or new ASINs that manual campaigns might miss.
Optimization Mistakes to Avoid (Auto Campaigns)
A few warnings when optimizing auto campaigns:
Don’t pause an entire targeting group too hastily: If, for example, “Loose Match” looks bad after only 2 days and 5 clicks, don’t rush to shut it off. It might just need more data or a couple of negatives to turn around. Pausing too early could make you miss out on potential sales. Give each group enough time/spend (relative to your price) to judge properly. That said, if after significant spend one group is clearly inefficient, feel free to eliminate it and put that budget into better areas.
Broad Match Negatives – Be Careful: Amazon doesn’t have a “broad” negative, only phrase or exact, but phrase-match negatives can act broad in effect. For example, if you negative phrase “shampoo” in an auto campaign, you just killed almost everything because nearly every relevant search will have “shampoo”. Use phrase negatives only for terms you are absolutely sure you never want in any combination. It’s usually safer to use exact negatives for autos unless a word is universally bad (like “free” or “refill” if you don’t offer refills, etc.).
Small bid changes over time, rather than huge cuts/boosts: If your auto ACoS is high, don’t react by slashing the bid 80% at once. You might kill all volume and then learn nothing. Instead, step it down gradually (e.g. -15%, then see). Same for increasing bids – large jumps can spike spend unpredictably.
Don’t neglect search term review: Some people set an auto campaign and forget it. They end up with months of wasted spend on irrelevant terms. Make it a habit to regularly mine search term reports. Set a schedule (e.g., weekly on Monday and Thursday) to add negatives and pull promising terms into manuals. This kind of routine (often called a negative keyword “scrub” and keyword harvesting) is vital for ongoing success.
Competitor names = negative: I’ll repeat this because it’s a big one – in auto campaigns, add your competitor brand names as negative (phrase) early on. Unless your strategy specifically is to target competitor terms (which you would do more effectively in a manual campaign anyway), auto campaigns waste a lot of money on competitor queries where conversion is low. In my experience, this one step can notably improve auto campaign efficiency over time.
With a well-optimized auto campaign, you’ll have a stream of converting search terms and a controlled ACoS/ROAS.
The next step is to take those findings and build out manual campaigns to drill deeper into the best keywords and product targets.
3. Sponsored Products Campaigns: Manual Targeting
Manual targeting campaigns give you the most control: you specify the keywords or products where your ads should appear.
This section is crucial because it’s where you focus your budget on specific targets, bid strategically, and scale your profitable search terms.
By now, your auto campaigns (or prior research) should have given you a list of relevant keywords and ASINs to target.
Before diving in, keyword research is key.
If you haven’t done comprehensive keyword research for your product, do that first (check out my detailed guide on Amazon keyword research for step-by-step methods and tools).
A great tool for this is Helium 10.
Having a strong keyword list will guide how you structure campaigns.
Keyword Research and Clustering
You’ll likely end up with a large list of potential keywords – from broad terms to long-tail phrases.
Organize these into thematic clusters before creating campaigns.
A cluster is a group of keywords that share a similar intent or core term.
Example:
If you sell a lion’s mane mushroom gummy supplement, you might have a cluster around “lion’s mane” (e.g. “lions mane gummies”, “lion’s mane supplement”, “lion’s mane brain health”, etc.). Another cluster might be “nootropic gummies” or “memory supplement” if those are relevant angles.
If you sell an air duster, you might identify clusters like “keyboard cleaner” terms, “electronics cleaning” terms, and “car interior duster” terms. Each cluster represents a distinct use-case or context.
Clustering helps in two ways: campaign structure and optimization insights.
By grouping similar keywords in one campaign or ad group, you can allocate budget and set bids appropriately, and later when you optimize, you’ll notice if one cluster outperforms another (maybe “keyboard cleaning” terms convert great, but “car duster” terms don’t – you could then adjust budget between those campaigns accordingly).
So, step 0 is: have your keyword clusters ready (you can do this in a spreadsheet).
Each cluster will become one manual campaign (or at least one ad group).
Note: Always include a cluster for your brand terms and potentially your main product name or model. Branded keywords often have the best conversion because people specifically search for your brand. You’ll want to protect those by running ads on them – more on that later in “Branded vs Non-branded”.
Objective and Main KPIs (Manual Campaigns)
A manual Sponsored Products campaign’s objective can vary: it might be pure sales and ACoS/ROAS driven, or could be to drive rank for specific keywords, or defend your brand space, etc.
Generally, though, the primary KPIs for manual campaigns are ACoS/ROAS and sales for the targeted keywords.
Since you are hand-picking targets, you expect them to be relevant and potentially high-converting, so you measure success by how well those targets perform:
Are they hitting your target ACoS or better (meaning you can scale them up)?
Are they driving significant sales volume (helping you rank organically for those terms)?
Secondary metrics: Click-through rate (CTR) on your ads for those keywords (an extremely low CTR might indicate the ad isn’t appealing or the keyword isn’t as relevant as thought), conversion rate per keyword (which you can see in reports), and impression share (if you care how much of the available traffic you’re capturing).
Use cases for manual campaigns:
Focusing budget on proven search terms: After running auto campaigns or initial research, you’ll have keywords that you believe convert well. Manual campaigns let you pour money specifically into those terms without interference from other random terms.
Ranking on strategic keywords: If you have certain search terms that are very valuable (high volume, or strongly tied to your product’s identity), manual campaigns (especially exact match) help you concentrate on winning those auctions. Running a single-keyword campaign for a top term is a common strategy to improve organic ranking for that term (through consistent sales).
Controlling ad spend per keyword: In manual campaigns you set bids per keyword. This means you can bid higher on your best keywords and lower on less important ones, rather than a one-size-fits-all bid in auto.
Brand defense and competitor targeting: Manual campaigns allow you to target your own brand terms (to occupy top spots when someone searches your brand) and also to target competitor product ASINs or category keywords more aggressively if you choose.
Testing match types: You can decide to run broad, phrase, exact matches separately to see how each performs (since Amazon treats the same keyword differently under each match type).
Now, let’s build the campaigns.
How to Structure a Manual Campaign
I recommend structuring manual campaigns in a very organized, almost modular way.
The key dimensions are: keyword cluster/topic, match type, and branded vs. non-branded, and often ASIN (product).
Here’s a structured approach:
One Campaign per Keyword Cluster: Each group of related keywords (each cluster) gets its own campaign. This ensures that the budget for that campaign is focused on a coherent set of terms. For example, Campaign 1 might be “Lion’s Mane Gummies” cluster, Campaign 2 “Nootropic Supplement” cluster, Campaign 3 “Memory Booster” cluster, etc.
Separate Campaigns for Branded vs Non-Branded: Your brand terms (including product names if they’re unique) should be isolated in their own campaign(s). These terms usually have much lower ACoS and you don’t want them mixing with generic terms which typically perform differently. Also, if you later use portfolio budgeting or want to see how much of your sales are from people searching your brand, it’s easier if they’re separate. So, for example, have one campaign just for “[YourBrand] [Product]” keywords exact/phrase, labeled as branded.
Single ASIN focus: Ideally, each campaign should advertise one ASIN or a tightly related family (like a parent with its color variations). Don’t put totally different products in the same keyword-targeted campaign. You want all keywords in a campaign to be relevant to all products in that campaign. If you mix products, your ad may show an irrelevant variation for a keyword, hurting CTR/conv.
Match Type Segmentation (Siloing): This is important. Within each campaign/cluster, do not mix different match types in one ad group. For instance, if you have a list of 20 keywords related to “keyboard cleaner,” you might create two or three ad groups: one for exact match, one for phrase, one for broad (if using broad). This way, performance data isn’t conflated and you can adjust bids knowing exactly which match type is working. I often even separate exact and phrase into different campaigns (especially if high volume) because budget distribution can favor one over the other. But at minimum, separate ad groups. This practice of isolating match types is often called “match type silo” strategy
Another approach is to have separate campaigns for each match type of the same cluster.
Example: “Keyboard Cleaner – Exact” as one campaign, “Keyboard Cleaner – Broad” as another.
The reason to separate into distinct campaigns (not just ad groups) is again budget control: Amazon budgets at campaign level, so if one match type eats budget faster, separate campaigns ensure each gets its full share.
Broad Match Modified (BMM) usage:
Amazon allows a modified broad match using a + before words (similar to old Google Ads).
For example, “+mushroom +gummies” means the search must contain both “mushroom” and “gummies” in any order.
This gives more control than plain broad, which could match to synonyms or related terms that don’t include the word (Amazon’s broad can be very broad, sometimes bordering on semantic rather than containing the same terms).
I suggest if you use broad match, use modifiers on the critical words.
In our example, “mushroom gummies” broad could show on “mushroom candy” (bad match), but “+mushroom +gummies” will not match “mushroom candy” because “gummies” must appear.
Essentially, BMM ensures the specified words are present in the shopper’s search query.
MATCH TYPE | KEYWORD | POTENTIAL SEARCH QUERIES |
Exact | "running shoes" | running shoes |
running shoe | ||
(minor plurals or misspellings only) | ||
Phrase | "running shoes" | best running shoes |
cheap running shoes for men | ||
black running shoes | ||
(must contain the exact phrase in order) | ||
Broad | running shoes | running sneakers |
shoes for running | ||
athletic footwear | ||
shoes for trail running | ||
(words can be reordered. synonyms can match) | ||
Broad Modified | +running +shoes | running trail shoes |
shoes for running | ||
men's running shoes | ||
(must contain all modified terms, in any order, even loosely) |
Known bug/caveat: Amazon’s implementation of broad match modifiers isn’t perfect.
Some advertisers (myself included) have observed cases where Amazon appears to ignore the + and still shows an ad on a query missing that term.
It’s not supposed to, but it can happen.
So, be vigilant.
Treat broad-modified almost like broad – watch the search terms coming in. If Amazon misbehaves, you may need to add negatives or even pull that keyword if it’s matching irrelevant stuff.
Still, BMM is generally more targeted than vanilla broad.
Campaign Naming:
As with auto, use clear names. Here’s a format I use:
“[Product Name or ASIN] | SPM | [Match Type] | [Keyword Cluster Descriptor]”.
“SPM” stands for Sponsored Products Manual.
For example: “Air Duster B0XXXXX | SPM | Exact | Keyboard Cleaner”.
If it’s phrase: “…| SPM | Phrase | Keyboard Cleaner”, etc.
If using broad-modified, indicate that (some people write “Broad+” or just Broad/M).
For campaigns containing brand terms, I include “BR” or “Brand” in the name, and for non-brand I may add “NB”.
E.g., “MyBrand Gadget | SPM | Exact | BR terms”.
This helps at a glance and allows filtering.
If you use tools like Pacvue or Skai, having “BR” vs “NB” in campaign names is extremely useful for applying different optimization rules and budget allocations to branded vs non-branded campaigns.
Budget:
As mentioned, plan for at least $10/day per campaign as a starting point.
But more importantly, consider the number of keywords and their search volume.
If one campaign has 50 keywords that are moderately high volume, $10 might be eaten by one or two of those keywords alone.
If you have limited budget overall, it’s better to limit the number of keywords per campaign so that each can get sufficient spend.
I recommend grouping maybe 10-20 keywords per campaign if budget is tight.
If you have an important keyword that has huge search volume (thousands per month) and the rest are low, consider even isolating that one into its own campaign (often called a Single Keyword Campaign or SKC).
Otherwise it could hog the budget and overshadow the other terms.
Practical Example:
For example, if in your cluster one keyword has 15,000 searches a month and the next has 3,000, that big one can dominate impressions.
By giving it a separate campaign, you ensure the other 3k volume terms in the original campaign get a fair share of budget.
Bidding:
During setup, set your bids per keyword.
Amazon will often show a suggested bid range for each.
I typically start bids around the suggested median, unless I know the term is super critical (then I might start on the higher end to ensure I show up).
Remember to set bids according to match type – exact might warrant a higher bid than broad because exact won’t get as much volume and is more precise. Broad I often bid lower because it can match to lots of variants, some less relevant.
Dynamic bidding strategy: As with auto, I recommend “Dynamic bids – down only” to start for manual campaigns. It helps prevent overspending if a keyword is off-target.
Once you have data, you can experiment with up and down if, say, you want Amazon to push certain keywords harder.
But down-only gives more controlled spending initially.
Placement multipliers:
For manual campaigns, you can also set placement bid adjustments (Top of Search, Product Pages).
At launch, keep them at 0% unless you have a specific strategy (like you absolutely want to appear top of search for a particular keyword and are willing to pay extra).
It’s generally wise to gather performance data by placement first, then you might find e.g. Top of Search conversion is 2x better than rest, which justifies a +50% bid for top.
Initially, though, let it run evenly.
Bonus – Broad Match “Exploration” Campaign:
If you decided not to run an auto campaign for a new product (perhaps because it has no data yet), an alternative is to run a Broad match manual campaign with carefully chosen seed keywords.
Essentially, you’re guiding Amazon’s exploration by providing the broad terms you want it to explore around.
Use broad match modified here heavily.
For example, rather than auto guessing anything, you might add “+wireless +earbuds” as a broad term.
Amazon then will only show queries that contain both “wireless” and “earbuds” (like “wireless earbuds for running”, “wireless bluetooth earbuds noise cancelling”, etc.).
This way you restrict the universe to something relevant.
This broad campaign can act similar to an auto campaign but with more directional control from you.
It’s especially useful in product launch phase where you have zero reviews – you might not want Amazon auto matching to competitor pages where you can’t compete yet.
Instead, you fish in a smaller pond of keywords.
And again, keep an eye out: sometimes Amazon’s broad match might still match things outside your plus terms due to quirks. Negatives are your safety net.
Common Setup Mistakes (Manual Campaigns)
Avoid these common errors when structuring manual campaigns:
Dumping all keywords in one campaign/ad group: If you throw 200 keywords of all different types into one campaign, two things will happen:
a) a few keywords (often the broadest ones) will consume most of the budget, leaving many with few impressions;
b) optimization becomes impossible to parse – you won’t easily know which group of keywords is dragging down performance or which are doing well. Always segment into logical groups.
Mixing match types in one group: As discussed, don’t put exact and phrase (and broad) all in one list. If “red mug” exact and “red mug” broad are together and broad triggers 90% of the spend, you might think “red mug” as a whole is bad, when maybe broad was matching to “coffee mug red” loosely and exact would have been fine. Silo them to isolate performance.
Not separating brand vs generic: Your own brand keywords will usually have ACoS near 5-10% (or ROAS 10-20x) if people search your brand specifically. If you mix those with non-brand (which might be 50% ACoS), the average could look okay, but you might actually be losing money on non-brand and it’s hidden by the cheap brand clicks. Also, if budget is shared, your brand terms could cannibalize budget (they get clicks easily and might use up the campaign budget early in the day). It’s smarter to isolate them. The same logic applies for defensive ASIN targeting (targeting your own listings to keep competitors from appearing) – separate those from offensive campaigns.
Too many keywords for the budget: If you can only afford $10/day, don’t target 50 high-volume keywords with it. You’ll just scratch the surface on many and not get useful data. It’s better to focus on a smaller set of the most relevant keywords so you can actually see results and optimize. You can always expand later once those are tuned or if you increase budget.
Not doing Single Keyword Campaigns (SKCs) for very important terms: This isn’t a mistake per se, but a missed opportunity. If one keyword is absolutely pivotal (either very high volume or very profitable), isolating it in its own campaign can maximize control. You can ensure it always has budget, tailor its bid without averaging across others, and easily track its exact performance. I often do SKCs for my top 5–10 terms.
Including irrelevant keywords out of hope: Sometimes sellers include marginally relevant keywords just to get more reach. E.g., including “hair conditioner” in a shampoo campaign because maybe some shampoo buyers search that. This usually backfires with wasted spend. Stick to keywords that truly match your product’s use case unless you have specific evidence to try a tangent. Use sponsored display or other methods for tangents.
Ignoring negative keyword overlap (campaign cannibalization):
When you have multiple manual campaigns, especially with broad and phrase match in play, you can have scenarios where the same search term is targeted by multiple campaigns.
For example, you have a Broad match ad group with “mushroom gummies” and an Exact ad group with “lion’s mane gummies”.
A user searches “lion’s mane gummies” – which campaign gets the impression? Potentially both are eligible. Amazon will choose one (often the one it thinks has higher ad rank at that moment).
This means you might be competing against yourself unknowingly, or one campaign’s data is getting skewed by queries that belong to another.
Solution: Use negative exact to silo campaigns.
A common practice: if you have a campaign for exact terms, add all those exact terms as negative exact in your broad campaign.
That forces the broad campaign to exclude any search that exactly matches one of your exact targets, leaving those to the exact campaign.
Similarly, if you have phrase and exact separate, add exact negatives to the phrase campaign.
This way each campaign/ad group has its own “territory” of queries.
Example:
Campaign A (Exact: “wireless earbuds”) and Campaign B (Broad: “+wireless +earbuds”).
Add negative exact “wireless earbuds” in Campaign B.
Now if someone searches “wireless earbuds”, Campaign A’s ad will serve (if it wins) and Campaign B won’t.
This prevents internal bidding wars and also means each campaign’s performance data is clean.
Once your manual campaigns are structured properly, you’ll start collecting data on which keywords are driving sales, which are clicking but not converting, etc.
The optimization process is somewhat similar to auto, but with more precision since we control each keyword/ASIN bid.
How to Optimize a Manual Targeting Campaign
Optimizing manual campaigns breaks down by match type group:
For Exact Match and ASIN Product Targeting ad groups:
These are straightforward: each target is explicitly what the customer must search (for exact) or which product page to show on (for product targeting).
There’s no ambiguity or variant matching (aside from plurals or close misspellings which Amazon auto-normalizes).
So optimization is mostly about bids and perhaps pausing/removing underperformers.
Steps to optimize exact/ASIN campaigns:
Look at performance per target: Go to the “Targeting” (or “Keywords”) tab for that campaign in Amazon ads UI (or download the report). Sort by Spend or by ACoS. Identify which keywords or ASIN targets have poor performance vs good.
Bid adjustments: A general rule of thumb I use: if a keyword has spent about your product’s price without a sale, consider lowering its bid by maybe 20-30%. If it’s spent double your product price with no sale, maybe lower dramatically or pause it (unless you have reason to believe a conversion could still happen, e.g. maybe 3 clicks at $2 each = $6 spent on a $50 item, that’s not alarming yet – always contextualize with clicks and CPC).
If a keyword’s ACoS is above your target, lower the bid. The goal is to bring ACoS down by reducing cost per click, even if it means fewer clicks. For instance, ACoS 50% but target 30% – try cutting bid by, say, 20%. Monitor if ACoS improves (though volume might drop, you care that when it does click, it’s at a cheaper rate so any eventual sale yields a better ACoS).
If a keyword is below your target ACoS (i.e., very profitable), you can raise the bid to grab more impressions. As long as it stays below target, you’re good. E.g. a keyword with ACoS 10% and target 30% – you might safely bid more to see if it scales, until maybe its ACoS rises closer to 30%. Essentially, you give more budget leeway to winners.
If a keyword has zero impressions or very low, your bid might be too low. You could gradually raise it or check if maybe your bid is fine but your budget capped out early (less likely with exact, more a broad issue).
Search term level check for Exact? For exact match, the search term is essentially the same as the keyword (except plural/singular, etc.). So there’s not much to mine – if an exact target “wireless earbuds” got clicks, they all came from “wireless earbuds” search. If you see weird search terms in an exact campaign, either Amazon considered them close variants or you accidentally had phrase/broad in there. Usually, exact is clean.
Adjust budgets if needed: If the whole campaign is doing great and often running out of budget midday, raise the budget. If it’s performing poorly and you’re in the process of fixing bids, you might temporarily cap budget to the bare minimum to limit losses until it’s optimized.
Product targeting optimization: For manual ASIN targeting campaigns (where you target competitor or complementary product ASINs), treat each target like a keyword. If some ASINs are burning money (people click your ad on that product page but don’t buy yours), consider excluding that ASIN (pause or lower bid drastically). If some ASIN targets convert well (maybe your product is a clearly better deal than a specific competitor), you could raise bids there or allocate more budget to that campaign.
Tip: Often when targeting competitor ASINs, look at their product reviews, price, etc. If you target a competitor who is much cheaper or much better-rated, your conversion might be low (people click but choose them). It might show up as a poor ACoS target. You can drop those and focus on ones where you have an advantage (e.g., target competitors who are higher priced than you, or with worse reviews, increasing the chance a shopper might switch to you).
For Broad and Phrase Match ad groups:
These require a mix of search term harvesting (like auto campaigns) and bid optimization:
Pull Search Term reports: Even though you chose specific broad/phrase keywords, Amazon will match them to various shopper searches. Go to the Search Terms tab under that ad group to see what queries actually triggered clicks/sales. This is essentially the same process as optimizing an auto campaign’s search terms.
Add negative keywords for any irrelevant search terms that crept in. If your broad keyword “wireless earbuds” matched to “wireless headphones over ear” and you sell only earbuds (in-ear), you might negative “headphones” (phrase) to prevent that. Clean out non-relevant stuff regularly.
Add negative exact for search terms that you decide to move to an exact campaign (harvesting). For example, your broad keyword “running shoes” might match a specific long-tail “best running shoes for marathon”. If that got sales, you might graduate that to an exact campaign, and then add “best running shoes for marathon” as negative exact in your broad group to avoid overlap. Broad keeps fishing for new terms, while your exact campaign handles that term now.
Monitor high-click no-sale terms: If a phrase/broad target is consistently matching to a particular search query that gets many clicks and no sales, you have the option to negative exact it in that same campaign to stop bleeding, even if the broad keyword itself might still be useful for other queries. This is how you refine broad campaigns over time – by stripping out the parts that don’t work.
Bid optimization for the broad/phrase keyword itself: This is a bit abstract because the performance of a broad keyword is an aggregate of many search terms. I generally optimize broad/phrase bids based on their overall ACoS. If a broad keyword has generated 5 sales with ACoS 40% (above my target 30%), I’ll reduce its bid to try to bring that down. If another broad keyword has ACoS 15%, I might raise it. You can also use a more granular approach by seeing which search terms under it are consuming most of the spend. For instance, if one broad keyword’s bad ACoS is entirely due to one particular search term, you could negative that term and maybe the broad will become profitable. But if the broad keyword as a whole just isn’t working (lots of variations tried, few sales), then lowering the bid or pausing it is fine.
Graduating to phrase/exact: Sometimes you use broad to explore synonyms or longer tails, and phrase to explore slightly less strict matching. Over time, you may find that certain phrases always convert and you can add them as exact targets in your exact campaign. The idea is to let broad cast the widest net (with modifiers to keep it on-topic), phrase a bit narrower, and exact the narrowest. Budget-wise, you often allocate more to exact since it’s efficient, medium to phrase, and keep broad on a tighter leash just for discovery.
Frequency of optimization: Just like auto, do weekly checks for search terms on broad/phrase. Add negatives and adjust bids frequently at first (e.g. weekly). As campaigns mature, you might move to bi-weekly or monthly fine-tuning. High-budget campaigns might even warrant twice-weekly optimization.
Branded vs Non-Branded optimization:
If you separated brand terms, you’ll likely find your branded campaign runs at a super low ACoS.
You might actually increase bids a lot on branded terms because you want to dominate those and you can afford to (they’ll still stay profitable typically).
Non-brand terms you’ll manage more tightly to hit targets.
Dynamic bids note:
If you kept dynamic down-only, Amazon won’t raise your bids ever.
If you find some exact keyword where you want to really push for top-of-search, you might either manually raise the bid or consider switching that campaign to up and down.
With up and down, Amazon might increase your bid by up to 100% for searches it thinks likely to convert (based on context). Just be aware, up and down effectively lets Amazon spend more of your money, so use it when you’re comfortable with the campaign’s performance or when you need more aggression.
As a rule of thumb, I keep manual campaigns on down-only bidding. If I want to gain Share of Voice (SoV) on a keyword or dominate top-of-search, I usually push using the Top of Search bid modifier instead of switching to up and down. It gives me more control — and avoids the usual ACoS spike that comes with Amazon’s “help.”
One more thing on Manual Product Targeting optimization.
This is similar to exact keyword.
Look at each targeted ASIN’s results.
Did it get clicks and no sales?
Maybe lower the bid or remove it.
Did one ASIN give you 5 sales at good ACoS?
Maybe bump its bid to show there more often.
You can also refine by using Amazon’s reports that show “advertised product report” or “purchase product report” to see if maybe your ad was clicked on competitor pages but the sale happened on a different ASIN, etc., but that’s advanced analysis.
Generally, keep the ASIN targets that work, drop the ones that don’t.
And consider product attribute differences – if none of the cheaper competitor ASINs convert, maybe focus on those where you have a price edge.
Optimization Mistakes to Avoid (Manual Campaigns)
Be careful of these when optimizing:
Over-optimizing too fast: If you cut bids too low, your ads might stop showing entirely. There is a threshold below which Amazon will give you almost no impressions. So don’t panic and drop a bid from $1 to $0.20 in one go because ACoS was bad for a week. You might kill the momentum and then never know if it could have improved. Gradual changes allow you to find the sweet spot.
Pausing keywords prematurely: Similar to auto, if a keyword is relevant but hasn’t converted yet, consider the context. Maybe it got 7 clicks and no sales, but if those clicks were $0.20 each and your item costs $100, that’s nothing – it might convert on the 10th click. If it’s truly relevant, a better approach than pausing can be to lower the bid to reduce cost and continue giving it a chance. When you pause a keyword entirely, note that if you resume it later, Amazon treats it somewhat like new (loss of history). Better to keep it active but low-bid if you think it has potential later (perhaps after you improve your listing or reviews).
Not considering listing issues: If a relevant keyword is getting a lot of clicks but no sales, don’t just blame the keyword. Check your listing when someone searches that term – is your product the right fit? Maybe your price is too high for that term’s audience, or maybe your main image isn’t appealing for that context. For example, if the term is “gift coffee mug” and lots of people click but no buy, perhaps your packaging isn’t gift-ready or your design isn’t what they expected for a gift. So sometimes the fix isn’t to negative the term, but to adjust your offer or content. If you can’t fix it immediately, you might down-bid and come back to it after changes.
Chasing rank blindly: If you intentionally overspend on a keyword for ranking (a strategy some use: accept high ACoS to boost organic position via “SEO rank juice”), be aware of the cost. That’s fine as a deliberate play (in launches especially, you might run at 100% ACoS or worse on a top keyword to climb to page 1). Just isolate those campaigns and keep track so you don’t accidentally count them as “failing” when actually it’s an investment. Generally, outside of that context, keep your targets in sight – profitability or awareness etc.
Forgetting to negative-match between campaigns: If you implement broad and exact in parallel and forget to add negatives to prevent overlap, you might see weird data. For instance, your broad campaign might be eating all impressions of a certain exact term because maybe its bid was higher. Then your exact campaign looks like it’s not performing (no impressions) even though you thought that term was important. This can lead you to think “broad is doing fine” but maybe broad had lots of waste under the hood you didn’t see individually. To avoid confusion, isolate via negatives as discussed. It’s a one-time setup task that clarifies things tremendously.
Not monitoring search term reports for phrase/broad regularly: Some think manual = set keywords and done. But phrase and broad can generate just as much junk as an auto campaign if not monitored. Keep up the negative keyword additions on a schedule.
Ignoring new keywords: Over time, trends change. A slang or new keyword might emerge in your niche. Don’t assume your initial list covers everything forever. Use tools periodically to see if any new search terms are gaining popularity that you should add to campaigns. Also, check your auto campaign (if running) and broad match reports for new gems.
By diligently managing your manual campaigns, you’ll drive your ACoS down and ROAS up on the terms you care about, and scale up the spend on terms that are profitable or strategically important.
Next, let’s move to Sponsored Brands (headline search ads), which add a new creative dimension to your Amazon advertising.
4. Sponsored Brands Campaigns
Sponsored Brands (formerly Headline Search Ads) are those banner ads that appear at the top of search results (and elsewhere) featuring your brand logo, a custom headline, and multiple products, or sometimes a video.
These are great for showcasing your brand and product line, and they often serve a slightly different purpose than Sponsored Products.
While Sponsored Products are very intent-driven and conversion-focused, Sponsored Brands ads are fantastic for brand awareness, storytelling, and driving shoppers to explore more of your catalog or brand page.
They come in a few flavors:
Product Collection (banner with 3 products and a custom image/logo/headline),
Store Spotlight (if you have a Brand Store with subpages, you can highlight those), and
Sponsored Brands Video (a video ad linking to a product). We’ll primarily discuss Product Collection and Video since those are most common.
Objective and Main KPI
The objectives for Sponsored Brands can be multifaceted:
Brand Awareness: These ads occupy premium real estate and can catch the eye. A big use-case is to make shoppers aware of your brand’s presence in a category. For example, if you sell coffee and your brand isn’t known, a Sponsored Brands banner can introduce “Acme Coffee Co. – Premium Arabica Blends” at the top of results, even if the shopper searched a generic term.
Multi-product Promotion: You can showcase 3 complementary products. Great for cross-selling (“get the whole set”) or offering variations (so user picks which they like), etc.
Store Traffic: If you have an Amazon Store (highly recommend creating one if you’re brand registered), Sponsored Brands can drive traffic to it. The Store can then educate and hopefully convert the shopper on one of your products.
Video engagement: The video format (which appears in search results mid-page typically) can demonstrate your product in use or highlight its value prop in a way images and text can’t. It’s very powerful for conversion when done right.
Main KPI:
Amazon reports a 14-day attribution for Sponsored Brands (longer than Sponsored Products’ 7-day).
The attributed sales can sometimes be muddy because, with a 14-day window, a shopper might click your ad, not buy immediately, and then buy later after maybe seeing other marketing.
Because of this extended consideration period and the focus on awareness, Impressions and Click-Through Rate (CTR) are very important metrics for Sponsored Brands.
I typically consider impressions as the primary KPI for upper-funnel awareness (did my ad get seen a lot by my target audience?) and CTR as a measure of how appealing my creative and headline are.
ROAS/ACoS is still relevant, but you generally interpret them more loosely for SB.
Expect SB ACoS to be higher (ROAS lower) than Sponsored Products in many cases, because not everyone who clicks a brand ad is in an immediate buying mindset – some are browsing your store or watching your video then maybe buying later.
Amazon internally suggests using SB to drive Store engagement which can lift overall sales not directly tracked to that click.
So I often have softer ACoS targets for SB campaigns.
For example, if my SP target ACoS is 30%, I might be okay with 50% on SB as long as it’s driving lots of impressions and new visitors to my brand.
Total ROAS is worth checking: see if overall account sales rise when SB campaigns run – that indicates some of that exposure is paying off organically.
In short: Use Sponsored Brands to increase visibility and complement Sponsored Products. Judge success by reach (impressions) and top-of-funnel engagement, in addition to any direct sales.
Note: Amazon has a “New-to-Brand” metric for Sponsored Brands which tells you what percentage of orders coming from the ad are by customers who haven’t bought from you before. That’s a great indicator of whether your SB ads are recruiting new customers vs. just snagging repeat buyers.
Ad Formats and Landing Page Options
When creating a Sponsored Brands (SB) campaign, you choose the format and the landing page:
Product Collection (Image + 3 products): This appears as a banner with your brand logo, a custom headline, and up to 3 product images. When clicked, the ad can either take the shopper to a Store page or a Product List Page (basically a search results page featuring just the products you included).
Amazon strongly recommends linking to your Store if you have one. In fact, an internal Amazon study found that linking SB campaigns to a Store got 22% better ROAS on average than linking to a generic product list page. This makes sense: a Store is a curated experience with rich content, whereas a product list page is just a bland selection of items. The Store likely converts better and encourages browsing multiple products. So, if you have a Store, use it as the landing page for Product Collection ads (you can link either to your Store’s homepage or a specific sub-page in your Store relevant to the ad).
Store Spotlight: This format is specifically for highlighting up to 3 Store sub-pages. For example, your ad could promote “Shop Men | Shop Women | New Arrivals” if those are your store categories. Clicking each goes to that section of your Store. This is useful if you want to drive category-level traffic rather than to specific products.
Sponsored Brands Video: This is a separate ad format (but still created under the SB campaign flow). It shows a single video (with no sound by default, user can tap for sound) and typically one product’s info (image, price, stars) to the right of the video. When clicked, it usually takes the shopper to that product’s detail page by default.
Pro tip: Amazon now allows the video to link to your Store instead of the product page. This is a newer feature – it means you can have a video that maybe tells your brand story, and clicking it drops people into your Store where they can see the full lineup. This is a unique way to get vertical videos (like those from TikTok/Instagram) in front of Amazon shoppers and then direct them to a broader experience than just one product.
Use video ads when you have a compelling demonstration or a lifestyle clip that can stop scroll.
They often have higher CTR than banner ads because motion catches the eye.
Also, video slots on search are still relatively few, so competition can be lower; I’ve seen very efficient ROAS from video ads in many cases.
Headline and Creative: Don’t treat the headline lightly – it’s basically ad copy. Make it benefit-driven or call out your USP (Unique Selling Proposition). E.g. “100% Organic Coffee – Roasted Fresh Weekly” as a headline does more than just “Acme Coffee Company”. Use that space wisely to give a reason to click. Also ensure your brand logo looks good (Amazon will pull it from your brand profile) – clear and not cropped.
Choosing keywords vs category targeting for SB
When creating a Sponsored Brands campaign, you can either target keywords (like Sponsored Products) or product categories/ASINs (similar to auto or product targeting).
Keyword targeting for SB: You bid on search terms. It’s usually the way to go when you want your ad to appear on searches like “best coffee beans” or “running shoes for women” etc. You’ll choose broad/phrase/exact here too. Typically, broad and phrase are useful for SB to get reach on various related searches, and exact for the few core terms you know do well. I often start with a set of 10-30 keywords, mostly broad/phrase to cast a wide net for brand awareness, then trim via negatives later.
Category/Product targeting for SB: This places your SB banner on product detail pages (yes, SB can appear on product pages in certain placements, usually as a banner near the top or middle of the page). When you select category targeting, Amazon will show your SB ad on detail pages of products in categories you choose (with some refinement options like price range, star rating filter). This is more of a display advertising approach – it’s less granular control than keywords, but it can get your brand in front of people browsing similar products. I might use this if I have a broad catalog and want to retarget or reach people looking at competitors broadly.
Note that SB ads on product pages don’t always perform strongly for conversion – someone already on a product page might ignore a banner. But it’s an option if you want max visibility.
Bidding for SB: There is no “dynamic bidding” choice for Sponsored Brands like Sponsored Products. You set default bids for each keyword or target, and Amazon uses them (with some algorithmic adjustments internally, but no setting for you). So bid wisely based on suggested bids and your goals. If you’re going broad for awareness, you might bid a bit lower to avoid spending too much on possibly less relevant traffic; if you have exact keywords that convert well, bid more aggressively.
Setting up a Sponsored Brands Product Collection campaign (example):
Let’s say I sell a line of skincare. I create an SB campaign titled “GlowSkin Brand – Moisturizer Line | SB”.
I choose Product Collection.
I select my Store as the landing page (specifically the “Moisturizers” subpage of my Store).
I add 3 products to feature: my best-selling moisturizer, a second variant, and maybe a complementary serum to upsell.
I upload a custom image (SB allows a background image now for the banner – highly recommend doing this because a lifestyle image can really improve CTR).
I input a headline like “GlowSkin – Hydrate and Nourish Your Skin Naturally”.
Then for targeting, I might add keyword targets like: broad “moisturizer”, phrase “face moisturizer”, exact “organic face moisturizer”, etc., maybe 20 keywords including some competitor brand names (SB allows targeting competitor brand keywords too – a common strategy for brand conquesting).
I set bids maybe at $1.50 for the broad, $1.75 for phrase, $2.50 for exact (just an example, actual suggested bids will guide you).
I launch that.
Now my banner might appear on searches like “moisturizer for dry skin” or “BrandX moisturizer” (if I targeted BrandX as a keyword).
Vertical Video SB example:
If I have a nice 15-second video of someone applying my moisturizer and smiling with glowing skin, I’ll create a separate Sponsored Brands Video campaign.
I’ll upload the video, link it to my main product page (or Store if I choose, but usually product page for video is fine unless it’s a brand story video).
I target some of the same keywords but perhaps focus on highly relevant ones (since video will only show on page1 for certain keywords).
I bid competitively because video slots are prime.
The creative advantage of SB cannot be understated – having a custom image or video that highlights your product’s use can increase click-through and also pre-qualify customers (they see what it is, so clicks are more likely to be genuinely interested shoppers, which can improve conversion after click).
How to Optimize and Evaluate Sponsored Brands Performance
Optimization for Sponsored Brands has parallels with Sponsored Products campaigns:
Search Term Harvesting: Just like others, you get a search term report for Sponsored Brands. Check what queries your SB ads are showing for and which ones drive clicks or sales. Add negative keywords for terms that are clearly not working or not relevant. If you did broad targeting, you might find some odd matches to eliminate.
Bid Adjustments: Adjust bids on keywords based on their performance. If certain keywords never get impressions, maybe raise bids or consider they might be too narrow compared to others. If some have a high ACoS, lower bids or negative them out if they’re outliers. Because SB can sometimes have sparse data (fewer clicks overall than SP), give it a bit more time or look at longer windows.
Impressions & CTR: If you’re getting very low impressions, your bids might be too low or your keywords too niche. Expand keywords or increase bids to get more visibility. If impressions are okay but CTR is very low (far lower than your Sponsored Products CTR for similar terms), that could indicate your creative or headline isn’t resonating. You might test a new headline or ensure your custom image is attractive and relevant. For video ads, a low CTR could mean the first few seconds of your video aren’t compelling enough (consider changing the thumbnail or the start of the video to something more eye-catching).
New-to-Brand Metrics: Amazon provides “New-to-brand orders” and “new-to-brand revenue” for SB. If your goal was to acquire new customers, see how you’re doing there. If NTB% is low (meaning mostly existing customers click/buy from the ad), maybe your targeting is skewed to your own brand or you’re reaching the same audience; to fix that, you might add more generic keywords or higher-funnel terms.
Store engagement: If you send traffic to your Store, check your Store’s analytics (in Amazon’s Stores interface, “Insights”). See how many views your Store got from the campaign and what the engagement was (page views, dwell time, etc.). If lots of people go but don’t click any products, maybe the Store page needs improvement (clear navigation, relevant products on that page, etc.). If they do browse and click, great – even if they didn’t buy immediately, you gave them more to consider.
Time Lag Consideration: With 14-day attribution, it’s useful to let campaigns run a couple of weeks before making harsh judgments on ACoS. Someone might click on Day 1 and purchase on Day 10, and that sale will suddenly appear in the campaign’s metrics later. If you kill the campaign too soon, you might not capture that. So when evaluating SB campaigns, look at at least a 2-week window (or longer) to judge ACoS/ROAS.
Budget: If your SB campaigns are doing well (in terms of CTR and contributing to overall sales), don’t hesitate to raise budget. These often have lower default budgets, but if ROAS is within acceptable range, more budget = more brand exposure. Conversely, if ACoS is too high and you’re experimenting, you might cap budget at a reasonable level until you optimize better so it doesn’t overspend.
Advanced: Audience retargeting in SB: Sponsored Brands now allows Audience bid adjustments (a new feature Amazon released). For example, Amazon introduced a built-in audience for “New-to-Brand shoppers” where you can boost bids to specifically reach people who haven’t bought from you before (source: Amazon Advertising). Using this, you could say “I want to bid 50% higher for impressions to new customers.” Amazon reported that advertisers who aggressively bid for new-to-brand in SB saw significantly more new customer orders. This is a bit advanced, but if available in your account, consider using it if new customer acquisition is a priority. (You’d find this under campaign settings or bid adjustments in SB campaigns – there might be a tab for audiences where you can toggle on New-to-Brand and set a bid boost, e.g. +100% for NTB shoppers.)
Category targeting optimization: If you used category targeting in SB, you can actually exclude certain brands or refine categories. For instance, if targeting the “Facial Skincare” category, you might exclude luxury brands if you’re a budget brand (or vice versa), to make the spend more efficient. It’s a bit of trial and error – if category targeting isn’t yielding good results, consider switching that campaign to keyword targeting for more control.
How to evaluate success properly:
Since SB is somewhat upper-funnel, a holistic way to evaluate is:
Overall sales trend: Did total brand sales (and organic ranking) improve after introducing SB campaigns? If yes, even if SB ACoS looks high on its own, it may be driving indirect value.
Brand metrics: Are more people searching for your brand name after seeing these ads? (You can check Brand Analytics for top search terms for your products, or simply notice if your brand campaigns volume increases over time.)
ACoS/ROAS within reason: While I’m lenient, I still set a threshold. If a SB campaign is, say, 200% ACoS (0.5 ROAS) and not clearly driving other benefits, I’ll rethink it. But if it’s 70-80% ACoS and I see overall growth, that might be acceptable as a marketing cost for expansion.
Impression share: For key keywords in your niche, see if your SB is appearing. A common strategy is to use SB to dominate the top of search for important generic terms while your SP ads also appear below. This one-two punch can greatly increase chance of conversion (customer sees your banner, maybe clicks something else, then sees your SP ad too – you look like a big player). Check those keywords manually or via brand analytics (Search Query Performance dashboard) to see if you are capturing a good chunk of impressions/clicks.
3 Reasons for Poor SB Performance (and fixes):
Low Conversion Rate: If people click your SB ad but don’t buy, the issue could be your landing experience. Maybe the Store page or product page they land on isn’t convincing (maybe the products featured aren’t the best match to what the ad promised). Or your reviews/price aren’t competitive – SB ads often get broader audiences who might compare you to others. Fix: ensure the landing page is optimized – e.g., if your ad headline touts “Natural Ingredients”, make sure the Store page immediately shows that benefit and the products featured align with that message. And of course, product fundamentals (good reviews, competitive pricing) need to be in place, or SB will just drive awareness but not sales.
Low CTR: If hardly anyone clicks the SB ad (impressions high, CTR low), then either the creative is not appealing or targeting is off. Maybe your ad headline is generic or your image isn’t attractive. Or you targeted a broad keyword that isn’t really relevant to your brand/products, so people see your ad and ignore it. Fix: A/B test a different headline or image. Ensure you highlight a compelling offer (“20% off” or “New Arrival” or a strong benefit). Also, refine targeting to more relevant terms or audiences so that when people see the ad, it aligns with their search intent.
Low Impressions: If you barely get impressions, you might be bidding too low, or your keywords are too specific/low-volume. Also ensure your campaign is approved (SB ads get reviewed especially if a custom image is used; any disapproval would zero out impressions). Fix: Increase bids, add more keywords (especially broad ones) to expand reach, or raise the daily budget if it’s too limiting. Check for any policy issues (Amazon will usually notify if the ad was not approved or has limited serving).
Sponsored Brands, when done right, can boost the effectiveness of your Sponsored Products by keeping your brand top-of-mind and capturing shoppers at different stages. They require good creative and careful messaging, but they can pay off in building a brand presence on Amazon that sets you apart from competitors.
5. Sponsored Display Campaigns
Sponsored Display (SD) campaigns are Amazon’s self-serve display ads, which allow you to reach audiences both on Amazon (product detail pages, etc.) and off Amazon (through Amazon’s network) with banner and display placements.
They are a newer ad type (introduced after SP and SB) and serve slightly different purposes: primarily retargeting and product targeting with display creatives.
Think of Sponsored Display as Amazon’s answer to retargeting ads you might run on Google Display Network or Facebook ads – but simplified for Amazon sellers.
Use Sponsored Display to supplement your search ads:
Attack competitor pages where you have an edge.
Retarget shoppers who showed interest.
Possibly reach new audiences via category or interest targeting.
But do so knowing it’s a more upper-funnel tool (especially off-Amazon retargeting) and keep an eye on efficiency so it doesn’t drain budget without returns.
Objective and Main KPI
Sponsored Display can actually serve a couple of objectives depending on how you set it up:
Competitor targeting / Product targeting (on Amazon): Here your SD ads appear on competitor product pages, much like a Sponsored Products ad could, but often in banner form under the “Add to Cart” or mid-page sections. The goal is to steal shoppers from those pages or at least make them consider your alternative. This is more of a conquesting/competitive strategy. KPI in this case is similar to Sponsored Products – direct sales from those placements (so ACoS/ROAS is a primary metric, as well as absolute sales and click volume).
Retargeting / Views Remarketing: This setting targets shoppers who have viewed your product (or similar products) in the past but didn’t purchase. Amazon can show them your ad later when they browse Amazon or even on third-party websites/apps that are part of Amazon’s ad network. The goal is to re-engage past visitors and remind them to buy. KPI here is a bit trickier – conversion rate might be lower, but if you can recapture some lost prospects, it’s valuable. You’ll look at ROAS but also perhaps new-to-brand orders or overall lift.
Audiences (interest or lifestyle targeting): Amazon also offers targeting of certain predefined audiences (like “in-market for laptops” or “beauty enthusiasts” etc.) under Sponsored Display. This is more awareness focused – reaching people who haven’t seen your product yet, based on behavior patterns. The main objective there is upper-funnel discovery. The primary metric could be impressions and maybe click engagement, since direct sales might be low.
However, to keep it straightforward, Amazon provides two main targeting modes for SD: Product Targeting and Views Remarketing (there’s also Interests and Audience demographics if enabled, but let’s focus on the main ones the user asked about).
Product Targeting (SD): This is analogous to Sponsored Products Product Targeting, but in display format. Your ad appears under the Buy Box or below the bullets on Amazon product pages. When setting up SD with product targeting, you pick specific products or product categories to target. The placements are pretty aggressive (just below the “Add to Cart” box is prime). Conversion window is 14 days like Sponsored Brands. I evaluate these kind of like Sponsored Products in terms of KPI: I measure their ACoS/ROAS on a 14-day attribution. If I’m targeting competitors, I want to see if I can get sales at a decent cost. Main KPI: Sales and ACoS (for that 14-day window).
Views Remarketing (SD): This targets audiences who viewed certain products. When setting up, you choose either “viewed my products” or “viewed similar products” and a lookback window (7, 14, 30, 90 days, etc.). These ads can show on Amazon (like the right side of the homepage, on detail pages as banners, etc.) and off Amazon on websites/apps. The conversion rate for these is typically low (as is common with retargeting – many people who looked didn’t buy for a reason). Amazon even allows optimizing these for “page visits” or “conversions”. If you optimize for “reach” or “visits”, Amazon will try to get impressions/clicks cheaply (possibly sacrificing conversion focus). If for “conversions”, Amazon will try to show to people it thinks likely to buy. I usually choose “optimize for conversions” to let the algorithm focus on likely buyers.
Main KPI:
Total impact on sales, because some conversions might not be directly attributed if they buy organically later.
But within ads, I look at TACoS (total ad spend vs total sales) after launching these to see if my overall cost of advertising as a percentage of sales improved.
The user mentioned one way to gauge if retargeting is working is to watch TACoS before/after launching them.
If TACoS improves (goes down), it could mean those “unprofitable” ads actually drove organic sales later.
Also, check detail page view rate or click-through on these.
If hardly anyone even sees/clicks them, perhaps your ad creative (it auto-generates from your product info, but you can add a custom image now for SD too) isn’t enticing.
One important note: Sponsored Display ads can go off-Amazon into Amazon’s network. Amazon recently added a control toggle for SD as well where you can choose “Expand reach to Amazon audiences off Amazon”. If you allow that, your SD ads might appear on news sites, apps, etc., targeting the audience you chose.
That can dramatically increase impressions, but clicks from offsite might not convert as well as on Amazon (similar to the SP offsite discussion).
We have the ability in Sponsored Display to actually see some breakdown (the campaign report will show placements “Amazon” vs “Amazon Publisher Network” or similar for offsite).
So for retargeting SD, main KPI for me is total sales lift or at least some conversions that wouldn’t have happened otherwise. It’s more of a supportive campaign type.
For competitor-target SD (Product targeting SD), main KPI is ACoS/ROAS on those sales, because that’s a direct spend vs return comparison on those placements.
How to Structure a Sponsored Display Campaign
Sponsored Display campaigns are simpler to set up, but here are best practices:
Choose your strategy (Product vs Remarketing) and don’t mix: If you want to do competitor/product targeting, make that a separate campaign from any views remarketing.
The interface likely separates them anyway.
They have different optimization approaches.
Product Targeting Campaign:
Start by advertising your top-selling ASINs only.
Why?
Because if you put your whole catalog in, some weaker products might spend without converting (especially if they have fewer reviews or aren’t price-competitive).
Your best sellers have the best chance to convert a competitor’s customer.
Example of Product Targeting Campaign
So create a SD Product Targeting campaign for, say, ASIN ABC123 (your best seller).
Choose product targets:
Perhaps start with a list of 20-50 competitor ASINs that you know you have an advantage over (cheaper, better rated, etc.). You can search on Amazon for your category’s top sellers or use a tool like Helium10’s Xray to grab a list of competitor ASINs.
You can also target a whole category with filters (e.g. target “Skin Moisturizers” category, filtered to products priced above $30 if yours is $20, meaning you target more expensive ones). This casts a wide net. It’s a good expansion after you try specific ASINs.
Consider separate campaigns per product or per small group of products if they are different category or performance.
For example, I wouldn’t combine a premium item and a budget item in one SD campaign if their target audiences differ.
Bid & Budget: These placements can sometimes get pricey.
I start with a moderate bid (Amazon gives suggested bids for SD too).
And I usually don’t allocate a huge budget to one SD campaign initially – maybe $10-20/day – because I want to see performance first; SD can spend fast if off-Amazon placements kick in.
You can always scale up.
Also, note SD product targeting has an option to optimize for “Page Visits” or “Conversions”. Choose “Conversions” typically, because you want sales, not just eyeballs.
Views Remarketing Campaign:
For retargeting, decide:
Do you want to retarget people who viewed your product (to try to win them back)? That’s “Views – your advertised products”.
Or do you want to retarget people who viewed similar products (not yours), which is more like prospecting new customers who considered things like yours? That’s “Views – similar to advertised products”.
You can also target past buyers of your product via SD by creating an audience of “people who bought X also buy Y” or using Amazon Marketing Cloud custom audience. But for now, assume basic retargeting.
I typically start with “views my products” for 7 days.
That hits people who saw my product in the last week but didn’t buy.
These are warm leads – maybe they got distracted or were waiting.
7-14 days is a good window; beyond that, the intent might have faded or they bought something else.
Conclusion & Next Steps
Amazon advertising is a journey, not a set-and-forget task.
We’ve covered everything from calculating your ideal ACoS/ROAS and setting the right bids, to structuring campaigns for different ad types, to advanced tactics for seasoned advertisers.
The key takeaways to remember:
Know your numbers: Always start with your profit margins and break-even ACoS/ROAS. They are your compass for all optimization decisions. Whether you use ACoS or ROAS, understand how they relate (ROAS = 1/ACoS) and set targets that align with your business goals. Use tools (like the provided calculators) to take the guesswork out of bidding and budgeting.
Structured campaigns win: Separate automatic from manual, separate match types, separate branded from non-branded. A well-organized account (with clear naming conventions) will save you headaches and money. It lets you pinpoint what’s working and what’s not, and scale or cut accordingly.
Start with control (Down-Only bids) and adjust as needed: Beginning with a conservative bidding strategy (down-only) gives you cost control. You can always layer on more aggressive tactics (up/down bidding, placement multipliers, bid boosts) once you have data. This phased approach prevents overspending early on and ensures you have a baseline performance to improve from.
Optimize iteratively: Amazon PPC is very iterative. Continuously harvest search term data, add negatives, tweak bids. Small changes over time lead to big improvements. Avoid drastic changes without data support. Develop a routine (for example, weekly negative keyword sweeps and bid adjustments, monthly campaign expansion or restructuring). Being consistent with optimizations is often more impactful than one-time overhauls.
Leverage all ad types to complement each other: Sponsored Products drive the core sales, Sponsored Brands enhance brand visibility and can funnel new traffic (with a 14-day broader window), and Sponsored Display helps retarget and secure your brand presence on detail pages. Using them in concert – as we outlined in the phased strategy – creates a flywheel effect (SP drives rank, SB drives awareness, SD protects and retargets). Don’t rely on just one ad type; each has its role.
Measure holistically: Always look at the big picture – Total ACoS (or TACoS) is crucial. If you launch a bunch of ads and your ad account ACoS looks bad but your overall sales doubled, that might be a win (total ACoS would reveal it). Conversely, if ads look super efficient but total sales are flat, maybe you’re not pushing enough. In addition to Amazon’s metrics, measure things like organic rank improvements, changes in keyword share (Brand Analytics can show you share of voice for keywords), repeat purchase rates, etc. Advertising doesn’t exist in a vacuum – its ripple effect on organic performance and customer behavior is what ultimately matters.
Before we wrap up, remember that successful Amazon advertising also relies on an optimized product listing and a quality product.
Ads can drive traffic all day, but if your detail page or product quality is lacking, conversion (and thus ROAS) will suffer.
So ensure your product titles, images, bullets, A+ content, and price are competitive.
Many optimization clues come from your ads: for example, if you get plenty of clicks on a keyword but no sales, that signals a potential listing issue (maybe your product isn’t meeting the expectation set by the keyword).
Use these insights to improve your listing which in turn boosts your ads performance – a virtuous cycle.
In conclusion, save this guide as a reference and step-by-step playbook.
Treat it like your personal “PPC Bible.” When in doubt, refer back to the principles and tactics we covered:
Calculate, then spend.
Structure, then scale.
Analyze, then optimize (and repeat continuously).
Implement these best practices, test new ideas, and don’t be afraid to tweak the strategy to fit your unique business and category – as we said at the very beginning, the only correct strategy is the one that works for you.
This guide provides the foundation and the advanced arsenal; it’s up to you to apply, experiment, and refine until you find “the one that works.”
Now, ask yourself:
Which of these tactics were you missing before, and how many dollars might you save or gain by implementing them?
The Amazon ads landscape is competitive, but armed with this knowledge, you’re ready to tackle it like a pro.
Happy advertising, and I wish you high ROAS and skyrocketing sales! 🚀
Good luck, and let’s grow your Amazon business together!
Comments