[Amazon Ads] Why I Choose to go with a Granular Amazon Advertising Strategy

Working with Amazon Ads I have come across two schools of thought. One approach is to condense campaigns and the other is to get super granular with the campaign build-out for Amazon sponsored ads.

With a background in PPC, primarily focused on Google and Bing search and shopping, Google Display, and Facebook Marketing, I have always tried to go for the granular build-out so that I have bidding and budget control. In my experience, this is still relevant even though the above-mentioned ad platforms have become a lot better in machine learning allowing more reliance on the algorithm and shifting focus to messaging, creative and funnel optimization. However, this was not always the case, especially in the earlier days. Compared to Google you could say that the Amazon Advertising Platform is still pretty young.

When I started working with Amazon Ads and was taking over accounts, I quickly noticed that there was no delivery control when the Amazon Advertising Strategy consisted of too many ad groups and keywords are bucketed into a single campaign. This factor does not allow for even delivery distribution with budgets being spent on a small percentage of Ad Groups and Keywords.

Taking this into account, I will go over how I restructured the account and optimized it on an ongoing basis.

I would love to hear other Amazon Advertisers take on this. I am not here to say right or wrong, just sharing this strategy that I have been able to effectively use in expanding an account.

Here we go :)

A Brief Overview of my Amazon Ads Account Structure

First off, each product was grouped defining it by product type/category and equal or similar break-even ACOS in the Amazon Advertising Platform leveraging portfolios and a clean naming convention.

This allows me to have a good birds-eye view of how a product performs holistically across Branded, Automatic, Manual, Ranking, Win Back, Expansion, Converting, Targeted, Variation, Catch-All, Defense and Retargeting Amazon Sponsored Ad campaigns.

Initially, I had each product grouping under one respective portfolio for each but to manage the Amazon Ad account build-out and number of sponsored ad campaigns efficiently, I had multiple portfolios for each product grouping allowing me to better manage campaigns by campaign type. This is where a clean and consistent naming convention comes into play (see further below).

Ok so back to the portfolio structure. I will focus on the structure here. Further down, I will cover some of my optimization methodologies/workflows.

Within each portfolio, I would have the following for Sponsored Products and Sponsored Brands (Amazon Headline Search Ads) and it would look something like this:

Branded

Automatic

Manual

Ranking

Win Back

Expansion

Converting

Targeted

Variation

Catch-All

Defense

Retargeting

Branded Amazon Product Ads — A Necessary Evil

Yes, with Amazon Ads you will also have to pay for branded search terms to protect your brand.

A quick note on non-product specific branded sponsored ad campaigns since these do not fall under any specific product grouping. I always build out a Parent brand campaign including Broad, Phrase and Exact keywords of the branded search terms. Bids are set low to high in the above order.

In addition, I will also create a product-specific branded search term campaign (call it the Child Brand Campaign) within each product grouping that only goes after branded searches that include flavor, product type, size, color, style, etc. within each respective product grouping. The bid for this will be set $1 or $2 higher than the exact match keyword in the Parent brand campaign. Doing this allows me to serve a relevant sponsored product to branded search terms.

Note: I always run both Amazon Sponsored Products and Sponsored Brand (Amazon Headline Search Ads) campaigns for all brand campaigns.

Having branded campaigns grouped by ‘product type’ variations also allows me to indirectly see if expansion campaigns or high impression volume campaigns are buying mind share and are serving as a lift in branded sales.

Clarification: This is a directional, not deterministic measurement and brand + product or variation of product searches are measured as a blended ACOS, hence a build-out that can be measured across product groupings, for each product type. For example, this approach also allows me to run defense Amazon Product Ad campaigns at a higher ACOS ensuring the difference in efficiency is picked up by stronger performing lower-funnel campaigns.

The Manual approach to Automatic Campaigns

There are two ways to structure automatic campaign ads on Amazon. One is to create a single automatic campaign for each product grouping. Alternatively, you can create four automatic campaigns under each product grouping, one for each match type.

The results have been mixed. Automatic campaigns with history seemed to do better as a single auto campaign compared to the blended ACOS across match type campaigns.

However, I have seen this vary by product grouping across Amazon Ad accounts. The theory with a campaign break out by match type is that bidding and budget can be controlled on a granular level so that I can funnel the budget to match type campaigns that are driving sales and reducing budgets on the lower performers. Again, I have seen mixed results across campaigns with history and across different product types on which one did better so I would always recommend testing this.

Get Granular with Manual Campaigns

Super condensed campaigns never made sense to me. This Amazon advertising strategy does not allow for budget and bid control and more often than not I have seen Amazon allocate most of the ad delivery against 20% (on the high end) of keywords. Or on an Ad Group level, I have seen Amazon distribute 90% of the budget to one ad group severely under-delivering on others.

For this reason, I will build out Phrase and Exact campaigns (not Ad Groups) grouped by topic which determines where I move keywords from Auto and Broad campaigns when data mining for keyword winners.

So in some way, I do follow the traditional Broad, Phrase and Exact match manual campaign build-out. However, these are broken out as Campaigns not as Ad Groups so that I can control bids and budgets.

So moving on, keeping the topic grouping mentioned above in mind, my manual campaigns are built out as follows.

The broad campaign starts off with relevant search terms set at a lower bid (higher than auto but lower than phrase) to find winning keywords.

As I find winning keywords, I will move them to exact and phrase.

The phrase campaign will include the same keywords as the broad campaign to start, but once I find winning keywords this campaign will be built out with additional keywords as well.

I love the idea of finding a winning keyword in my search term report and maxing it out in an exact campaign, but data has told me this is not necessarily true. So much for simple. I don’t just add winning keywords to the exact campaign since I have often seen exact campaign keywords underperform but do really well in the phrase campaign and vice versa. If a keyword converts well and is hyper-relevant to what the user is looking for it will convert but naturally if working in a crowded category will also have many other players bidding on the same Exact keyword.

In addition, exact match keywords have lower volume and don’t necessarily account for minor variations of that keyword. Taking this into consideration, it makes sense to serve the same keyword in phrase campaigns for three reasons:

1. Higher search volume

2. Can catch more variations

3. (For now) less competition

Sidebar: Back in 2013 I saw a similar trend with Google Ads. Many best practices focused on a granular exact match build-out which still has its place today. But, as advertisers adopted this strategy going after exact match keywords became inefficient due to lower search volume (less inventory) and an increase of competition. This opened the conversation back up to going after phrase and broad match modifier keywords which had more volume and less competition. I wouldn’t be surprised if this same trend will happen on Amazon Advertising Platform since compared to Google, its ad platform is still relatively young.

I also don’t add winning Keywords as Negative Keywords to the Auto campaigns once they have been identified and placed in their respective manual campaign following the (somewhat) traditional method of Auto ➡️ Broad ➡️ Exact ➡️Phrase since a clear bidding hierarchy will ensure visibility on converting keywords lower in the funnel.

Create Ranking Campaigns To Support Product Listings (Aisle Take Over)

What I call Ranking campaigns for Amazon Ads consists of campaigns targeting focus keywords for the product listing. First, let me clarify what I mean by focus keywords. Focus keywords usually consist of approximately 20 keywords that are:

(1) Super relevant (ideally) high volume

(2) They have low to medium competition

These ‘focus keywords’ are integrated as far up in the listing as possible followed by a list of up to 100 keywords and keyword phrases that are integrated through your listings as secondary keywords.

Since I want to rank for the first group of focus keywords and given that conversion rates from paid and organic clicks impact product ranking, I will create ranking campaigns with an allowed ACOS increase so that I can make sure to be visible when users search for these terms. I do this leveraging both Sponsored Brands and Manual Campaigns targeting Exact and Phrase match keywords/campaigns.

Amazon reps have referred to this as Aisle a takeover. Essentially, you are capturing as much real estate as possible with Amazon Ads and your Organic listing on the top of the first page.

This has worked well, but you will still have keywords trying to eat up most of your budget. To manage this, I will incrementally bring down the bids to make sure the campaign is balanced, but if I cannot find the ACOS sweet spot for offending keywords I will turn these keywords off and move them to a win-back campaign.

Scrape Lower Placements for Relevant Keywords with Win Back Campaigns

A win-back campaign consists of keywords that are super relevant but have not been able to perform in a manual, automatic or ranking campaign.

These campaigns are grouped by product and pushed live with low bids. The theory is similar to catch all campaigns, covered further down, that with lower bids you might not get the best first-page placement, but you do get placement further down the page or on the second page which can convert into clicks and sales. These are not super high volume campaigns- but they do scrape the bottom of the barrel and drive sales with a profitable ACOS.

Expand Your Ads on Amazon with Expansion Campaigns

I would build expansion campaigns off of keyword research using Helium 10. For this campaign type in the, I look at keywords related to the product but not necessarily specifically the product I am advertising. This is not talking about the ‘audience targeting approach’ where you target people searching for products your target market would search for even if the product has nothing to do with your specific product. In contrast, I am more focused on closer offshoots. So if I sell keto bars I might also focus on sugar-free bars or healthy bars.

Audience targeting as I like to call it, on the other hand, is serving ads for people that are not necessarily looking for your product but fall under your target market. Similar to Facebook audience targeting where, for example, if selling a Jacket I would not target people that have an interest associated with Jackets but rather would target people that have expressed interest in Jacket Brands i.e. North Face or Stores such as Burlington Coat Factory. You can do the same with Amazon Ads. If you are selling diet food products why not try and run Sponsored brand ads to people looking for recipe books specific to that diet?

This build-out can get a bit crazy but it’s a good blanket approach to find winning keywords and keyword ‘topics’ that might be less competitive than super relevant keywords. It’s also a must to tap into once you start hitting an order/revenue volume plateau with Amazon Ads.

For expansion campaigns, I have typically built out one manual campaign for Exact and Phrase respectively with each campaign containing four ad groups and a maximum of 20 keywords each. The reason for this is that you want to make sure all keywords and variations get enough delivery exposure so that you can find the winners.

Sidebar: Contrary to many advertisers who strongly believe in condensing an account, I have seen keywords in condensed accounts get no delivery with only a small percentage of keywords in a given ad group getting any traction. I see that as a lot of missed opportunities.

The expansion build-out can quickly add up to many campaigns, as you can imagine, and to maintain this I would perform at least two optimization sweeps per week and use some form of automation to minimize wasted spend and manage my Amazon Advertising Cost. Given that you need enough click data to make accurate optimization, this will initially increase your overall ACOS so make sure to set the proper expectations, monitor your total ACOS (TACOS) and allow some breathing room so you don’t end up making too many reactionary decisions.

Buckle-Down on Low Hanging Fruit with Converting Campaigns

Converting campaigns are campaign breakouts for sponsored products and sponsored brands targeting converting ASINS and converting search terms pulled from the search terms report. I won’t go into each type here since they will be covered in their respective campaign types below.

Find the Silver Lining with Targeted Campaigns

We all know that Amazon is full of competitors and that if you are lucky enough to have an early movers advantage they will come soon enough. However, you can leverage their data using targeted campaigns.

Targeted ASIN campaigns are great for targeting similar or competitor products on Amazon. I break these down into two categories. Targeted ASINS for expansion which essentially consists of researching similar products and competitor ASINS and Targeted ASINS pulled from search term reports. The Targeted ASIN campaigns are built out the same way in the Amazon Advertising Console as I would build out Keyword Expansion Campaigns where I will have a maximum of four ad groups in each campaign with 20 ASINS in each ad group to ensure exposure.

As these campaigns start to deliver I will adjust or turn bids on/off once enough data has been collected. Statistical significance thresholds for optimization are set based on the number of conversions generated from all clicks (ASINS producing conversion rate + % allowance) relative to the product listings historical conversion rate.

Converting ASIN campaigns are built out using ASINS from search term reports. Again I follow the same build-out as above to ensure exposure. Converting search term campaigns follow the same structure as converting ASIN campaigns. Again I will follow the same structure as mentioned previously where I group four ad groups 20 keywords each for converting search terms pulled using historical data.

Both expansion and converting ASIN campaigns described above are for Sponsored Products but I will build out the same campaigns for Sponsored Brands (Headline Search Ads Amazon); I have seen this do really well for driving sales with the Sponsored Brands (Amazon Headline Search Ads) campaign type.

Create Relevance with Variation Campaigns

Product variation campaigns are very specific, targeting long-tail keywords when users are looking for a specific product flavor, size, color or type. These are not necessarily high volume Amazon Ad campaigns but this approach allows me to be very specific in what product gets served with super high relevance. Given the matching accuracy of this campaign type, I am usually not hesitant in holding back on bids.

Catch-All Campaigns Will Make Sure you Don’t Miss Out

Blanket campaigns or Catch-All campaigns are similar to Win-Back campaigns. I will create these for manual, auto and sponsored brand campaigns. This strategy takes broad high volume keywords and sets a super low bid. At times I find that the bid is to low and I will increase the bid incrementally until I start getting delivery.

What’s a little fuzzy here is what products to serve. I have served all products as a default but seen a better performance by filtering down to the brand’s hero products. Although it’s easy to define a hero product, I like to back my assumption up with data and narrow products down to ones that have strong conversion data and high order volume.

Protect Your Brand on Amazon with Defensive Sponsored Product Campaigns

Defensive campaigns are there to protect competitors from showing on your product listing page. Using Brand Analytics available in the Amazon Advertising console, I can easily find out what products are frequently purchased together and if it happens that one of these products is from a competitor I can take one of my brand’s products that is the same or as close to the same and replace with our own.

For example, if you have a product with four flavors you can show alternate flavors targeting each flavor. Or if it turns out that people often purchase competitors’ chocolate products when purchasing your bars you can decide to show your own chocolate when people are viewing your brand’s bars.

Again these are defensive campaigns so I do allow for a higher ACOS since they help protect the brand by displaying Amazon Sponsored Ads on your listing pages. However, I do look at brand analytics and non-amazon customer purchase data (if available) to determine what products are frequently bought together then test multiple versions scaling back or turning off the lesser performing product.

Capture Lower Funnel Purchasers Leveraging Retargeting with Amazon Display Ads

Retargeting or Sponsored Display Ads are great to capture the bottom of funnel purchasers. Although this works as a retargeting campaign, Amazon will show your Amazon Display Ad to people that viewed your product or similar products to yours. I would have one campaign for each product grouping and start with a $1 bid adjusting bids in the amazon advertising platform respective to performance. (At the time of writing this post you still could not add this campaign type to a portfolio in the Amazon Advertising Console, kinda annoying I know). ACOS performance has varied for this campaign type but given the timing, bottom of funnel targeting, and that they also show for people that have viewed competitor's products I tend to give some ACOS leniency for these as well.

The Importance of a Solid Naming Convention for your Amazon Advertising Strategy

I also want to briefly cover the importance of a consistent naming convention. Since each product grouping will consist of multiple portfolios and it is important to be able to get a quick birds-eye view of performance across a product line when looking at the last 7, 14 and 30 days of performance in the Amazon Advertising Platform. For this reason, I make sure that each grouping has the same format, for example, the following campaigns would be distributed through Expansion, Automatic and Converting portfolios:

MSP — PRODUCT A — EXPANSION

ASP — PRODUCT A — LOOSE MATCH

SB — PRODUCT A — CONVERTING SEARCH TERMS

MSP — PRODUCT A — CONVERTING SEARCH TERMS

Having ‘Product A’ in each campaign name allows me to easily filter down that specific product grouping both in the Amazon Advertising Console and in a search terms report.

A quick ramble: Digital advertising is never linear, consists of multiple touchpoints and many micro-conversions from exploration to purchase even on Amazon which is a lower funnel, direct response, and intent-based platform. Although with Amazon you don’t have the nuances (read pain in the ass) of the walled garden effect (think measuring cross channel Google to Facebook) Amazon is still pretty basic in being able to provide detailed insights into a cross campaign and paid to organic attribution. This makes it hard to see which campaign, ad group or keyword impressions and/or clicks assist in driving conversions. Having a consistent naming and account structure provides filtering ability that allows me to directionally get top-line insight into profitable ACOS and account growth on a product grouping level which is super valuable when working with campaigns that need leeway on ACOS given their respective strategy.

Optimization Methodologies

Managing a large number of campaigns across Amazon Product Ads, Sponsored Brands (Amazon Search Headline Ads) and Amazon Display Ads can be tricky, especially if you have little room in allowing an ACOS inflation which invariably will happen as you let delivery run to find what is working.

There is no way for me to manage this account structure manually so I use a mix of automated and manual approaches to efficiently optimize Amazon Ads across the account on an ongoing basis.

Automated Rules (Sellics)

A quick note on Sellics. It worked pretty well for me but I only used it to manage bids not move keywords (outside of moving keywords with delivery but no conversions to negative) I love the idea of automating the movement of winning keywords but I have found the tool to be limited for a complex account structure that is not operating within the simple and traditional Automatic ➡️ Manual ➡️ Broad ➡️ Phrase ➡️ Exact structure.

Some notes on the settings I use:

• Attribution is set to look at data based on a 5-day attribution window
• Some target ACOS are set slightly higher if we have low sales volume and need to get momentum
• The click threshold is calculated using the product conversion rate. For example, if a product has a 20% conversion rate 10 clicks should produce two conversions. A 1.5x click threshold would allow 25 clicks at 2 conversions without making a bid adjustment.
• Brand and defense campaigns are not connected to any rules since they help protect the brand.

Ok now that I have briefly gone over the settings here are some of my primary automated rules.

Automated Campaign Rules

On an Ad Group level:

• Lower bid by 20% if ACOS is 7.5% greater than the break-even target ACOS at 1.5x times the click threshold.
• Increase bid by 30% if impressions are less than 200 over the last 5 days
• Increase bid by 20% if ACOS is 7.5% lower than the break-even target ACOS at 1.5x times the click threshold.
• If clicks greater than 2x times the threshold and no conversions move search term to negative

Manual Ranking Campaign Rules — Keyword Level

• Lower bid by 30% if ACOS is 10% greater than the break-even target ACOS at 2x times the click threshold.
• Increase bid by 30% if ACOS equals the break-even target ACOS at 2x times the click threshold.

Manual Campaign Rules — Keyword Level

• Lower bid by 30% if ACOS is 5% greater than the break-even target ACOS at 1x times the click threshold.
• Increase bid by 30% if impressions are less than 200 over the last 5 days
• Increase bid by 30% if ACOS is 5% less than the break-even target ACOS at 1x times the click threshold.

Notice that for the Ranking campaigns I built rules with more leeway given their purpose as opposed to the general manual campaign automated rules.

Let’s move on to some manual optimization sweeps I apply to an account on an ongoing basis.

Manual Optimization Sweeps

With manual optimizations, I will start on the account level and drill down as I apply each filter variation starting with the campaign manager.

In the campaign manager first sort campaigns by spend. For campaigns with an ACOS greater than 70% drill down to each ad group and keyword then sort by ACOS and drop bids on all keywords with ACOS > 70%.

Next, I want to drill down to high offenders (I should have caught the glaring ones in the step above) and high performers to see if I can break out and dedicate budget towards some winners.

Campaign Manager Review (This also means reviewing campaign ACOS filtered by product grouping)

Sort by: Campaign ACOS > 70%

Sort by: Campaign ACOS < 30%

Campaign Delivery + Ad Group Spend Distribution (Look at the delivery of each campaign over the last 7 days.) If there are winning ad groups break them into their own campaign.

On high ACOS days look for offenders from the previous day — I treat this with caution since last day performance is not necessarily accurate given the attribution lag. Being too quick to make decisions on last day performance can be reactionary and cause for preemptive adjustments. However, having said that, for new campaigns/keywords you can at times quickly find an aggressive keyword that is set to a high bid causing an uneven distribution of delivery in which case a bid adjustment or removal of the keyword is justified.

I will also want to perform a general keyword review across the account using the Search Terms report.

Keyword Review

To bring down ACOS across the account I want to look at recent offenders:

On a pivot table — Filter clicks greater than 20 with orders < 1 and turn off/move keywords

On a pivot table — If spend > 40 and orders = 1 bring bid down by 50% or turn off/move keywords

Above was primarily focused on sponsored products so I will isolate sponsored brands as well and asses their performance.

Sponsored Brands Review

Sort by spend

Drill down on campaigns with ACOS greater than 70%

Apply bid down and bid up rules

Bid Down Rule

Keyword ACOS > 70%, Spend > $20

Keywords Spend > $20, no conversions

Bid Up Rule

Keyword ACOS < 30% Conversion minimum 2

And finally here are some general search term report analyses I will perform typically done on the product grouping/campaign type level.

Twice Weekly

1. Pivot table filter by automatic campaigns by product
2. Sort descending by highest spend
3. Add as negative if spend > $20 and no conversions or add as negative if ACOS 100% +
4. If relevant but high ACOS make sure keyword is included in MSP with low bid before adding as negative to Auto (Or add to win-back campaigns)
5. Look for repeat words that are not relevant and add as phrase negative
6. Add high impressions 1,000+ no clicks to negative (auto only)
7. Sort ascending by ACOS and move winning ASINS to targeted campaigns and winning keywords to general search terms campaign (cross-check and remove duplicates)

When filtering for duplicates make sure to also:

1. Lower bids on high ACOS keywords/ASINS
2. Increase bids on low ACOS keywords/ASINS if 2+ orders
3. Keyword report — greater than two orders + healthy ACOS increase bids

Conclusion

Although this Amazon Ads build-out can seem overwhelming and complex (which it is) starting with a clean structure will almost organically flow into this Amazon Ads Strategy. So how did this work out in terms of numbers?

After the initial product grouping/portfolio build-out of the Amazon Advertising Console, I was able to operate the account MoM with a 30% ACOS and a 15% Total ACOS (which was healthy for this category). Once I started to aggressively kick in the expansion of this account the account ACOS went up to 50% and a 22% Total ACOS in the first month with a 15% revenue lift. Into the second month, I saw a high 40% account ACOS but a reduction of total ACOS to 18% with a trending 30% revenue lift.

Building out your Amazon Ads account, especially in a competitive landscape is necessary if you want to scale, but naturally, as you do this, you will at first see fluctuation in performance so make sure to have a solid grasp of profitability and set the proper expectations.

Hopefully, you were able to take something away from this. I would love to hear what other Amazon marketers think of this Amazon Advertising Strategy.