Amazon PPC Optimization: A Step by Step Guide to Find Converting Keywords

Amazon pay per click campaigns are some of the most cost-effective advertising platforms around. For the Jungle Stix product, we are getting some conversions for as little as a dollar a click. I would easily skip a trip to McDonald’s Dollar Menu if that dollar could be put towards selling a product!

When you create a purchase-focused marketplace like Amazon’s, it’s no surprise that a lot of random customer searches will lead to a sale. The question then becomes, which keywords convert well, and how much should we be willing to pay for them?

In this post, we will dig in to Amazon’s Search Terms report, and find keywords that are profitable for us.

You can find profitable keywords using our new tool, Keyword Scout, which is included with Jungle Scout.

Review of Jungle Stix Pay Per Click Campaigns So Far

You can read Greg’s post about creating Pay Per Click campaigns when he first launched Jungle Stix in December.

The strategy was basically to run two campaigns simultaneously: one campaign would be adding a bunch of relevant keywords that we got from Google’s Keyword Planner, and the other campaign would simply be Amazon’s automated campaign.

  • Keyword Bids: were set at a maximum of $0.75  per click.
  • Match Type: All of the keywords were set to Broad Match. So our ad would appear for searches like  “bamboo sticks” could also
  • Campaign Budget: Our Daily Budget is set at $100 per day.

Here is a look at the Campaign Manager Dashboard, from the day the campaigns started on 12/12/15 to today. You can see the Lifetime reports on the right:


And here is a look at each campaign, starting with the Manual Targeting:


And the Automatic Campaign:


The campaigns have pretty much been on auto pilot since then, and have been profitable.

The Average Cost of Sale is 16% and 17% respectively, which is pretty good, but there is still some room for improvement. Paid Search campaigns are all about optimization, so there are some tweaks that we can make to save money and convert more.

In this post, I will review how we can analyze the Search Term reports to find profitable search terms, and how we can reorganize our campaign structure to reduce our costs.


Step 1: Find the Relevant Data

Compared to Google’s well-polished AdWords platform, Amazon’s paid search platform is like the smart slacker: the data and intelligence is in there, it just takes some coaxing to get it out.

The first step to finding the converting keywords is to pull the Search Term Report. You can find that under the Advertising Reports tab of Seller Central:



Once you have your campaigns set up and running consistently, mark you calendar every 2 months to download this Search Term data. Unlike AdWords, which retains and displays all of your campaign performance history, Amazon only shows you the past 60 days. I guess it’s true, Bezos doesn’t like to share!

However, once you download the data, it renders in a text file, looking like this jumbled mess of mess text:


In this state, it is of little value to us. Let’s make some quick changes to make it more useful!


Step 2: Clean the Data

This step is pretty straightforward. Simply copy the text document, and paste it into your Excel spreadsheet.


On to the next one….


Step 3: Organize the Data

Alright, you’re still here! Now we’re approaching the fun part….the pivot tables! Pivot tables are a feature of Microsoft Excel and Google Drive that allow you to reorganize large data sets to be easier to read and understand. They can be very powerful, and also very easy to create.

From your spreadsheet, you are simply going to click “PivotTable”:


It will automatically select the data that you have pasted, and you can paste it onto a New Worksheet. So without changing anything, you can just click “OK”:

pivot_table_prompYou should be brought to a blank sheet: endless possibilities, but can be confusing where to start.


As you can see in the top right, the columns from the Search Term Report are listed as fields. These are the fields that I select in my Pivot Table:


Note: I choose not to include any calculated averages, like click through rate, average cost per click, Average Cost of Sales. It is better, and more accurate, to calculate these manually. We’ll review how to do this later.

Once you select the fields to add to the report, you will need to organize them. This is how I organized the fields for this pivot table (it’s as easy as drag-and-drop between the fields):


And after you click OK, your Pivot Table will look like this:


Once I am here, I copy the whole sheet, and paste it to another brand new sheet. However, when pasting, make sure to “Paste Special” and “Paste Values”, so we don’t have to deal with the formatting of the Pivot Table.


At this point, let’s add in some additional columns for helpful metrics that we can calculate.

This includes the following:

  • Click Through Rate = Clicks / Impressions. Helpful to know how many customers were interested in your ad and clicked it. If it is low, then it may be an unrelated keyword, or the image/headline isn’t engaging enough to click through.
  • Average Cost Per Click = Total Cost / Clicks. Helpful to know how much, on average, you are paying for a click to your product page. You want to manage your average cost per click, and bids, depending on how much it costs you for each conversion.
  • Conversion Rate = Conversions / Clicks. Helpful to measure the relevancy of the keyword, based on how many people who land on your page actually purchase your product.
  • Cost Per Acquisition = Total Cost / Conversions. Helpful to know whether you are making or losing money for a given pay per click conversion.

Additionally, I like to filter the column headers so that we can easily sort the data depending on which metric we are looking at. This is how you filter the data:


Lastly, you can make it cleaner visually by adding the appropriate number type for each column, whether dollars, percentages, or integers. So now that things are organized, let’s dig in and find see which keywords are converting well for us.


Step 4: Analyze the Data

If you have made it this far, kudos to you! To make it more interesting and engaging to you, and so we’re on the same page, you can download the worksheet that we have put together with the actual Search Term data for 2/19/16-4/18/16.DOWNLOAD THE WORKSHEET HERE!

Here are a few steps that can help us get more mileage from the keywords that convert well, and those that are bringing unqualified traffic that doesn’t convert.

One quick yet important side-note: how much are we willing to pay for a new customer? Here is how we will piece that together.

  • Retail Price: $20
  • Cost of Goods Sold (Including all shipping): $6
  • Amazon Fees: $9
  • Net Margin = $20 – $6 – $9 = $5

Theoretically we could spend up to $5 on marketing and still break even. That is assuming that the selling price stays at the current $20 price point. Good to know! Amazon’s Pay Per Click terminology would calculate that as an Average Cost of Sale à $5 / $20 or 25% ACoS.

So our break even Cost per Acquisition is about $5, or said differently, we can shoot for a maximum ACoS of roughly 25%. We won’t outright pause any keywords that are above these thresholds though, because there are some additional benefits to getting customers, like a boost in organic ranking, an increase in reviews, and a lift in organic conversions as a result of more reviews.


Pay Per Click Optimizations

Alright, now it’s time to start optimizing the campaign!

Step 1: Adjust Bids and Pause Keywords

As I mentioned earlier, our Jungle Stix campaigns were both Broad Match campaigns. So we left the targeting completely up to Amazon’s algorithms for which search queries our ad would appear for. The benefit of this is that we would come across some good keywords that we didn’t expect. The negative could be that we end up paying for a lot of irrelevant clicks that don’t turn into customers.

To filter for these keywords that aren’t working, I like to first filter for the keywords that are costing the most money. Are there some search terms there that are completely out of left field that are irrelevant yet generating clicks and costing ad spend? A quick scan through the list, and I don’t see any.


I would then want to scan over to Column J, Cost Per Acquisition. Are there some keywords that are very expensive? It looks like “marshmallow sticks” has had two conversions, at $18.55 per sale. That is eating into our profit margin like a rabid monkey. No thank you.

So we have a few options here:

  • We can reduce our maximum Cost Per Click bid, which is currently $0.75. At $17 per conversion, we are basically at 3x our Cost Per Acquisition threshold of $5. So if we reduce our max bid by 3x, we would theoretically be getting conversions at $5 and breaking even. That would mean that our maximum bid would be $0.25.
  • We can pause the keyword altogether. We have only generated 2 sales with this keyword over the past 60 days, so we probably wouldn’t see a noticeable dip in overall sales if our ads don’t show for this keyword. And we would not be losing money on each sale.
  • We can keep it running as is. This is the aggressive play for the seller who wants a sale at all costs. In some ways, selling on Amazon is a land-grab of sorts, so there is a rationale to this.

You can choose whichever path is most comfortable for you, but those are the thought processes when looking for the poorly converting keywords.

On the flip side, there are some keywords that are your money makers. These are the keywords that have brought paying customers and profits. So you want to make sure that you are showing prominently for these search queries.

Keeping the same filter, of seeing the search terms that you have spent the most for, I see “wooden skewers” has generated 10 sales, at $1.44 per sale:


So make note of that, and you’d want to increase the max bid on that keyword. The average Cost Per Click is only $0.37, and our max bid is currently at $1.00, so increasing it won’t really do much, but if competition emerges for this keyword, at least you will continue to win the bids.


Here are some other helpful filters to check out

You can also sort by the Conversions column, from highest to lowest. This will help you pull out the converting keywords, and make sure that your bids are high enough.

You will also want to check the converting keywords, and check your Max Cost Per Click bids against your Average Cost Per Click.


The basic concept is that you want to make sure that your average cost per clicks aren’t close to your max bid for any good keywords. If it is, then that means that you may be losing out on impressions due to a low bid, and if the keyword converts, then you may be missing out on sales. Filtering from Largest to Smallest, it looks like our highest average CPC is $0.75, and our bids are at $0.75, so if it is a good keyword for us, then we may want to adjust the bid.

You adjust the bids for each keyword (in manual campaigns) in each Ad Group, like so:



Step 2: Narrow The Keyword Targeting (change from Broad match to exact match)

One adjustment that we can make to reduce our spend is by reducing the number of Broad Match keywords and create a new campaign for Exact Match or Phrase Match keywords that are converting well.

For example, for our Broad Match keyword “bamboo sticks”, we are matching to all of these customer searches.


If we change this to Exact Match or Phrase match, we would have our ads only show for “bamboo sticks” for exact match, or a phrase containing “bamboo sticks” in that order, like “9 inc bamboo sticks” or “bamboo sticks 100”. You will have much more control, and specific targeting with Exact Match, which means that you can separate out the non-converting keywords and save some money.

This would help us reduce our spend, assuming that we follow through on the next step.


Step 3: Create New “Match Type” Campaigns

In order to help organize your pay per click campaign structure, it is helpful to separate your match types into different campaigns. Some people like to separate into different ad groups, but I think you can control your budgets more easily if they are in separate campaigns.

Create a new campaign, which you can call, “Jungle Stix Exact Match” and add all the keywords that you identified as good converting search terms. After you do this, you will have two campaigns, one for broad match and one for exact match. Now you are starting to really tighten up your pay per click campaigns. Just one more step for the fully polished campaigns.


Step 4: Check For Keyword Cannibalization

This next step is a Google Ad Words best practice, I have not yet verified if it makes a difference in Amazon’s platform though. It is having your keywords competing against each other, and increasing the price you pay for a click. If for example, we have “bamboo sticks” in two different campaigns, they are essentially increasing the necessary bid to win an ad auction.

Here’s what you do…

Take all of the Exact Match or Phrase Match keywords from the new campaign that you created. And add them in as “Negative” keywords in your Broad Match campaign. You would do that here:

negative_campaign_Or if you prefer to separate your keywords by Ad Group, you can add Negative Keywords at the Ad Group level:

negative_ad_group_levelNow you will still be getting your broad match keywords for “bamboo sticks” searches,  like “110 bamboo sticks” for example, but if the customer searches for “bamboo sticks” and you have created an Exact Match campaign that includes that search term, then your ad will appear. This is helpful because you now have much more control over the search terms that you appear for, and the price you pay for a click.


In Conclusion

Thank you, friend, you have made it to the end! Hopefully now you have a grasp of how to pull helpful search term data from Amazon, reorganize it, and pull insights from it to save you money and generate more sales. This was a long post, and there may be some details that you have questions about, so please drop them in the comments section so I can clarify anything. And if you have other suggested pay per click tips that you can share, please do! I will cover other PPC related posts in the near future.


96 comments on “Amazon PPC Optimization: A Step by Step Guide to Find Converting Keywords

  1. Hi there,

    Great article, thank you so much!

    I’m an FBA seller exclusively on Amazon Canada, I have one product right now and I’m about three months in. I’m trying to optimize my campaigns without losing out on clicks/sales. Obviously the Canadian market is so much smaller and less active than the US market. Therefore, it takes longer to validate keywords and accurately review campaign metrics. I am running an Auto and Manual Campaign with four different Ad Groups in my Manual, one fore each match type (broad, phrase, exact) and one with a combination of matches.

    Being three months in, my Auto Campaign has a lifetime ACoS of 78% and my Manual Campaign has a lifetime ACoS of 100%. Obviously this is not good and needs to be improved 🙂

    The majority of my PPC spend is going toward a short generic keyword that is also producing most of my sales. I’m worried that, if I reduce my bid or add it as a negative exact keyword to my Auto Campaign, that my sales will decrease and I’ll lose some of the traction I’ve gained.

    I’m still a newbie with this so any feedback I can get from an experienced seller would be invaluable.

    Thank you very much!


  2. Great thanks Dave. One more thing, on Amazon seller central the number of impressions showing for the campaigns doesn’t match when you download the results using the advertising reports. Is this because Amazon doesn’t show on the report all the times the ad was shown but not clicked on?

    1. I’m not sure I follow, Paul. But I’ll see if I can’t go over it.

      The impressions might be for both organic and advertised campaigns on your business reports, then when you download your advertising reports, it’s just for your sponsored ads. That’s why the impressions in business reports would seem to be higher.

      That’s my best guess without understanding more about the situation.

  3. Excellent post. I have a question, on that report you had $211.06 worth of keyword clicks that generated zero sales.
    How do you determine when to add the keyword as negative? The problem i see is that a lot of them may only have a couple of clicks so it’s hard to forecast when it will generate sale in the future. Individually they aren’t costing a lot but collectively you’ve got $211 completely wasted which can make a massive difference to a profitable campaign or not.

    1. Hey Paul,

      Great question. I usually wait for about 20 clicks before I think it’s a dud. And it’s all Pareto principle. The ones that do perform well usually make up for the cruddy ones.

  4. Thanks for sharing the guidelines of converting keywords. Your post is freaking helpful as you have explained all the technique in a step-by-step manner and the way you insert the screenshots of each step is really appreciable.

  5. I do not understand how this indicator is obtained in step 4?

    Conversion Rate = Conversions / Clicks. Helpful to measure the relevancy of the keyword, based on how many people who land on your page actually purchase your product.

  6. Hey Guys,

    I have a question regarding sponsored ads.

    I currently sell jewelry on Amazon and I am not receiving a great result on Amazon sponsored campaigns. Currently we carry over 1000 sku’s and i am not sure how to structure my campaigns properly. I have tested multiple add structures where I have tried to group similar products. Does anyone have experience in marketing a multiple products sku’s on amazon if so what is your take on this.

    My current sales in almost 0.5 products a day from sponsored ads, I am not sure why is this happening because I see my competiton selling almost 20-30 orders per day alone with sponsored ads.

    1. Hey Ian,

      One approach that we like to use it to target the same keywords with different match types to figure out which match types work best, and then optimize the campaigns from there. We find that some keywords perform better with certain match types and this is a good way to find out.

      You still have full control of your budgets for all ad groups and campaigns – so it doesn’t mean you end up spending three times as much so long as you set it up correctly and optimize your campaigns on an ongoing basis.

      Many thanks,

  7. Hello,
    Why should I create new match type campaigns?
    What’s the difference between applying match types in different campaigns and in the same ad group?

    1. Hey Jack,

      It’s mostly personal preference in terms of how you want to structure and organize your campaigns.

      We like to create different ad groups for different match types on the case study products – and see which match types perform best and then optimize from there 🙂


  8. Hello,
    This article is really helpful.
    I have a question that should I use ACoS or CPA to determine my campaign performance?
    Both of them are about spend and revenue.

    1. ACoS is just your CPA translated into a percentage based on the sales price. So I’d look first at CPA to determine how much you can spend, and then turn it into a ACoS so you can quickly review it on your campaigns page.

  9. Hi,

    Thanks lot for deep detailed PPC knowledge.
    A bit confused some data in Amazon Search Term Report.

    Question 1:

    Orders placed within 1-week of a click (7 Day Total Orders(#)). Actually, we download and analyze Search Term Report base on 60 days data. Why this data show us 1-week orders? 7 Day Total Orders means Average 1-week orders for 60 days OR 1-week data for last week OR any other meaning?

    Quesiton 2:

    Other SKU units Ordered within 1-week of click(7 Day Other SKU Units (#)). Means A keyword in Campaign A for product A, finally customer click this keyword but then buy another product B in my Amazon Shop, right?

    Question 3:

    How deal with keyword that 0 Conventions, High CTR and Low Spends. Pause OR Reduce bid?

    Question 4:

    Still no idea for Good convention Keywords! Increase Bid? Any formula?

    I am appreciated it if you give me a clear reply.

    Joshua Zhang

  10. Hi,

    Thanks lot for deep detailed PPC knowledge.
    A bit confused some data in Amazon Search Term Report.

    Question 1:

    Orders placed within 1-week of a click (7 Day Total Orders(#)). Actually, we download and analyze Search Term Report base on 60 days data. Why this data show us 1-week orders? 7 Day Total Orders means Average 1-week orders for 60 days OR 1-week data for last week OR any other meaning?

    Quesiton 2:

    Other SKU units Ordered within 1-week of click(7 Day Other SKU Units (#)). Means A keyword in Campaign A for product A, finally customer click this keyword but then buy another product B in my Amazon Shop, right?

    I am appreciated it if you give me a clear reply.

    Joshua Z

    1. Question 1:

      Orders placed within 1-week of a click (7 Day Total Orders(#)). Actually, we download and analyze Search Term Report base on 60 days data. Why this data show us 1-week orders? 7 Day Total Orders means Average 1-week orders for 60 days OR 1-week data for last week OR any other meaning?

      It means that the sales that you made within 1 week of a customer clicking on your ad. So if you had a customer click on your ad on Monday and they didn’t buy it until Saturday, then it still counts for calculating conversions.

      Quesiton 2:

      Other SKU units Ordered within 1-week of click(7 Day Other SKU Units (#)). Means A keyword in Campaign A for product A, finally customer click this keyword but then buy another product B in my Amazon Shop, right?

      I believe so, yes.

  11. Worth Information, I am working as a Digital Marketer in USA. The tips on keyword cannibalization will be definitely very useful for me at my work for better performance. I even printed it out to use while at work. I had a question to ask, are the fake clicks charged on Amazon PPC?

    1. Rachel,

      From what I understand, Amazon has a pretty good system for avoiding “PPC click” abuse. So if a competitor goes on and starts slamming your ads, you won’t get charged a bunch of times.

  12. Super article very informative, engaging! Quick question: Do you suggest creating an ad for each product i.e. Batman Easter basket, spiderman easter basket etc. versus creating group ads i.e. Easter baskets? Which then means I’d be listing all keywords associated with each product in the group campaign? Any drawbacks with one versus the other?

    Also, you’re stating that for Amazon PPC I should run both Manual & Auto Campaigns for each product and include my keywords from the manual campaign into the negative keyword section of the Auto campaign group.

    Do you have any resources on Amazon Keywords Strategy and Listing Optimization?

    Many Thanks!

  13. Dear both,

    thanks for sharing.

    I am clear about the Keyword research and PPC.And i have a question that usually how many keywords will be used??If i selected 500 keywords what it will happen??


    1. Hey Denny,

      If they’re strong enough keywords, you’ll get clicks and hopefully sales. More often than not, a lot of the keywords you’ll select may not get a click it all. But it’s better to cast a wide net and see what you get versus just putting out a few and hoping for the best.

  14. Hi guys,

    Love this tutorial. I’m on a Chrome book and only have access to Google Spreadsheets. Do you have a step by step on how to do this in Google? I’m not an Excel genius so it doesn’t translate easily.

    Cheers, Alina

    1. Hey Alina,

      Great question! Personally, I use excel, too (it’s way easier to create a table using Excel than it is in docs), but I know a lot of people love Sheets (myself included). During the marketing portion of our Excel series, I’ll put together a simple how to on putting into Google docs.

      In the meantime, you might try organizing the data using the “data” dropdown, then click “sort sheet by column _”. I usually sort by “conversions within 1-week” and highlight all keywords that converted at 10% or higher for products with pricepoints around $25 or less, and 5% for products with price points around $25 or more.

  15. Hi,
    thanks for your detailed information for PPC.
    i have runned a compaign, 3 keywords,setted as broad, the ACOS is 5%.
    should i set a new compaign ,set the 3 keywords as exact?
    really confused now.
    looking forward to hear from you.
    thanks in advance.

    1. Hey Echo,

      That’s an extremely low ACoS which is great.

      Usually, yes, you might want to move your broad keywords that do well to exact match campaigns. Exact keywords are often treated more favorably and get a better cost per click. However, what we like to do is run three separate ad groups for our keywords and bid on the same keywords as broad, phrase and exact match. Different keywords actually perform better using different match types, so if you run all three concurrently you can see which match type brings the best results per keyword, and then refine accordingly.

      Three keywords is also very minimal, so unless you plan on adding further keywords to your campaigns I would actually leave them as broad so that you can spot other keyword opportunities in the search terms report.

      Hope this helps. If you haven’t already done so, I would advise running an automatic campaign for a week or two, (Amazon will decide what keywords to bid on for you). This is the quickest way to get real keyword insights about what terms shoppers are using to find your product.

      Good luck 🙂


  16. I’m diving into Amazon PPC for the first time, and this is really helping me get a grasp of things while I flounder around trying to figure it out. Thanks a lot! 🙂

  17. Quick question: If you put bamboo sticks as negative for broad match while putting bamboo sticks in phrase of exact match, how will I still be showing up for broad matches like “110 bamboo sticks” along with exact match “bamboo sticks”? Wouldn’t “110 bamboo sticks” not show up since I put “bamboo sticks” as negative in broad?

    1. Hi John,

      thanks for dropping by.

      that is strange, you shouldn’t really be showing up for any bamboo sticks search queries, as the broad match negative of “bamboo sticks” should negate your ad from showing.


  18. Great post, thanks for sharing.

    How do I add a “campaign’ column when creating the pivot table? It seems to part of the “row labels” the way you have it set up.



  19. Hello Gen,

    Amazing article One of my friend wanted to get some PPC services and I thought of providing a helping hand for him. I read your article and came to know more about PPC and its effects. Just loved what you shared. Looking forward for more articles. Keep Posting.


  20. Awesome article. BUT I see 1 potential major flaw/problem with negative keywords and wonder if you have solved it?

    Lets say your ppc made sales when a customer typed the term “French bamboo stix” based on the broad keyword “bamboo stix”.

    When you add “French bamboo stix” to an exact match ppc campaign, and then create the negative keyword “French bamboo stix” in your broad campaign, won’t amazon ppc disqualify the exact match from showing ads based on relevancy and no longer show your ad for that term unless you actually have all the words “French” “bamboo”, and “stix” in your listing?..

    So if you don’t have “French” in your listing for example, you would lose out on the exposure from that search term altogether if you limit it to exact match? This would imply that you would have to find a way to add every “exact” search term into your description which may not be practical.

    Would love to learn your thoughts on this! Thanks for the amazing articles.. SOOOO helpful!

    1. Hey Regina – if you have separate campaigns for exact match and broad match then adding a negative keyword to a campaign will only affect that particular campaign.

      You are correct that with exact match keywords you need to include all of the exact variations. But you could do this by mining for useful keywords from a broad match campaign and then moving them over to an exact match.

      Broad match will generally gain more exposure and impressions but you will usually spend a little more because there is much more variance in what keywords can trigger your ad.

      Hope that helps 🙂

      1. In reference to the last paragraph of Step 4:

        What is the point of adding an Exact Match for “bamboo sticks” (or any other keyword for that matter) if we are still keeping the Broad Match campaign for “bamboo sticks”? The Broad Match campaign would already make our ad appear for “bamboo sticks” so I don’t get why we are separating it out into an Exact Match campaign AND still keeping the Broad Match campaign for it. In either case our ad will still be showing for all of the same keywords so what is the use? I thought the point of the Exact Match was to remove all of the useless keywords that the Broad Match is showing us for.

        Also, are Exact Match and Phrase Match the same thing?


        1. Hey Phil,

          As Greg mentions in this recent the webinar, he likes to target the *same* keywords with different match types and separate these by campaign. This is because he finds that some keywords perform much better under certain match types. So by targeting both this way you can see which is working for you. This is just another way to mine for data to find out how to maximize results from your campaigns.

          But you are also correct, you can indeed use exact match keywords to remove the chance of your ads showing for irrelevant searches that can crop up with Broad Match. It just depends on how much money you want to spend and whether you are still trying to capture more information about how to best optimize your campaign.

          Exact and Phrase match are not the same, see this table for a better explanation!

          I would recommend reading some of our newer resources on this topics, here’s a couple to get started:

          Hope this helps,

          1. I see that there was a recommendation to include bamboo sticks into exact/phrase match while putting that as a negative in the broad match. Then it proceeds to state that it will show both exact match ‘bamboo sticks’ and broad ‘110 bamboo sticks’. If the broad match was put in as negative, how would this be possible?

  21. Hi!
    I have sold 13 units without running an advertising campaign. Is it possible to find the keywords that customers used to find my product? Or is it only possible after you run a campaign? Thanks!

    1. Hi Andrea,

      Because your product is being found organically, there is no way to find the exact keywords. However, with some thorough keyword research you can start to get some good indicators. One easy way to find out is to do some searches in Amazon yourself for keywords related to your product, and see how well you are ranking in the Amazon search results.

      Here’s some useful resources on keyword research and optimizing your product listing for Amazon search:

      And here is a blog post about leveraging Amazon PPC to get more data (such as keywords):

      If you do decide to run campaigns, you will be able to see the actual search terms that people have used to find and click on your ads within Amazon in the keyword reports.

      Hope that helps!


  22. Hi,

    Excellent Article

    Just one quick question, I followed step by step of the instructions but i only got the clicks of each keyword from my “automatic” campaign, but from my manual campaign I didn’t get the clicks of each keyword I only got the total clicks from the whole campaign

  23. Hi,

    I am getting a little confused when we are importing the search term reports, do we just take one of them or is there a way to import all of them and combine them in the excel file? I tried manually copy and pasting the text files one at a time into Excel but it seemed because each text file I copied in had its own headers that it skewed the information. I tried a individual report without trying to combine all and it seemed to give a similar looking excel spread as Gregs. So again do you combine all of the search term reports or do you have to do this process one file at a time to make up for all 6 months?

    My second question was when when I tried copying my pivot table and putting it in a new sheet how do you input the additional information like click through rate, average cost per click, etc?



  24. Hello Gen,

    Just a quick question: can I use Average Cost Per Click to see how competitive the exact key word is? For example,the keyword that with 0.9 Average Cost Per Click must be more competitive than the one that with 0.4.

    Thank you.

  25. Hi,

    great article! However when I follow the setup shown in the article, I don’t get the “Campaign Name” column on the right hand side.

    I tried moving the campaign name box in the pivot table under “values”, but it is only showing a count of it, I can’t get it to display the actual name of the campaign.

    Any ideas?

    Thx again for a great article !!


    1. Hey Oliver,

      Try dragging ‘campaign name’ to ‘rows’ instead of values, as well as ‘customer search term’. Then put all of your other data into ‘columns’!

      Then you should get your campaign name and the customer search terms in rows on the right hand side in the sheet.

  26. I’m by no means an expert on PPC but have seen some success, and have campaigns in the single digit to high teen ACoS.
    I go back & forth on the whole ‘cannibalism’ argument. Like others, I spoke with an Amazon Sponsored ad rep, who said ‘yes the same keyword in different campaigns, will mean you are effectively bidding against yourself’, so I did the 3 campaign types and any keyword ‘moved up’ to the next campaign, would be set as negative keyword in a lower campaign.
    However, Brian Johnson at Amazon PPC Troubleshooting, the cannibalism effect is not real, so you are not bidding against yourself. For now I’ve gotten rid of all negative keywords, run all 500-600 keywords in 3 campaigns (broad, phrase & exact) as well as having auto turned on. For now, I set auto & broad at a low bid & daily budget, phrase at 50%-100% higher than both of those, and exact at 50-100% higher than phrase. I also have BID+ turned on only for exact. I figure there is enough of a gap in my bids, that if for some reason cannibalism is true, there would probably be another bid between my 2, so I would not be bidding against myself. I’ve also stopped pausing non-performing keywords / campaigns, but rather keep lowering their bids, so as to still get imps.
    My thinking (may be incorrect) is that if exact can get you on 2-3 spots on page 1-2, phrase does the same for pages 3-4, and broad on page 5-6 etc, you hopefully have the first 5-6 pages covered for those that do go down that far.
    I only ever looked at ACoS as a metric, but will also start to look at CPA now.
    Thanks and look forward to your thoughts on cannibalism.

    1. Hi Errol,

      Thanks for stopping by and sharing your experiences!

      You have an interesting bidding strategy, I have not considered something like that. For example, if you search for “bamboo marshmallow sticks”, I see Jungle Stix get the top sponsored position, and also on the bottom or second page. Is your theory that those ads are coming from different campaigns?

      If so, that is an interesting way to cover more potential searches in a cost-effective manner. We haven’t tried different ads to verify which ad is coming from which campaign, so would love to hear about your results!


      1. Hi Gen,

        I’m curious did you implemented this technique of adding exact/phrase negatives to Broad and Auto campaign. What are the results?

        After reading this article I implemented to my campaigns with exact/phrase negatives to both Broad and Auto campaign. The overall performance of my advertising campaigns was very good but I noticed something very weird about my product’s search rank on the main keywords. It looks like everytime it gains a higher position is suddenly pushed back 7-15 positions for apparently no reason. I’ve had days when my product jumped from page 2 to page 1 (because of good sales number) and then back to page 2 or even 3. In the last days it was stabilised on the down side of page 2 and then slowly moving up. I thought that maybe this madness is over and now that is slowly moving up will remain there. But not, once it got to position 2 or 3 (in 2-3 days) on the page 2, it suddenly dropped back to last position. And it was a good sales day.

        I don’t know but it looks like is being penalised for something. And I found some info stating that when you set negatives for some keyword, amazon’s search engine will consider that is not longer relevant for your product so you lose organic ranking for that keyword. And if that is true, it would be the worst thing you could do for your best keywords.

        Now I’m getting all the negatives out and see what will happen next. Do you have any inputs on this?


        1. Hey Adrian,

          thanks for the update, very interesting and observant. I didn’t know that Amazon penalizes negatives in that way. I actually have not yet made the change to this account, however, we are working to put together a webinar on PPC which may help you with more insight! we’ll keep you posted on this.


    1. Hey Ben, good catch, thanks and corrected….it was correct in the spreadsheets and screenshots, typo in the post text though.

  27. Another question: most of my sales seem to come from one-off search term searches, meaning I get a lot of individual sales from terms that were only searched for once so I can’t really say that other people will use that term again. Therefore I can’t really draw conclusions from it nor can I eliminate other words because, who knows? Those might be the terms that convert next week. Am I just stuck with mediocre results? Any way I can improve results?

    1. Are you running broad match, phrase, or exact match campaigns? I would keep running it if it sounds like some of them are converting. The way to minimize your “mediocre” results is to comb through your Search Term Results and flag any irrelevant matches that you are paying for…this will reduce your overall spend so that you are devoting your budget only to relevant keywords with a high likelihood of converting.

      1. I ran a broad campaign for awhile to get a list of keywords to determine their relative success. From there, I ran 3 campaigns – a Phrase Match of Keywords that had already converted to at least one sale, an Exact Match of Keywords that converted to at least one sale (same words as the first campaign so I definitely can see some room for improvement) and Exact Match of words that did not convert – the thinking is that just because they haven’t converted yet doesn’t mean they won’t convert. And unfortunately, all three campaigns perform similarly which is counter-intuitive. Those words that have converted in the past don’t necessarily convert again and those that didn’t convert seem to have just as good a chance of converting on any given week. But only a handful of words are driving multiple sales. It makes it difficult to add budget or remove budget since there doesn’t appear to be any rhyme or reason.

  28. Great post with plenty of detail – One part I didn’t understand was Part 3. You say not to add the ACoS or other items that require a calculation, because it’s more accurate
    to do the calculation yourself but then seemed to gloss over adding those items back in without really showing how to do the calculation manually. Can you clarify?

    1. Hey Russ,

      Sorry for the confusion….

      Because the ACoS is by definition an average, if we average those figures it wouldn’t be a true average (you can read more about that here:

      So this would apply to the Click Through Rate, Average Cost of Sale, Conversion rate, and Average Cost Per Click. You can see the formulas in the post above, or you can see how I calculated it in the Excel download as well. Once you make the formula, for example divide the “Cost” column by the “Clicks” column to get the cost per click, you can automatically fill in the whole column by dragging the bottom right corner of the excel box all the way down.

      Does that make sense? Sometimes it may just be easier to play with the data to see it. let me know if you have any remaining questions.


  29. Hi Greg,

    Great post. I was wondering if you ever tried to take it 1 step further and use your “exact match keyword list” on other platforms, such as Adwords, BingAds or FbAds?

    1. Hey Dan,

      We have not run ad campaigns on other platforms outside of Amazon. However, I have seen some Amazon dynamic remarketing that has shown me ads for Jungle Stix on other sites. So maybe that counts to help drive traffic from outside of Amazon to our page.


  30. when i go to the link i got for yesterdays webinar it says sorry the webinar link has expired, so where can i see the replay?

  31. On Keyword Cannibalization: I run on a continuous basis both Manual and Automatic Campaigns. Do I add keywords that I put in Manual campaigns to Automatic Campaign Negative Keywords?


    1. Yes, you are correct: add keywords from your manual campaigns as a negative keyword in your automatic campaign, that way you are not bidding on the same keywords twice.

        1. Hey Julian,

          You could do that, but bear in mind that you are targeting the keywords using broad match in your manual campaign, therefore there may still be some overlap in what you are bidding on in your manual and automatic campaigns.

          To prevent this, you could add your negative keywords in the auto campaign as phrase match, which should help to prevent a few more of these broad variants being used in the Auto campaign.


  32. Thanks guys. You share the most insightful, actionable content in the field. Oops, I have doing ‘keyword cannibalization’ – Thanks for misidentifying the mind boggling Search Term download. Very timely post for where I am at.
    Love and Gratitude

    1. Hey Sharon,

      Hopefully these tips on keyword cannibalization will save you some money and increase your conversions!


  33. this was sooooo freaking helpful. everyone keeps talking about how we need to use Amazon sponsored ads effectively but no one has layed it out and explained in a step by step manner. thank you so much!

    1. Hey Jaya,

      Thanks for the note, glad to hear that it is helpful! What other parts of Amazon sponsored ads do you have questions or confusion about? We can try to address those in future posts.


  34. Hi Ken/Greg,

    Great post useful information. Would it be possible to record a video showing the keyword search process please? What do you use to get keywords merchant words google keyword tool? How did you decide which ones to go after? I.e is it purely the high search volume words, or do you included all to start with? That process would be useful.


    1. Hi Jo,

      We have done a few videos and posts on keyword research.

      Here is a written post and video where we do research for the Jungle Stix listing:

      Here is a post on PPC best practices and keyword research:

      And here is a post on 3 Keyword Research tools:

      This should give you a solid foundation to find some good keywords!


  35. Thank you so very much for this! It is long but it is nicely broken into small parts which makes it so much easier to follow. I feel more confident about PPC now, especially when it comes to analyzing data and the different types of campaigns.
    I just have a question, how do I factor the sales generated with promos because these sales will reflect in my ppc?

    1. Hi Maribell,

      Glad to hear that the article helped you!

      If I understand your question correctly, your promotions would be handled in the Promotions tab of “Advertising”, whereas your PPC would be handled in the “Campaign Manager” section, see here:

      Does that address your question?


  36. Hi guys

    Thanks for the great information as always I am getting stuck in step three of organizing the table for example Cost Per Acquisition = Total Cost / Conversions. I don’t see total cost in my search term report anywhere. Is total cost my cost per unit to amazon warehouse or is it the same as total spend column?

    Also I see conversions column in the search term report but not as an option on the pivot table. How are you grabbing the data for the conversions column and coinciding it with the data in the pivot table.

    Thanks for again for the help, you guys do great work.


    1. Hi Greg,

      Sorry for any confusion, the “Total Cost” is shown as “Total Spend” in the Search Term Report. You are correct that it doesn’t account for the COGS or shipping, just the advertising costs. If you want to add in those figures to get a true CPA, you may just want to create a new column and add a constant figure to the “Total Spend” column.

      For conversions, I used “Orders placed within 1 week of a click” from the search term report.

      I hope that answers your question, let me know if there’s anything else!


  37. Can’t I just use negative key words to get rid of wasteful terms? For example the term “milk powder” works pretty well but a customer search term of “goat milk powder” is a waste of $. Can’t I just put “goat milk powder” as a negative keyword in my Broad campaign to eliminate this wasted $. Thus I want to keep my broad terms and just subtract out the ones that don’t pay. Does this work as well as your exact match suggestions?
    Thanks for your tutorials!!!

    1. HI Andy,

      Yes, in your example you could just add “goat milk powder” as a negative and keep the broad match “milk powder”. However, what if someone searches for “baby milk powder” or “milk powder container”? Your ad could show for both of those, and I’m assuming that they are both irrelevant to your product.

      Just creating an exact or phrase match campaign allows for greater control of what your ad shows for, and therefore you are able to save money and increase your click through rate and conversion rate as your ad is presumably much more relevant. I hope that makes sense! Let me know if you have any questions.


  38. Hi John,

    Though it may be a bit more manually work, you can filter by Campaign, and then keep or delete whichever campaigns or products you wish to look in to further. I think that may be the best option.

    Hope this helps!


  39. Great article thanks. How can I get a search term report for individual products? I have a few products and the search term report pulls all product campaigns making the pivot table not possible.

  40. Thanks for this in depth article on PPC. The strategies are solid and well thought out. I have a question about Step 4: Check For Keyword Cannibalization. When I try to implement this strategy of creating a new campaign (phrase) with the “winners” from the auto campaign, it seems like phrase and auto campaigns gets less conversions and higher ACoS. Any explanation you can think of for this?

    1. Hi Jamal,

      Thanks for reading the article.

      About your question, did you add all of the “winners” in a separate Phrase Match campaign, and then also add them as NEGATIVE Phrase Match in your auto campaign? If you didn’t do that, that may be the cause of the increase in costs, as you are essentially running two campaigns that are competing for the same keywords. If you add the negative phrase match, then you are only competing once for the keyword, and your cost per click will likely decrease, as will your cost per acquisition.

      Hope this helps, let me know if you already added the winners as negatives.


      1. Thanks for the reply Gen! Actually, your article is spot on and I simply did not wait long enough for the data (sales) to show. I did add the “winners” and make them negative in the auto campaign and just as you stated, the acos has gone down on the new phrase campaign. I am still running the auto and getting sales there as well so I guess the next step is wait a while and continue to pull keywords out of the auto and make them their own phrase campaign.

        Any idea/rule of thumb on when you personally go “phrase” vice “exact”? Or is it just a feeling out process on what the data reveals to you?

        Thanks again!


  41. Hi Greg,

    Excellent post. I think I saw Amazon also offer a pivot report.
    How do you group by keywords in Excel like you did in step 2?

    1. Hi Ori,

      Here’s a screenshot of how the PivotTable is set up so you can see.

      Basically, just add “Keyword” and “Customer Search Term” to the “Row Labels” field in the Pivot Table and you can see all of the actual customer search terms for the Broad Match Keyword. Does that make sense?


Leave a Reply

Your email address will not be published. Required fields are marked *