Did Changes to Amazon Keywords Affect Product Launches?
From August to October 2018, things started getting strange in the realm of performing product launches based on keywords.
Up until August, you could give me any combination of products with keywords and I could give you a realistic giveaway strategy. I would tell you how many giveaways I would need to sell over a particular time period to send a product to page 1, and I would be successful 10 out of 10 times. If you haven’t read Part 1 of this blog series, I reminisced about the glory days of keyword launches using the giveaway method when ranking a product was much easier.
Could I predict how well something would “stick” on page 1? I’d say while not perfect, I’d guess right quite often. Some people would say that products with giveaway launches would not “stick” because they were being “penalized.” However, this notion is so far from the truth.
What helps your product stay on page 1 ranking is, and always been, your organic conversions. If your product is able to rank on page 1 for a keyword, but no one clicks on your listing nor buys your product, it will most definitely drop off of page 1 faster than you can say “unsuccessful product launch.”
Additionally, your product will lose top ranking if it’s not converting as well as the others on page 1, but that is another topic.
If you need help creating the best possible keyword list to for your listing and product launch, check out The Ultimate Guide to Amazon Keyword Research below:
For this post, I want to show you the data I gathered from doing 133 keyword giveaways from the first week of August 2018 to the first week of October 2018.
I wanted to see what factors were having the most significant effect on the product launches I was attempting. I tracked THOUSANDS of data points and tabulated everything by hand (I told you last time I am a data junkie!).
My data points included the following:
- Age of the listing
- Length of giveaway launch
- Type of URL used
- Number of times the keyword was in phrase form
- If the keyword was in phrase form within the title
- Cerebro IQ score
- Amazon Advertising relevancy score
- Number of exact and broad searches the keyword has
- Number of detected sponsored ads for the keyword
- Discount percentage
- CPR number
For those who do not know, CPR stands for “Cerebro Product Rank;” its the formula that Manny developed to estimate how many products you would need to sell at a heavy discount on a certain search in order to have the best chance to get to page 1 for that search term.
Below are the overall conclusions I reached after reviewing the results of these 133 tests:
Storefront 2-Step URLs
When things were still “normal” In August, I did nine keyword giveaways with the Storefront URLs.
Four of them got to page 1 even though I only averaged 50% of the CPR number. The fifth also got to page 1, but I had made a mistake and accidentally given away 370% of CPR.
The other four keywords averaged position 103, but only 20% of CPR was used. In these cases, I was mainly just trying to get products ranked in the top 306 so that I could run promotions that came from search and buy promotions, which is what the Heatseeker URL did.
In September, I did seven keywords with a Storefront URL. Two of them only managed to reach the bottom of page 1 with a CPR of 120%. The rest of the keywords only reached page 3 with an average CPR of 125%. (See how things changed from August to September?)
In October, I only did one test with a Storefront URL. After a CPR of 216%, it went from being unranked to page 20!
Hidden Keyword (ASIN 2-Step URL)
Started testing in September with 12 keyword campaigns total. Three of the keywords got to page 1 with an average CPR of 60%. The rest got to an average position 50 (around page 3) with an average CPR of 105 %.
Field ASIN 2-Step URL
Started testing at the end of September using four keyword campaigns with this URL type. Two of the keywords ranked on page 1 in the top SIX positions with an average CPR of 120%. Other two were page one, position 18 on average with only a CPR of 58%.
Brand 2-Step URL
Started testing at the end of September with four keyword campaigns using this URL type. Three keywords ranked on page 1 with an average CPR of 93%. The last keyword went from position 50 to position 24 with only a CPR of 55%.
This URL is not a traditional 2-Step URL and is not easy to generate. It is this difficulty that has kept us from including it on our Helium 10 Gems page with our helpful URL generators. However, for years this URL has been my go-to method for keywords that are already ranking.
In August, I performed 57 keyword launches with Heatseeker URLs. The average landing spot was position 12 on page 1 with an average CPR of 70%. Out of the 57 keywords, only 8 did not get ranked on page 1.
In September, I performed 20 keywords launches on this URL type. The average landing spot was position 24 with an average CPR of 118%. Out of the 20 keywords, 5 did not rank on page 1. Like before, you may notice a difference between August and September.
In October, I performed 17 keyword launches with Heatseeker URLs. The average landing spot was position 39 with an average CPR of 59%. In retrospect, I did not do well in estimating how many coupons would be redeemed. However, a Heatseeker URL with an average CPR of 59% would have ranked to my product on page 1 almost every time in March 2018, for example. Out of 17 launches, 8 did not rank on page 1.
Launches by Length
- 7-8 day launches: The average landing spot was position 25 with an average CPR of 84%
- 3-5 day launches: The average landing spot was position 41 with an average CPR of 104%
- 1-2 day launches: The average landing spot was position 24 with an average CPR of 81%
- If the target keyword phrase was in phrase form within the product title, the average landing spot was position 18 despite an average CPR of 50%.
- If the target keyword phrase was NOT in phrase form within the product title, the average landing spot was position 28 with an average CPR of 94%
- If the phrase was in phrase form SOMEWHERE in the listing at least four times, the average landing spot was position 22 with an average CPR of 22%. I ran eight promotions in this small sample size.
- If the phrase was in phrase form 1 or 2 times in the listing, the average landing spot was position 23 with an average CPR of 74%.
- If the phrase was NOT in phrase form ANYWHERE in the listing, average landing spot was position 27 with an average CPR of 96%.
You can clearly see certain trends that I was able to derive from all of these tests
- Something happened from August to September that made things more difficult and less consistent for ranking.
- Storefront URLs would still work on some keywords, but not as well on others. The differences seem to have been impacted by whatever changed on Amazon.
- Using a target keyword in phrase form within the product title tends to give you more bang for your buck as far as the number of giveaways needed to rank on page 1 is concerned. This keywords placement also seems to be affected by having the key phrase repeated more than two times in the listing.
What else did I take away from this test?
- Older listings beyond one or two-months-old were more difficult to rank on page 1.
- There wasn’t much difference in how high you could rank if you took the keywords that had over 1000 competitors, as opposed to the ones that had less than 1000 competitors.
- The average position rank achieved was between 22 and 27.
- The number of sponsored ASIN’s that were detected for any keywords did not carry any significant impact one way or another.
- Ranking on page 1 was not easier with lower discount coupons as opposed to higher.
- The larger discounts (85%+) had a slightly better landing rank than the ones with less of a discount (60-84%).
- If you check the Amazon Advertising relevancy score, there were plenty of examples of products being above 80% score, but could not rank on page 1, but many below 80% score were able to rank on page 1.
- The average number of days it took to reach peak ranking from the middle of the promotion was three days.
The biggest takeaway for me was that my old formula was not bulletproof anymore. I could no longer predict whether or not I could rank a product on page 1 through a promotion with 100% accuracy. Additionally, more units were now needed to rank on page 1, which is closer to CPR than before.
The ironic thing is that throughout the testing period, some sellers were saying that product launches with keywords weren’t working anymore, or that people were getting penalized for doing it. However, that has never been the case.
Four of the top launch companies out there are still doing 800-1000 keyword launches daily. There would be an unprecedented outcry if all of a sudden this form of product launch stopped working all together.
I hope you found this experiment enlightening and that you learned something useful about product launches (and why they still work). In the final part of this blog series, I’ll talk about my most recent tests to see if there are any “new normals” that sellers must get used to.
In all honesty, I’m not sure what the outcome of these tests will be. I have run another 50 or so tests since the above writing this post, and I am currently tabulating the data. Stay tuned my fellow data junkies for the conclusion into this insight experiment on the changing product launch landscape!
Do you have questions about my findings, or have data of your own to share? Let us know done in the comments below!