Until now the majority of my blog posts have been actionable marketing information about Amazon, how to start an amazon business, how to set up a PPC campaign and even the full on-page optimisation guide that went live last month. But I’ve never gone into extreme depth on the Amazon A9 Algorithm and how to optimise perfectly for that.
Today I’m going to dive into this topic in a little more depth for the more advanced Amazon sellers & anyone looking to understand more about what Amazon’s search engine actually is, how it works and what we can do about it. Remember to email me if you have any questions & share!
Table Of Contents
- 1 What Is The Amazon A9 Algorithm?
- 2 Amazon A9 Vs Google
- 3 Where Is The A9 Algorithm Is Going?
- 4 Updates That Could Be Added To A9
- 5 Conclusion
What Is The Amazon A9 Algorithm?
Amazon A9 is the name for Amazon’s search engine. The division that builds, does maintenance and makes improvements to the search engine is called the A9 division. Officially according to wiki they are a different company that provides “search engine and search advertising technology.” it’s called A9 as a play on words of “algorithms” as there are 9 letters after the letter A.
Whatever you make of the name, the software development behind the company is sharp. With a highly experienced team A9 is definitely where Amazon are going to stay with their algo needs for the near future.
Deciding what Amazon A9 is and determining how sellers & vendors can benefit from it are 2 very different things. In this post we are going to outline where Amazon A9 is in relation to Google, where it is likely to go and most importantly exactly what you need to do to help you sell more on Amazon.
Amazon A9 Vs Google
Amazon is a long way away from where Google is today. Looking at the relevancy of each product search we can clearly see the disparity. The issue is Amazon don’t really care that much. Their search engine is based on products. This means that the products that are shown are done so based on a relevancy + popularity metrics, very similar to what Google USED to run on. Anyone remember PageRank?
Earlier this month I did a guest post on how Amazon looks exactly like Google did in 2010 (pre Panda & Penguin updates.) – In this I went into detail on why Amazon’s A9 algorithm looks exactly like Google did a few years ago. – So what does this have to do with selling more stuff on Amazon?
Where Is The A9 Algorithm Is Going?
In my opinion there are 2 core options Amazon can take and honestly I do not yet know which one they will decide, only they do.
With Google it’s usually very clear what they are going to do (usually the opposite of what they say), but with Amazon’s A9 it’s a bit more complicated, but for a number of reasons it is only going to be 1 of 2 clear paths.
1.) “It’s Good Enough For Now”
The investment into A9 drys up. Although the investment into improving the search engine will never officially “stop”, the first core option of Amazon is that they look into investing into other methods. With Amazon purchasing whole foods last year their primary goal might be on tackling the local market. Meaning a search engine isn’t all that important. They are also looking to AI solutions which even Google have officially stated won’t be in their algorithm for another decade or so.
How long will “now” be?
Why they would say this? Why would they leave it with “good enough for now” and how long would that be?
All super important questions from Amazon’s point of view, but unlike Google, Amazon doesn’t make 95% of it’s profit through ads which are as a result of the relevancy of the search engine. Amazon needs to focus on a number of technical business elements even things from how to light warehouses is something they are looking to improve to why people buy one product with another. Those investments are going to take priority as they are generally seen to have a better ROI and hence get priority.
What I can personally see happening if Amazon were to go down the “good enough for now” route, is they wait for better tech solutions around search and simply buy the company that created it. This is very similar to what they did when A9 purchased Snaptell back in 2009.) It makes sense and it seems to be the Amazon way nowadays too!
2.) “Massive Improvements & Investment Into A9”
Of course the other core option is to massive invest into the technology & A9. This would solve a number of Amazon’s issues related to low quality listings & even low quality sellers from abroad using dropshipping techniques leading to a number of VAT & Tax issues that Amazon are currently dealing with. This is actually one of the reasons UK sellers (including myself) are starting to become frustrated with the Amazon platform. But of course Amazon is too big to fail now.
Obviously the investment would be high. But the results could be incredibly valuable to Amazon. Leading to (of course) a higher average customer sale. This is a common goal to Amazon, increase the amount of money people spend at our website…. That’s all it boils down to.
The artificial intelligence; there’s a lot of debate at the moment around AI and how that’s going to be adapted into Amazon. (Although Google has officially stated they use no AI in the algorithm and they seem to do fine…) Amazon have made claims and stated they would like to integrate more and more AI into the platform to improve the overall efficiency of the website. This of course makes heaps of sense until you see the astronomical cost this has. But Amazon is Amazon for a reason.
There are a few core areas that optimisation experts know Amazon could improve with simply algorithm changes and updates.
Updates That Could Be Added To A9
Panda – Thin/Duplicate content
The first update that comes to mind is the first massive Google update called Panda back in 2010-11. This affected websites that had thin or “poor” content according to Google. This was a massive update that ended up hitting close to 12% of total webpages (can you imagine the sheer size of that?)
The winners from this update were sites that wrote content specifically for users. Longer form posts (such as this one) created to try to educate or advice users. The losers were the content farms that pulled duplicate content from other websites, re-wrote it using spin-tax and essentially “didn’t add anything useful to the internet.” – I made that quote up but I’m sure it was stated by some Google spokesperson back then.
AMAZON NEEDS THIS UPDATE!
There are so many listings that are thrown together, zero regard for conversions, zero regard for semantic search and essentially will just lower the price and get as many sales through discounting a product to make a quick buck. Essentially – A thin or low quality product listing. Amazon could learn a lot from the Google panda update as, as much as this did kill a lot of people’s earnings, it did indeed clean up the web as we know it.
Penguin – Low quality sales & “spam”
The second most obvious update would be another one from the ranks of Google (although the target metric would be changed.)
Penguin was another update that Google implemented in April 2012. The date is stuck in some SEO’s heads as D Day, as this single day wiped out people’s earnings entirely. I personally knew SEO’s earning 5 figures a day from the search engines, every cent of which was gone the day after. The single biggest culling of spam (probably) of all time!
Google targeted a very single metric: BACKLINKS. A backlink is a link that points from one website to another. Google implemented backlinks as a huge part of their original pagerank algorithm. This is in fact what differentiated Google against their competitors (back when they had some). Whilst I’m over-simplifying this essentially; websites that had very spammy link profiles (that had obviously been used to “game the system”) were penalised and their organic rankings decreased rapidly. Of course this led to a massive reduction in traffic and hence earnings.
Amazon Could Do With This To.
Although not as essential as the Google panda style update they should be implementing asap. A penguin style update wouldn’t target LINKS, instead Amazon would target SALES. How did these sales come in? Were they artificially created? Did they use massive discounts? AKA: Are they trying to game the system?
If the answer to these questions is yes then their should be a flagging metric that they use penalise product listings.
I have started to see a very small shift towards this as of late. Although nothing is official from Amazon yet, things like using discounted product sales to spike best seller rank (and hence organic rankings) is starting to become a lot less effective than it used to be. This could be a move towards this style of rankings.
Although this is not based on any Google update of the past, reviews are a huge issue inside of Amazon. They are used for social proof, meaning if reviews are “faked” then people will buy products that aren’t that good, increasing the refund rate, all bad for Amazon’s top and bottom line.
Amazon did recently purge a number of reviews. Sellers lost 10-30% of their reviews that were considered low quality or faked. If this trend continues, instead of doing this on a monthly or yearly one time basis, they may start to integrate smarter AI to draw the lines for them. It is only a matter of time in my opinion.
Having worked with algorithms for over a decade now it’s pretty clear to me that Amazon does not only look like the Google of yesteryear but it really IS IT. The relevance metrics, the Best Seller Rank metric, sales counting as binary 1 or 0 instead of based on quality and a number of other metrics & points that just align so perfectly to what Google used to look like. This in itself excites me as based on the updates that could be implemented into the Amazon A9 algorithm and what that would result in (more complexity in ranking) it would benefit my clients and my personal Amazon businesses drastically. If you have any questions or would like to view our process, please get in touch.