“How does Amazon determine product rankings with its search algorithm?” you ask.
That’s a great question!
Just like any business with a “top-secret” formula or recipe, this very “special sauce” is one of Amazon’s top and best-kept secrets. Yes, some may proclaim they have tested this or that, but when it boils right down to it… only Amazon knows for sure what is baked into their algorithm. They are unwilling to share it with others, but who can really blame them?
Did you know Amazon has even topped Google as the number one shopping search engine? Yep. That’s reason enough to want to keep that recipe top-secret, which is probably part of the reason why the algorithm is constantly being tweaked and updated by Amazon.
But here’s one of the tastiest bits of knowledge… It’s called A9 Product Search. We definitely recommend you check this link out (It’s a very short read):
Here’s what we’d like to point out as most relevant:
One of A9’s tenets is that relevance is in the eye of the customer and we strive to get the best results for our users. Once we determine which items are good matches to the customer’s query, our ranking algorithms score them to present the most relevant results to the user.
Our ranking algorithms automatically learn to combine multiple relevance features. Our catalog’s structured data provides us with many such relevance features and we learn from past search patterns and adapt to what is important to our customers.
We strive for continuous improvement of our ranking algorithms. We continuously evaluate them using human judgments, programmatic analysis, key business metrics and performance metrics.
Wow. Just, wow. Who thought data mining and programming could be so interesting? This algorithm is some pretty futuristic stuff. This thing takes us one step closer to obtaining true A.I. within our lifetime.
This tells us two simple things:
- Relevant results are pulled from Amazon’s catalog of product listings.
- Results are scored and placed into order of what is most relevant to the user.
There have been many questions regarding whether or not factors, such as social media & external links are factored into the algorithm and, as far as we know, Amazon is only concerned with internal factors. But nobody, outside of Amazon, really knows for sure.
So what are the internal factors?
Based on what we learned from the A9 article, (specifically, the last sentence) Amazon’s internal factors include:
- Human judgments: Great to know, but not very useful to us since there are no specific guidelines given for how “judgments” are determined.
- Programmatic analysis: This is also good to know, but without the parameters and details of the analysis we are again left in the dark.
- Key business metrics: Amazon is a business and businesses are driven by the bottom line. Meaning everything Amazon does is with one goal in mind. That is to maximize profits and revenue. But what does Amazon use to develop, track, and measure their key business metrics?
- Performance metrics: They finally threw us a bone. We do have some information about performance metrics within Seller Central.
“We continuously evaluate them using human judgments, programmatic analysis, key business metrics and performance metrics.”
The next time you input keywords into Amazon’s search engine and hit enter, take a look at the right hand side. There is a “Sort by …” section visible and it defaults to “Sort by Relevance.” Sure, you can change the sorting order to other options but most customers don’t even realize or care that it’s available.
Here are some of the things we think the algorithm takes into consideration:
The higher your sales rank the better your relevancy score when customers search for products they want to purchase. This is like the question “what came first, the chicken or the egg?” More Amazon sales improve your overall sales rank and a higher sales rank improves your organic sales.
Amazon has the sales history for every product in its catalog. I can see multiple ways to use this data, so it’s very possible sales history somehow plays a part in the algorithm as well.
CTR (Click Thru Rate)
What happens after a customer puts in a search term for the product they are thinking about purchasing? Products are scored by the algorithm and listed in ranking order of relevancy. Once the customer sees the list Amazon tracks what they click on next. This is CTR in a nutshell and a vital piece of the algorithm.
After a customer clicks on the link to your product listing is a purchase made? If a purchase is made it improves your conversion rate. Likewise, if a purchase is not made your conversion score goes down. Other factors we associate with Conversion rate are Time on Page, Bounce, and Exit. Getting customers to your product’s listing page is only half the battle. You need to strive for a high Conversion rate once they view your product.
What do you do when you shop for something? Putting compulsive buys aside, you most likely do a little research first. Amazon is no different and customer reviews for your product are very important in swaying a potential buyer either positively or negatively. Amazon tracks both the number and quality of customer reviews. It’s possible the reviewers themselves are even scored and weighted accordingly to their review history and quality of reviews.
This goes hand-in-hand with customer reviews and the research phase of potential buyers doing their due diligence. Answering customer questions is just good customer relations. We don’t have any hard evidence this is tracked or part of the algorithm, but why neglect it when it’s so easy to do? It shows you care about your customers and you’re not some robot salesman. In the world of online anonymity, answering questions gives your company a human connection with your customers.
In today’s consumer culture it is just part of human nature to look for the best deals. Customers hate to feel like they were sucker punched and taken for a ride by a seller who is price gouging. You don’t want to get into a pricing war with competing products, but the algorithm does use price when scoring and ranking relevancy.
I’m not a fan of keyword stuffing. It’s not really necessary. The title’s keywords are combined with your other keywords used for optimization and product searches. There is no need for duplication. Whenever I come across a lengthy product title with too many keywords strung together, making it almost unreadable, I move on. While it is important to have good keywords in your product title, it is questionable how much mileage you will obtain by going overboard. It is best to have an easy to read title that is both professional and quickly describes your product.
Since Amazon is “online” shopping it is important to have a clear, high definition main image of your product. If the main image is not clear and does not provide sufficient detail customers will move on and your conversion rate will plummet. You didn’t hear this from me, but supposedly the “hover to zoom” feature increases your conversion rate! In order to take advantage of this feature make sure you have at least one clear, high definition image that is 1000×1000 pixels.
There’s not any evidence to support secondary or lifestyle images are part of the algorithm. However, they don’t hurt your cause either. Do whatever you can to enhance the customer purchasing experience with your listings.
If you are selling a product that has multiple colors, multiple sizes, or something similar it is best to create just one listing with multiple variations. This is done through Parent-Child variations within Amazon. The benefit to doing this is combining many of the aspects already mentioned. The most evident is combining all of the customer reviews into one centralized listing. This allows for all of those reviews to obtain a much higher collective total instead of smaller quantities spread out over multiple listings.
In truth, this could be a topic for a separate article and be just as long, if not longer. Amazon provides each product with five separate 50-character search term fields. Don’t worry about commas and don’t duplicate keywords from your title in these fields. Also, drill down to the lowest level of category for your product. There is a handy little template tool within Amazon’s inventory upload section that can help you with this.
The main point to remember is to provide as much clean data for your product listing as Amazon asks for. For example, include Product Specifications. The devil is in the details folks. The more details you can provide Amazon the better.
What do I mean by clean? Scrub your data for inconsistencies and eliminate misinformation. If your title, main image, product description, and bullet points contain conflicting information you are in for a world of hurt. Even if you convert customers into purchasing, it is very likely many will become unhappy when what they received didn’t match what they read or saw pictured for the product.
Be sure to include meaningful keywords throughout the listing. Again, stay away from unintelligible keyword stuffing. Write for a human being and include all of the feature rich content your product deserves.
Speaking of a world of hurt, negative feedback is a whole separate beast when it comes to the algorithm. It is believed Amazon tracks all negative feedback and weighs it equally the same… equally bad. In regards to algorithm relevancy, negative feedback kills your score and ranking. Remember what I said earlier about keeping good customer relations? The potential for negative feedback is why going the extra mile for your customers is always a good thing.
If you do a lot of searches on Amazon you will notice that those products with easily recognizable brand names in the title seem to consistently rank higher than those without. A lot goes into marketing and establishing a brand name. Just ask Nike, Coca-Cola, Xerox, and even Amazon! If you can include a brand name in your title then do it. This helps both new and repeating customers more easily search for your product.
Amazon tracks purchase orders processing from start to finish. With the number of Amazon Prime members growing at exponential rates we strongly recommend your products ship FBA (Fulfilled by Amazon) to the customer. This eliminates a lot of the processing slowness associated with FBM (Fulfilled by Merchant) orders.
An added benefit of utilizing FBA services is if something goes awry with the shipping Amazon will take responsibility for it and won’t ding your score. Amazon is really pushing Frustration Free Packaging. This tells us packaging matters to both Amazon and customers, so it might be a part of the algorithm.
Within Amazon’s “Seller Center” there is an Account Health section. You should be very familiar with all aspects of this section. If you are not, then place this at the top of your to do list. Order Defect Rate, Cancellation Rate, and Late Shipment Rate are all part of the algorithm. Amazon gives you the targets for each of these criteria, so just make sure you stay within those target ranges. Again, negatives are bad, very bad, so stay in the green for Account Health.
There are various subsections that are viewable and are tracked either short term or long term. Some are tracked for 7, 30, and 90 days. Of course Amazon has your full account history starting from day one, when you opened the account, so only Amazon knows how long the algorithm’s look back period is for various internal criteria.
If you run out of stock you lose potential sales. Customers and Amazon both hate it when a listing is made unavailable due to being out of stock. In short, don’t ever let this happen! Some believe re-ordering above a set number of units can help your algorithm score. We recommend running a lean inventory without stocking out. You should aim to turn over your complete stock for a product listing every 30 days. This will help with long-term storage fees and other issues Amazon is bringing to 3rd party sellers attentions.
This list is probably smaller in number than what you might have seen in the past. Each of the above numbered subjects can easily be a topic for discussion all its own. However, we tried to give you a quick overview that hopefully you’ll find useful. As with any such list trying to tackle the important question regarding what goes into Amazon’s algorithm your mileage may vary.
Remember, Amazon is always evaluating and updating the algorithm to improve customer relevancy.