SEO ranking factors for Amazon [Podcast]

Moz interviews Downstream CEO about the Amazon advertising ecosystem, including SEO ranking factors for Amazon.com.

Salim Hamed

Brian Childs of Moz and our CEO Connor Folley sat down to talk about SEO ranking factors for Amazon.com. The interview below is a transcript of the podcast episode over on MozPod.


Brian Childs: Hello, and thank you for joining me for another episode of Mozpod - the podcast for SEO, brought to you by Moz Academy. My name is Brian Childs and I will be your host. today on the pod we are speaking with Connor Folley about SEO for Amazon.com. Connor is the founder of Downstream. Downstream provides catalog management, reporting, ad placement and bunch of other solutions for retailers wanting to maximize their potential on Amazon. you can read up on them at downstreamimpact.com and I will put that link in the show notes for you. I really wanted to have Connor on the podcast due to his insight into SEO differs between Amazon and more classic search engines. And the answer is a lot. There is a lot of differences and it is interesting, and I hope you guys enjoy this conversation with Connor.

So, Connor’s team has done a bunch of analysis into what we probably call ranking factors for Amazon, and his experience while working with Amazon gives him insight into how the process for ranking gets developed. By the end of the class you can drop the term A9 on your friends with confidence and be hipped to the inner workings of Amazon algorithm development. When Connor first mentioned A9 when we were recording, I thought he was talking about the size of a letter for a printer. So, we are all on the same boat here. Turns out A9 is where they are building and optimizing the engine we all search when looking for products on the Amazon platform. I think something else worth listening for is the way that people behave on Amazon and what that means for the funnel. I tend to think of Amazon pages as being down they’re at the bottom of the funnel only, and Connor really opened my eyes as to how wrong this assumption is.

Hey Connor, thanks for joining Mozpod.

Connor Folley: Hey, thanks for having me!

So, Connor Folley used to work at Amazon. Maybe you should tell us a little bit about your background, so you can fill us in on sort of where your perspective is coming from.

Happy to do that. I have been in brands and digital marketing since I got out of college either on the client or the agency side. But the most relevant to our conversation is former Amazon marketing leader left several years ago to pursue some opportunities to offer my expertise to vendors and sellers, and amplified their traffic and sales on the platform. But would like to at the outset be clear that nothing that I am sharing today is proprietary or inside information based upon my former Amazon experience. All of this is observational and things that I have gleamed while working outside of Amazon. My current role is co-founder of Downstream offering expertise to vendors and sellers who are really looking to ignite their traffic and sales on the platform.

It’s interesting to see the development of this ecosystem that Downstream amongst others is contributing to. This discussion today will be really interesting for the listeners. We can really focus in on what are the differences between Amazon and normal Google search, but maybe kind of frame up first about how big is this. A lotof people think search and they think google or Bing, but tell us more about search with Amazon and what’s going up with that.

Yeah, it’s huge. It’s a sea change in the way the product research is done online. More product research is now done on Amazon than every search engine combined, and they really could not be more different as search platforms. First of, Amazon is retail platform fundamentally. now it’s becoming a branding platform because so much product research is done there. Second, the customer search behavior is quite different on Amazon than it is on Google, and I can elaborate on that. Also, the way that keywords are kind of allocated across the head and the tail between Google and Amazon, Amazon just has a very heavy concentration of volume in the head, and they are SEO optimized very differently from one another. 

So, when you say that there is a lot more product search, what do you mean by that? How would a non-technical person look at that? What is the difference between a product search which is what they do on Google?

Good question! I can frame that up with few examples. When customers go to Google, they are fundamentally going to Google and asking questions they are seeking answers to, vs Amazon customers are typically coming and querying on product attributes. To illustrate this, you look at the top searches on Google vs Amazon related to dogs. The top 5 keywords on Google are going to be Dog, Dogs, Rottweiler, Dog breeds and Golden Retriever. On amazon it is much more product focused -  Dog Bed, Costume, Dog food, dog toys, dog crating and on down the line just one product after another and getting more and more specific. Customers are just coming in, researching products and buying products. So, for those brands that think that they can just take their Google search programs and just copy and paste them over to Amazon, it’s really going to be poor use of resource, because they are fundamentally different platforms. 

So, would you say then that do people start with an Amazon Search? because it sounds like if I am asking questions about like, you know the maybe the dog example is little weird, because I wouldn’t be like what should I get my dog for Christmas, but maybe that what would be what it is. It’s kind of like Dog ideas or ideas for dog gifts or something vs when you are choosing a specific one and researching one. Is that all happening on Amazon? are people asking questions and kind of going through the buyer journey there or is that all happening on Google and they go to some decision point and then go to Amazon.

That’s a fantastic question. I think that much of that kind of buying guide top of the funnel kind of shopping and search behavior. Maybe you would see somebody coming to Google and using a buying guide or a top 10 lists or looking for some sort of superlatives where they can determine what kind of product to get. A lot of that is satisfied through Amazon’s merchandising widgets. So whereas you might go to Google and ask what are the best Dogfoods that are grain free, on Amazon you are just going to search for grain free dog food, you are going to use the left-nav dynamic filters to go to 4-stars, look for particular brands, start reading customer reviews, and you are going to be able to navigate you way through the search interface using those tools. 

Ok, so even though it’s so product centric, people are still quickly going through this funnel down into the actual decision-making phase relatively quickly and doing that on the Amazon platform itself?

Correct, just using the tools that Amazon makes available to them. It’s also worth pointing out that there’s quite a few differences with amazon Search platform, that are really critical. With Google you can point to a URL all day long vs on Amazon you are pointing to an ASIN or a product, and that product retail page. So, there is a whole network of retail metrics that somebody has to take under consideration there. Lost Buy Box meaning are you out of stock and a third party is now winning that buy box, speaking to in or out of stock, are you Rep out of stock, are you pointing to a derail page or you don’t even have any product available to sale, do you have poor reviews, are you an add-on item, are you Prime-exclusive? All of those are different metrics that you must consider, whereas on Google these simply aren’t factors. 

So, would you consider those to be ranking factors? Like is it in or out of stock, sounds like that would be a ranking factor, right? If it’s out of stock then they are not going to present it as an opportunity for someone to click on, is that what you call as ranking factors on amazon, or is it just some other words for thoseattributes.

To some extent I would call those ranking factors, but they are more like gates. They certainly are one of the most critical factors in considering what to surface within search, but it’s really a matter of off or on. If you are out of stock and there is no third party with an offer on that particular buy box, so there is no availability, it’s simply going to be search suppressed and not show up, which is not something that you have to concern yourself on Google of course.

So, you talked a little bit about technical differences on how maybe these different ranking factors and things like that. But can we talk more about the plays and the opportunities that people have with their Amazon account. Are they just putting keywords, is it keyword based or is it like the algorithm is deciding on what getssurfaced or is it just like I searched for dog bed and you have the word dog bed in the title? What are some of the attributes that people need to consider when they are thinking about trying to rank on Amazon.

First of, Amazon search is owned by Amazon’s A9 team, which is a totally separate team than what you have located here in Seattle, down in Palo Alto. They have built a pretty interesting mouse trap with Amazon search. Th structured data within Amazon’s catalog allow A9 to use many relevance features and the algorithm is constantly learning and combining different relevancy features based upon past search patterns and customer behavior. So, the algorithms are constantly changing. but one of the most critical things to understand about Amazon search is that it’s not one search algorithm. There really is a separate algorithm for every significant query on Amazon, and they are all adjusted for the unique challenges presented by each individual category. So, when we talk about how products rank within search, it’s really a 3-step process. the first step is about populating this consideration seat of ASINs based on relevance and suppression. Suppression we touched on briefly is about in-stock, is their an image for your ASIN and a few other factors, but that’s whether or not you are going even be able to be considered to be surfaced within a search

So that’s what you mean by a Gate? so Gate number 1 is like do you even have an image and are you in stock, if not then it’s just binary either yes or no?

Exactly, you are either off or on, and that really one of the factors that does not get enough consideration in my view when people talk about their Amazon SEO. They spend so much time focusing on content and keywords and all these hacky little tricks when it is fundamentally a holistic platform and you would be a lot better spent going back and working with your operations team to make sure you are in-stock consistently, so that Amazon can surface your ASIN.

So in that step 1, are those things like images, is it clear for those things for someone who is trying to set up their presence on Amazon that these are the priority, and that thee are the stage one priorities about how the algorithm is going to work, or is it kind of obscure for people to realize that’s so important

It can be pretty obscure. As with most things with Amazon, it can be a confounding black box, and it is not always clear what are the most critical inputs. But if you don’t have an image up for your ASIN, it is pretty clear when you go to search that it is not surfacing. I should caveat that there are a few categories - business and industrial, books, automotive, music, where few that don’t have an image and the products will still surface. There’s even some categories where you don’t have bullets or certain other content, you won’t surface either. For example in footwear, if you don’t have variations for sizes you may be search-supressed. So then when we are talking about the step 1 of suppression of relevance to get you into the consideration set, that the second component of that is content, and your content is going to include your title, bullets, description and then your brand, artist or author. With that Amazon’s search engine index is like the index in the back of a book, but instead of looking up a word and finding all the pages that contain it, Amazon look sup a word and finds all the products that match and their degree of text match. Title is the most heavily weighted piece of content. It’s going to consider the first 70 characters within that title. Now some less reliable resources out there will tell you to absolutely stuff your title full of content, but I would certainly not advise this because there is a variety of other factors that come into play here, most important being customer behavior. If you make your product listing appear sketchy and cram it full of keywords, you are going to critically start losing clicks and conversions that are important in your Amazon SEO.

You kind of mentioned like authorship is also being something. Is their something about the potential credibility like other than ‘Hey this was a good experience and I would rate this seller’, is their something about the content itself that points to reliability or trustworthiness or those kinds of factors Are those trust actors something that show up within this content?

Absolutely! so ratings and reviews are critical components that are heavily weighted. If it is third party offer that is the only one available for particular ASIN, they will also consider your seller ratings within that. Other content features include bullets, descriptions, browse hidden keywords. These are the hidden keywords that you are able to use when setting up your ASIN, and critically important. It’s about 250 characters that you are able to submit, and can represent spelling variants, synonyms, hypernyms related to products that customer might use to search for your products, and there is a whole art and a science to how you build these out as well.

So it sounds like maybe it’s similar to what folks may have done for Google and other common search engines where it’s like ‘Hey I am going to have my SEO section on my homepage that’s going to have stuff that doesn’t work anymore but it’s like maybe that still works with these hidden keywords on Amazon ‘Hey you do actually have to provide a bunch of different alternatives, focus area or help the algorithm understand the value of the page is’?

I think that’s a good analogy. There are a host of watch-outs here though, you find that often times people are trying to game these and trying to use a competitive keywords or things that might be loosely relevant, but you really don’t want your products to be surfacing where it’s not likely to generate a click, because the second step when we talk about SEO is customer success. That is the most important factor in your SEO ranking which is unique to amazon. So in that first step there is relevance based on your content and you get into the consideration set, and then what A9 is does is that it takes into consideration all these success metrics to determine which ASIN it is going to surface where within a search. So they score this based on a few different customer success metrics - click through rates, add to carts, and that’s also where the customer reviews come into effect. Then they weigh each of these differently based on the category and generally based on the average selling price of the products. These success metrics re weighted daily and that’s why things like Rep OOS are so critical, because if your ASIN is supressed for a day and you are not able to generate any clicks, it can heavily impact the overall search ranking for that ASIN. 

This is redone daily across the entire Amazon platform?

Yeah that’s correct. the first product that shows up in the search has the highest likelihood of generating most clicks and so it normalizes everything for position. I think one of the really interesting components of customer success inputs is thinking about how heavily Amazon weights the clicks within this equation. You would think that add-to-carts and purchases would be just the ideal metric to use here, but hey are really too scarce of an event to be weighting very heavily. clicks are just a much more prevalent signal, specially when you get it into these higher ASP categories, like if you are selling something in upwards of 75 bucks, using purchases is not going to be an ideal metric, but the clicks would be a much more prevalent signal to weight. 

Is that because Amazon realized that people are going through this learning process with it, where it’s like they view part of their role is not just selling you the dog bed but also being a source of information about gifts for dogs.

Exactly, and that’s customer journey towards the purchase. I earlier referred to every significant query having it’s own algorithm and we observed an instance of this earlier this year, when we noticed that within Consumer Electronics Amazon made a very observable change on how they were ranking on some of their branded terms because customers were being met with just a sub-optimal customer experience, and in this case we noticed that it was on branded terms within tablets. Previously, a customer might come in and search for say a Samsung tablet and just be met with a search results page that was absolutely loaded with tablet cases, whereas the customer intent was much more likely that they were coming and looking for an actual tablet but they had purchases at that point too heavily weighted, and you are just going to get much higher purchase frequency on tablet cases that are much cheaper than tablets. So we saw them making changes that were very observable, and changed the weighting there and soon when you searched for Samsung tablets, the tablets started surfacing at the top of the results. 

BIt makes sense right, that with a lower price points you will have more purchases and it seems like that’s not the best thing for that case. So, after content considerations, suppression relevance and these customer success metrics, what the gate number 3?

So the gate number 3 is what I refer to as simply greed. We have observed in our testing that all thing being equal there is a final step that Amazon is applying here, and it is pretty clear that is based on very high conversion and purchase opportunity. So there is a final weighting that we have observed that if Amazon has an opportunity to make significantly more money on a particular ASIN, then that one is going to win the day and it will be very difficult to unseat it. 

That’s funny. It sounds like with Google shopping results, it get’s their someday. It sounds like the weighting of the customer success metrics is more important or would you say based upon your observations or analysis that this conversion and value to Amazon takes precedence?

I think success metrics are really paramount here because that’s what is going to get you to that first page of search, and from there it’s a slotting game and things will be moving around within that page. So it’s critical to ensure that you are not supressed within search and you are available to drive those clicks, your pricing is competitive, your content is compelling, and you are running the promotions to make those conversions. 

So if you and I both go searching for dog beds. Am I going to see the same ordered results, assuming that we sat side by side in the same room and connection using generic computers? Would we get the same results searching for other dog toys? are the results personalized in some significant way?

I have done more incognito searches on Amazon than I like to admit. There are a few inputs here that could inflict a little bit of variability. Amazon is just constantly web-labbing all across the platform, and so it’s never quite sure when you are in or outside of one those particular web labs. 

So when you say web labs this is like them running tests or what does that mean?

Yeah, you are the gerbil and they are running tests on you whereas somebody else might be out of that set. Additionally, in certain CPG categories that feature a high volume of pantry products, pantry will suppress a listing if that ASIN is not available in the pantry fulfilment nearest to you. So if you are in a geography that is distant from me, and that toothpaste is not available to the FC nearest to be my IP address, but it is available in the one near you, you would see it and I would not. So there is a caveat there. some products are Prime exclusive and that’s generally what Amazon does based on profitability, where they try to reduce the volume of unprofitable ASINs.

That’s specific to Pantry?

Yes, if you are on normal Amazon.com, it’s still going to surface. That’s just exclusive to Pantry, because they are trying to drive profitability with that program. 

Interesting about the location and the fulfillment aspect. From your perspective, let’s say we have a listener out there that is trying to think of developing their Amazon presence, what is really the first thing that they should do or a common mistake that people start out with?

My guidance would be to focus on the fundamentals. First understand what the Amazon equation is, it is Sales equal to traffic times conversion, and traffic we often call glance views. Glance view is an Amazon term for detail page impressions with a fast track offer. So traffic times conversion equals sales, it’s pretty straight forward. so have a plan for driving traffic and don’t just rely on SEO and customers discovering your products unless you have a tremendous amount of branded search frequency. So unless you are Nike you better have a plan for driving traffic to your products, and understand what drives conversion when they get to your detail page. Get your team aligned and engaged because on Amazon you are operational metrics are also marketing metrics. So you need to have reliable in-stock, low vendor lead times, and have your operations willing to handle those small initial orders that you are going to get, because the first POs to start receiving from Amazon are pretty underwhelming. But it grows from there and you start spin up velocity on the platform and it will grow in time. 

So Connor, really great overview of Amazon and how it works. Some of the different factors that people might consider as well when trying to get started themselves. You run this business Downstream, who comes to downstream and at what point is it valuable to consider a service like Downstream to help with all of these different aspects of managing Amazon presence?

We do with brands of all sizes. I have worked with brands on the platform from biggest vendors in the world to mid-size vendors to ambitious vendors who are willing to invest and lean in into the platform. But everyday Amazon is investing in how they can drive more profitability out of their retail programs by doing more with less resources. AI and machine learning get a tremendous amount of press, but What receives less press is really more critical to the brands that sell on Amazon, is that Amazon is doing much of this efficiency by shifting a significant amount of burden of managing and driving that vendor’s catalog back on to these brands. And they are doing this via an ever-increasing number of self-service platforms. AMS, Amazon’s paid search platform is primary example of this. It is something that previously would have been managed by a marketing head within Amazon who would sell a merchandising package and execute it. Now that is a self-service platform as are many others. Additionally, Amazon is now the number 1 product research platform in the world, getting more product research than all search engines combined. Yet many brands don’t have their sales force managing this. So that’s not an optimal way of managing the largest product research platform in the world. So most brands found that they don’t have the resources nor the expertise to manage all of this. So, many are aware that they need our services or services from organisations similar to Downstream. It’s just a matter of getting buy-in from leadership and having the brand and sales forces come together and aligned on how best to move forward on this platform that has become a hybrid brand and retail platform.                   

It’s so interesting to think about how Amazon is changing. The way that you have described here, shows how comprehensive it is and how essential it is to people’s buying behavior beyond just simply ‘I am going to go look at a particular product and click on the purchase button’. It’s like there is lot more going on there that people need to think about. Connor, thank you so much for your time. Really appreciate you coming on the Mozpod and chatting about Amazon. It’s been really awesome. Thank you!

Thank you! It’s been a pleasure.

Stay competitive.

Subscribe to keep up as we share our insights on the entire channel for brands running competitive campaigns.

Product

See the Downstream advantage.

Pricing

See our flexible pricing plans.