First a disclaimer. I don’t have any secret information from the company itself. This article is based on general information about how commercial search engines work and then reasoning about which of these apply to Teachers pay Teachers’ search algorithm plus some experimental findings.
Search algorithms of other online commerce sites
There are tons of online shops and search algorithms out there, and first we have to pick some of them for a closer look. We want those sites to have some things in common with TPT such as
- The site is selling something
- There are a lot of individual sellers on the site who have their own shops.
So, we won’t be comparing TPT to Google, however interesting search engine it is because it’s not selling products directly to consumers. I have picked three large online shops to compare: Amazon, Etsy, and eBay. All these three are public companies, TPT is privately owned. To set these to some sort of scale: Amazon has over a million workers, eBay has over 10,000. TPT has over 2,000 and Etsy over 1,000. So, you could expect Amazon’s algorithm to be the most refined, because they have more resources to put into it. For the rest of the companies, the fine tuning of algorithms depends on how much effort they put to them. But for purposes of this article, we assume that TPT’s search is about the same level as eBay’s and Etsy’s and all those look up to Amazon’s, thus it’s important.
Below, I list factors that matter for the search algorithms in some of these sites. After each factor, there are one or many of these ecommerce sites which utilize these factors. This information is based on sources that can be found from the end of this post. Note that if a site is not listed, it doesn’t necessarily mean that it doesn’t use that factor. This list is in no particular order.
Title relevancy
(Ebay, Etsy) How well the product title matches the thing the buyer is looking for. This is more important if the buyer is using longer search terms, it would be hard to rank high by using some very general search terms like “fractions”, even if some people may do that exact search. The best long term strategy is to use titles that describe your product the best way, so use all the words that are relevant as long as it’s easily readable to a human buyer and don’t repeat words in the title. Algorithms generally punish repeating keywords too much.
Seller authority or track record as a seller
(Amazon, Ebay, Etsy) This is definitely a ranking factor in all those marketplaces. In TPT context, that would mean the quality of your shop. TPT keeps the seller’s star rating at the upper corner of every shop’s page, so surely they think it’s important. That being said, TPT’s ranking system has been quite monotonic for years resulting that almost all old sellers have full star ratings. However, the system has been changed in order to get more variety to feedback. From all of that, we can reason that for e-commerce, shop reputation is generally important. Maybe it isn’t that important in TPT yet but surely it will be in the future.
Organic sales
(Amazon) An organic sales is when a buyer searches something, finds and buys it. If the buyer finds the product via promotion, that’s not an organic sale. There is no 100% way of knowing if TPT distinguishes organic and promoted sales in future search results, but it’s something that may have some effect.
The price
(Ebay) The price most likely affects the Amazon search too, as the Internet is half-full of sites about how you find the right price point on Amazon. It’s reasonable to believe that this also affects Teachers Pay Teachers’ search, but from an ecommerce site’s point of view it’s not that simple. People will generally buy cheaper products easier, but if products are too cheap, the revenue is harder to get and people tend to think less of products that have too low price. So, an intelligent search algorithm designer would favor middle priced products.
Off-site sales
(Amazon) Are you bringing more customers to the site by advertising your product? Then, you are a good seller! Maybe you even get some boost in the search engine. My own experience with TPT is that I rarely get sales from social media sources such as Pinterest. But there are times when I focus more on Pinterest marketing and more traffic from Pinterest is coming to my shop at these times. Even if those people are not buying right there and then, also my organic sales are higher at these times. So, it might have an effect. (Wishlisting is also a possible contributor here.)
Quality of product description
(Ebay) This is easy to implement in TPT, because you can always write a good product description. Some of the other factors are not in seller’s direct control. A good description has everything you need to know about the product and also all the relevant keywords. But again, stuffing the same keyword over and over again is punished nowadays.
Conversion rate
(Amazon) Conversion rate is a ranking factor on Amazon and because TPT lists it as one of the product statistics it’s probably a thing in TPT, too. Logically, the bigger, the better.
Listing of terms of service
(Ebay) This idea of being clear about terms and service could translate into TPT language in licensing policy. If TPT takes this to account, it is likely to favor those products that have a clear licensing policy written down in the description.
Listing quality
(Etsy, Amazon) If a buyer makes a search and then clicks a product on the listing, that counts towards listing quality. In TPT, if there is such a feature, good thumbnail pictures will help you with this.
Recency
(Etsy) Other things being equal, a new and fresh product is better than ten years old. It may not feel that way because other factors tend to favor the older ones. I have heard rumors that updating a product in TPT would have a positive impact and that is in line with recency.
Other interesting search factors from these sites:
- Words at the beginning of titles are considered more important than words at the end.
- Past intellectual property infringement issues can have a negative effect.
- Store location in relation to buyer
- Shop’s information page about the seller
Experimental findings about Teachers Pay Teachers’ search algorithm
Because your search history may affect the results I did some test searches when I was logged out of the site. If you use some popular search term, the results in the first page have pretty much everything right, so you really start to see more interesting patterns when browsing more pages down the search results until the matches are not “perfect” any more. Here is something I found:
This is from page 9 with a search “Adding and Subtracting Fractions with Unlike Denominators”. This may sound a bit arbitrary thing to search, but in fact it is one of the popular search terms every now and then. Grade level at this search was three. (There is no actual reason for that, I forgot to take it off after the previous test.) I didn’t take the whole page in the pictures, but the idea is that most of the resources have only a couple of stars, while the resources in page 1 would have hundreds of stars. But there are few exceptions, pointed out with arrows.
It took me a moment to figure out what was dropping these resources down. I have covered the seller names in the pictures, but I checked them and they were perfectly top rated, established sellers. The problem was that these products have very few pages. The ones that cost $1.00 had only one page per product. So, the lesson is, that if you have very few pages accompanied with a high per page price, your resources will not do well in the search.
Summary
Based on logical reasoning, we could expect at least the following factors to have something to do with the order of Teachers Pay Teachers’ search results:
- Title relevancy
- General shop quality
- Organic sales
- The price
- Off-site sales
- Product description
- Conversion rate
- Clear listing of license terms
- Customer clicks in search listing
- Recency
- Price per page
Hopefully this article was helpful! Have you experienced some weird rankings in Teachers Pay Teachers search? Leave a comment!
Sources: Amazon’s A9 algorithm, eBay optimising listings, How Etsy search works.