In this first in an advanced series of posts dedicated to eBay’s most recent search engine, Cassini, I provide the main purpose of the technology, a brief history of it, and a defense of its importance.
When was the last time you searched, “eBay SEO,” in Google?
If you haven’t done so recently, go ahead and do it now.
As our website is new, you won’t currently find us near the top of the results (something we intend to change!).
What does appear, however, is enlightening.
Because of how much it sucks.
Whether it’s one of the following —
- CrazyLister’s myth-laden crapfest that actually recommends sellers use other companies’ trademarked terms in their eBay store names,
- Shoplo’s misspelling frenzy (“Terrapeak,” anyone?) that blatantly confuses eBay’s search engine with its default algorithm,
- Danna Crawford’s misguided piece that tells sellers to optimize their listings by using item description heading tags and keywords in the saved names of their listing images,
- or even eBay’s own sparse SEO page that suggests sellers add alt text to their photos (when eBay is already doing it for them)
— it’s clear the eBay community is being consistently misled, unintentionally or otherwise, in the name of search engine optimization.
In addition to SEO itself, those who write about the subject inevitably end up talking about a little thing called, “Cassini.”
Most eBay sellers have heard of it, and most “gurus” feign an expertise in it.
But the truth is, I see so much online confusion about Cassini that I decided to write my own series about it.
If you found this article and were hoping for a handful of easily-implemented tips and tricks, or heaven forbid, wanting to learn how to “crack Cassini,” you are in the wrong place.
At List Rank Sell, we are only interested in the truth — even when it gets heavy and a little cumbersome. Whatever the case, we don’t do things like they do.
And although I’m by no means a Cassini expert, I also know more than they do.
So let’s get started.
What is Cassini?
Cassini is a search engine, just like Google or Bing. A search engine is basically a portal to a database of information, enabling users to query its data using keywords, images, and in some cases, even vocal commands.
Cassini is a search engine, just like Google or Bing. A search engine is basically a portal to a database of information, enabling users to query its data using keywords, images, and in some cases, even vocal commands.
The main difference between eBay and, say, Google is that eBay’s search engine, Cassini, returns eBay product listing pages instead of webpages or other internet resources.
But in both cases, the engine accesses an index of information important to the searcher. The better the search engine can return its results, the more likely the user will come back to do another search.
Once Cassini has decided upon its results, the role for ranking them falls to Best Match — eBay’s default algorithm.
Best Match is a complex ranking formula that takes the set of listings given to it by Cassini and then determines which should rank first on the page, second, third, and so on.
But what listings appear in the first place is not decided by Best Match. They’re up to Cassini.
Cassini completes this task by using machine learning to understand search “intent.” In other words, it does its best to determine what buyers actually want.
It’s not easy for technology to know, for example, that I’m looking for the “God of War” DC Comics series (or maybe even the more recent Dark Horse Comics series) and not the popular video game.
Cassini tries to know just that by learning from many things, including my own search behavior.
To be honest with you, its ability to do so currently isn’t what I’d like it to be.
However, it must be understood that, in spite of its issues, eBay’s search engine has still come a very long way.
A brief history of Cassini
Much can be learned about eBay’s search history from two excellent resources. The first is a decade old eBay presentation entitled, “The eBay Architecture.”
By 2002, seven years into the company’s development, “eBay search had reached its limits” and was in need of something new.
After a handful of early off-the-shelf solutions, eBay had developed a unique set of search requirements that required a homegrown solution.
The answer was eBay’s first major search engine, Voyager — which, like Cassini, was named after the NASA space probe by the same name.
Veteran eBay sellers will remember Voyager as an unsophisticated, Web 1.0 search engine that was extremely literal in its interpretation of buyer search behavior.
According to Hugh Williams, the chief architect of Cassini:
“Voyager…delivered impressively and reliably for over ten years…it was architected before many of the modern advances in how search works, having launched before Microsoft began its search effort and before Google’s current generation of search engine.”
You simply cannot talk about the history of search at eBay, specifically Cassini, without referencing Williams’ prolific writing and work. It was his job to help eBay turn the page away from Voyager and enter the modern search age.
As early as October 2010, Williams and his team at eBay began to build eBay’s second major search engine, making “fundamental decisions about [its] architecture and design principles.”
By August 2011, Williams described the Cassini project as “on the scale of any major search engine re-write a company has done.” Williams’ goal was a “world class search engine” that could be “easily extended without the need for another massive reset in 10 years’ time.”
Finally, after some confusion about its initial rollout, eBay CEO John Donahoe officially announced in January 2014 that Cassini was now live on every eBay platform (except Korea).
Since that time, Cassini has provided the backbone for everything search-related at eBay.
And that’s too important to talk about like it can be easily manipulated by a high schooler with basic HTML knowledge.
Why Cassini matters
Cassini is at the very heart of a seller’s success, and most of those talking about it online simply don’t understand it.
The typical “expert” speaks so blithely about having Cassini figured out, they do this impressive software system a great injustice.
Consider just these few items:
- When it was under development, Oren Etzioni, chief executive officer of the Allen Institute for Artificial Intelligence, considered Cassini to be a direct threat to Google.
- Since its initial rollout, constant improvements have been made to Cassini, including regular increases in its processing efficiency.
- In order to better run Cassini and its resources, eBay is literally developing its own custom-designed servers and data centers.
I’m the first to admit that eBay’s most recent search engine is by no means perfect — as we will see throughout this series.
But we won’t just see how fallible Cassini is.
We’ll also see how truly central it is, not only to search efficiency on eBay, but to making sales.
And that’s why we all do this in the first place.
Keep your chin up.
