eBay’s Voyager platform currently drives the company’s massively complex algorithm – the core offering attempts to connect buyers with the best possible product/service through an underlying “Best Match” algorithm. If you have spent hours searching for that product/service you desperately seek – only to conclude your 2-hour exhausting search with 13 or so possible unconvincing options in your watch list, you have experienced the shortcomings of the current platform. However, there is good news for you: eBay’s ultra top-secret Cassini platform is just around the corner.
As a dedicated, passionate, and experienced eBay seller (beardsley wellington), I have the pleasure of providing a glimpse into what appears to be a possible working version of the eBay’s Cassini search platform in raw form. As a courtesy to eBay, the publicly accessible URL and the related eBay lab name have been omitted. While I am not a super tech blogger, I would like to believe that I am fairly intelligent enough to present the following data, along with a few reasonable conclusions and several predictions.
In an act of shameless, but hopefully forgivable self-promotion, the following are several screen shots – all reflecting default values for the query “ties,” along with a filtering the results for the eBay seller beardsley wellington.
While it is yet to be determined if this is in fact a working version of Cassini, the interface stationed within an eBay labs sub-domain is likely a time machine into the near distant future of things to come for eBay.com.
eBay’s biggest current problem is extremely complex in nature, but fairly simple to explain. I will describe the issue in Alabama terms:
Buyer’s perspective – there is a lot of stuff out back for sale that I probably want at a great price from a great seller, but I cannot see it with all of this junk everywhere – so I guess I will settle with this from this seller because I have to go pick Billy Joe up from school. I may or may not come back to spend more money because I am not truly happy with what I got the first time. I think I may go to that property named after that river the next time I get a paycheck.
Seller’s perspective – I am on a tropical deserted island with Luxury goods. I can see my potential buyers cruising by on the titanic heading straight for a glacier. I am screaming at the top of my lungs to try to save them, but I have to come to the realistic conclusion, they will never hear me. I think I am going to go sell my Luxury goods at Uncle Donnies – he’s got a great setup near the Amazon.
eBay must scale to flatten the environment in order to increase the probability of creating meaningful, relevant connections. eBay knows this – exit Voyager and enter Cassini.
WHAT IS CASSINI?
From a buyer’s and seller’s perspective, I believe the Cassini Platform can best be described in two words: Resolving Relevancy.
The Following is a fantastic summary of the Cassini Platform during a 2011 tech conference by Hugh E. Williams – Vice President of Search, Experience, and Platforms for eBay Marketplace: click here
THE DETAILS OF MY DISCOVERY
Within a publicly available sub-domain of eBay labs, the company appears to be testing one keyword each day within a search interface never before reported. The purpose of the testing remains unknown. The following screen shot was taken on September 6, 2012. Please reference the bottom of the High-level Category Listing. which shows duplicated Categories for the Clothing, Shoes, and Accessories (Category 11450) and the Books (for some unknown reason). Please note the acronym “QA” before these category titles.
Ironically, the keyword tested on September 6, 2012, was “Alden Boots” (Please reference Example #1 below), which is an extremely scarce product within Category 11450; thus, linking the “QA” with the corresponding daily search. The acronym “QA” most likely represents Quality Assurance. It is unknown, but likely, quality assurance testing was being performed.
The following are 4 daily keyword searches that I observed and documented, including today’s daily search of a super sweet Bane Mask that I was almost tempted to bid on (kidding….I’m stuck on dressing up as Vladimir Putin for Halloween):
1. Date: September 6, 2012 – Query:“Alden Boots”
2. Date: September 12, 2012 – Query:“Halloween”
3. Date: October 9, 2012 – Query:“iPhone 5 Unlocked”
4. October 14, 2012 – “Bane Masks”
The “New” Best Match – Based off eBay’s daily query results, the search results appear restricted in nature, but in fact, I believe each individual eBay buyer will have unique search results relevant only to each user id. For example, if two different users sitting next to each other using two unique user ids were to search the same query at the same exact time, the query may yield different results. The result – organic relevant personalized search results in real time. I believe being a top-rated seller and the value of Detailed Seller Ratings (DSRs) will become less relevant and meaningful in search placement – in the long-term.
Limitation of Search Results to less than 200 per query – I believe eBay is preparing to go FREE listings across all Categories with the launch of Cassini in 2013. Based off eBay’s daily query results, this search algorithm appears to be acting as a smart filter in fulfilling eBay’s core platform offering – connecting buyers with the best possible sellers available at any given moment in time.
One extremely interesting item to note – even if a given search result yields 10,132 returns, the result may only display a theoretical 62 listings to a user. This is reflected above in the October 14, 2012, “Bane Masks” search query, but with different numbers.
This is even stronger evidence that reflects the inevitable of what appears to be Cassini.What could this possibly mean for Craigslist? The application of the smart filter at local level will be interesting to watch and Craigslist should view this as immediate direct threat…
Good/Bad Related Searches & the Query Builder – eBay appears to be incorporating several other current eBay Labs into this interface for these Gadget offerings – including Buzz, Catman, and the Listing Title Optimizer Tool.
Incorporation of what appears to be a New Reputation System…finally – Emosi Sosial, a company eBay acquired several years ago, used to have an active home on the eBay labs homepage circa 2008 (See one of the screenshot’s below), then disappeared. The first screenshot below depicts the Emosi Sosial bunny icon next to each seller’s user id. This most likely implies a Emosi Sosial value has been assigned to each user id, which in turn – would have an effect on the user defined TVS beacon. Maybe this will answer a theoretical question: What’s the relevant (to me only) difference between 2 top rated sellers selling identical products with over 1,000 Feedbacks combined 100% Positive Feedback?
The TVS beacon appears to be a gadget customizable tool that allows users to filter results by 1 user-defined seller preference or the default value (reference the Manual Graph TVS setting below) in addition to 10 defined values (all listed in screen shots below). This allows users to customize based off a 100 point scale based off of Trust, Relevance, Value, and Diversity. One feature I was disappointed about (from a totally geeky perspective) – the search was not “instant,” like google instant. For each individual query after changing the TVS beacon, you must press search.
WHAT IS REPEX? Additionally and extremely interesting (although possibly irrelevant), I also discovered an eBay lab for “REPEX,” otherwise known as a “Reputation extractor Prototype,” circa 2008. Similarly, like Emosi Sosial lab appears to have disappeared within the same time frame. With Emosi’s emergence in this search interface, it is possible this application is incorporated in what appears to be to be this new form a of reputation system.
Both of these applications appear to have the same end goal – gathering every piece of possible available information about a seller in the effort to connect with the perfect buyer at any given time.
Sellers beware! It could appear that luke warm positive feedback now is being interpreted in the new smart filter described above.
Additionally, there have been several recent privacy revisions – one specific revision which specifically resulted in the recording of customer service phone calls (not just for “quality assurance purpose”) as most do.
Consider this as a possibility – the text conversion of your phone call where you may have insulted a customer service agent could possibly impact where your product lands in the search results. Most would think this is a stretch conclusion, but after examining the data for the past month, I believe any and all user ebay id user data to be fair game.
HIGH LEVEL – WHAT THIS POTENTIALLY APPEARS TO MEAN FOR BUYERS/SELLERS?
Buyers – Relevant Optimized Organic Personalized Search that is continuously getting better with every user click.
Sellers – While most eBay sellers would like to believe Cassini is only a satellite in deep space orbit, you must accept and prepare for the upcoming change/transition to Cassini model. Please check back at my blog, as I plan on detailing what you should be doing to prepare for the potential upcoming changes.
Both Buyers and Sellers – Among a handful of potential impacts, I find this to be the most groundbreaking: Assuming an efficient market, it appears theoretically possible price equilibrium could finally be attained for an online auction, over the long-term…..good bye winning that $50 item for 0.99 with free shipping during Kickoff of Super Bowl Sunday!!
While it is not public knowledge when Cassini when the working version of Cassini will go public (given the Cassini project has already missed it’s initial deadline – click here), I believe the company will launch post Christmas 2012. Mid-January 2013 makes the most sense to me….
Additionally and from an experienced seller’s perspective – it’s even possible given all the current irregularities/inconsistencies within the recent selling process, there are pieces of Cassini operating within Voyager right now. In effect, softening the blow of the implementation on short-term revenue expectations, which will most likely decrease with the incorporation of Free Listing across all categories.
BIG PICTURE – Most pit eBay solely against Amazon. I believe eBay biggest competitor is itself. The e-Commerce race will be won by those organizations who can accurately predict our behavior (i.e. what you and I want with out available capital) with a high degree of probability including but not limited to the following ways:
The acquisition of this information will result in future advertising revenue in unpredictable forms.
We interface daily with Facebook “Likes,” “Wish Lists” among multiple platforms (all of whom appear to have copied Amazon), and Geo-location tracking with that fancy new Apple iPhone 5 you got a few weeks ago. Where does all that data go?……into understanding what you want. Cassini will do this for eBay in the long-run – when combined with eBay’s historical sales database.(/p>
Regardless of whether or not this information presented is confirmed by eBay to be some working form of Cassini, I still strongly believe the information herein to be relevant to eBay buyer’s/seller’s in understanding what is in the pipeline. Given all of the variables presented and some left unpresented, I strongly believe this to a working form of Cassini – regardless of how eBay responds.(/p>
Sadly, you or I may never know. One giant problem that will exist for eBay even after Cassini- Transparency.