|EcommerceBytes-NewsFlash, Number 1933 - December 09, 2008 - ISSN 1539-5065 1 of 3|
Raghav Gupta was a Research Scientist at eBay Research Labs from 2002 - 2008, where he invented, evangelized, built and released some key search technology. Raghav left eBay in October to launch his own start-up, GeoTerrestrial, which is building a next-generation dynamic location-based-services platform. We checked in with Raghav to find out what it was like to work on eBay search and learn more about his new venture.
While he touched upon relevance in search, Raghav chose not to answer specific questions about how to improve performance in eBay's Best Match because of its complexity, he said, and because his broad intention was to steer discussion away from problems and instead try to arrive at solutions.
AuctionBytes: You've worked on some of the most important eBay search technologies, tell us about the significance of search features to a site like eBay.
Raghav Gupta: If a service has to only deal with a strict fixed-price catalog and a few thousand unique products, search is quite trivial, both in visual presentation and backend implementation. But add in freely structured item listings and auction based formats, the picture suddenly changes. On top of that you add volumes measured in the hundreds of millions with thousands of items expiring and getting listed every second. Together these aspects create conflicts not easy to amalgamate, both in presentation and implementation.
AuctionBytes: Were you give free reign to pursue your own creative ideas, or was it a more structured environment?
Raghav Gupta: In late 2002 when I joined eBay as an engineer, as far as creative freedom was concerned, it was a gulag. A lot has changed since then. There is actually a lot of good innovation happening nowadays in terms of demos and prototypes and contests, but hardly anything worthwhile ever makes it out. The personal cost of having to push something down the approval and implementation pipeline is so great that very few are able to persevere. And whatever does get out usually suffers through so many Dilbertian compromises that it is missing the core aspect of the original idea.
Personally, ever since I built the Related searches product in 2005, I have enjoyed pretty much a free rein to pursue research ideas. But the control (and the vision) is forfeited as soon as an idea gets accepted to be built and released to production.
AuctionBytes: What was the most interesting project you worked on while at eBay Research Labs?
Raghav Gupta: It depends on how you define "interesting". If it means "technically challenging", automated language translation would be it. If it means "potential to vastly increase eBay's reach", then I think AdContext would be the one.
AuctionBytes: In reading over some of your writings, you come across as part research scientist and part philosopher. You have written, "Only a deeper introspection reveals that behind every piece of data, behind every click, there is an individual human being, who possibly earns his/her living on eBay." How important is it for scientists to think about the people who are using the tools they create?
Raghav Gupta: The definition of scientist is so broad I think it highly depends on the field. A weapons scientist probably has very different thought processes compared to a medical researcher. At the end of the day, a paid scientist in a company is just another employee. Any employee could choose to agree or disagree with the direction the employer is taking.
AuctionBytes: Can you always foresee how users are going to use your tools, have they ever used them in a way you never envisioned?
Raghav Gupta: No one can ever envision in advance all the ways how something will be used. Should the "pen" not have been invented because it could be used to write bad things? Minds harbor malice, not machines.
Conversely, possibility seeds intent. If someone openly sold a one-button device that vaporizes the world instantly upon pushing it, it would have been pushed by now.
AuctionBytes: You led the "relevance sort" production implementation built on the prototype you developed. Can you explain what relevance sort is on eBay?
Raghav Gupta: I cannot speak for what it is on eBay, I can speak for what I think it should be. "Relevance" in a search context is not a property of the result being displayed, it is a measure of how correct the individual user thinks the result is. The goal of a quantitative search function has to be to provide a user the tools in an intuitive, uncluttered manner so he or she can quickly pull up what is relevant to that user. In an auction based scenario, additionally, users must never get a nagging feeling in the back of the mind that they've "missed something".
The difference between web-search and eBay search is that web-search is based on relatively static documents conveying mushy abstract information, and the search engine attempts to approximate the perceived relevance as a number, because it has no other choice. Whereas on eBay, every result has quite a few absolute hard physical properties like end-time or price or seller feedback score etc, and "relevance" of a result in this case can vary vastly from user to user. Therefore, in my view, the goal should be to provide the means directly to the users to be able to create their own definition of relevance for a particular search.
AuctionBytes: Does the "Recent Sales" factor in Best Match have anything to do with your vision of Relevance Sort, and in your opinion, should it play a role in Best Match?
Raghav Gupta: No matter what the "factor", I think individual users should be able to choose the factors they want to sort/filter their result by, without anything being imposed. In that scenario, you will agree that the more factors available, the better. Imagine being able to combine keywords and factors as you see fit to create your own perfect search. Play with it, save it, advertise it, email it, whatever works.
AuctionBytes: You developed "Contextual Keyword Extractor," aka eBay AdContext. Is this live to the site, and do you think it will be significant to eBay's advertising revenue?
Raghav Gupta: We worked on this way back in 2005, and I personally believe it could have been the single most important win-win product in the lifecycle and growth of eBay and its community. It had all the means to expand the reach of eBay outside its own website, out onto the entire web. It was the one thing that could have really competed with Google Adsense. Some readers will remember the headlines it made when it got announced during eBay Live 2006. Unfortunately, I have no idea what happened to it or where it went, or where it is now.
AuctionBytes: What is eBay doing right these days in your opinion?
Raghav Gupta: eBay is a large company, and even I don't know all the various departments and the work they do. Contrary to popular belief nowadays, eBay is actually full of hard-working upright folks. Remember, behind every perceived evil corporation there are employees who are working hard to make ends meet, and who actually do the right thing whenever faced with a decision.
AuctionBytes: And what is eBay doing not so right, in your opinion?
Raghav Gupta: I've left, so anything I say would be hypocritical, and probably even turn out to be incorrect.
AuctionBytes: What went into your decision to leave eBay?
Raghav Gupta: At my age, 6 years is a long time to be in one place.
AuctionBytes: Many users have seen your blog post, Ode to eBay. What were you trying to say with that poem?
Raghav Gupta: That poem was the content of an email I sent to a select group of eBay employees the day before I left. Since it got out on the net, I later decided to put it up on my own blog. The target audience while writing it was eBay employees.
AuctionBytes: What do you see in the future of eBay?
Raghav Gupta: I have a lot to say in this department, but I'll try to keep it short. I won't talk about what eBay should be doing, instead I would like to talk about what I think the community should be doing.
Here is my thought process. Every seller on eBay is in effect running a business. The primary target of a business is to make money. By definition a business has the right to do everything it takes to make more money, as long as it doesn't break the law. The legal responsibility, or accountability, stops there. Imagine if a small new seller comes to an established seller of similar products and says "I expect you to increase your prices so my products will get a chance to sell. Its your responsibility to help me while I'm starting out". Absurd, and totally against the concept of running a business, isn't it?
The point being, all things considered, eBay is just another for-profit corporation in the business of making money. Think of it just like another extremely large seller. It has no legal responsibility to "care" about you. As long as its shareholders are happy with its progress, it need not listen to you at all. I know it sounds depressing. Enron collapsed, Yahoo is dying a slow death, stuff like this will keep happening. This is the fundamental flaw in building a life-supporting small business which is so horribly dependent on another business for its very existence. All the risk, but none of the control.
The bright side? The success of eBay has validated the business model, demonstrated the huge market, and also created this large close-knit community.
Let me also repeat something most of you already know. The entire wealth generated by eBay actually has all come from your pockets over the years! The salary of all the customer support people, the engineers, the ad-campaigns, the government lobbyists, the CEO's corporate jet, the server hardware, the cost of running the data centers, it all comes out of your pockets. This is not a bad thing, rather it actually shows that you as a group hold the underlying controlling stakes in this operation.
But if you are paying all the expenses, how come you have no control over what your money does? More importantly, if you have the money, why are you giving it to a third party like eBay at all? Why would you give it to any third party at all?
To me the success of eBay proves that the community as a whole has enough resources to build and operate its own alternative. Instead of paying a 3rd party to use its selling platform, pool your resources into a non-profit organization, elect your leaders democratically and task them with hiring the right people to build and manage your own infrastructure. As the platform matures, large groups of sellers can migrate to it en-masse. Where the sellers go, buyers will come.
It is not as difficult as it might appear to some. In my view it is probably the best long-term alternative where users can focus on building their businesses and compete with others rather than constantly having to fight the system itself. All that's needed in the beginning are a few motivated, capable and trusted members of the community who can kickstart this thing the right way. Once the foundation is laid, an explosion of goodwill and support from unexpected quarters will put it all on auto-pilot.
AuctionBytes: Tell us about your new site, GeoTerrestrial.
Raghav Gupta: GeoTerrestrial aims to be the default service provider for all location enabled services. Friends near you right now? Recent allergy outbreaks in your area? A store holding a short term sale and you happen to pass by? The GeoTerrestrial platform will power your phone with the means to answer all these questions, and more.
AuctionBytes: Do you see GeoTerrestrial ever becoming as big a company as eBay?
Raghav Gupta: If I play it right, yes. The response from users of our current release for Windows Mobile is beyond expectations. Sure the economy is in bad shape, but we need time to build anyway. In a way it is a plus for me because it will flush out competitors who don't have it in them to persist through this downturn. Besides, over the years I've learned to trust my own gut instinct, I also have a whole bunch of well-wishers, and most importantly I have an unbounded capacity for hard work. Players against me are up against some very tough odds.
Comment on the AuctionBytes Blog
About the author:
Ina Steiner is co-founder and Editor of EcommerceBytes and has been reporting on ecommerce since 1999. She's a widely cited authority on marketplace selling and is author of "Turn eBay Data Into Dollars" (McGraw-Hill 2006). Her blog was featured in the book, "Blogging Heroes" (Wiley 2008). Follow her on Twitter at @ecommercebytes and send news tips to firstname.lastname@example.org.
You may quote up to 50 words of any article on the condition that you attribute the article to
EcommerceBytes.com and either link to the original article or to www.EcommerceBytes.com.
All other use is prohibited.
Email this story to a friend.
1 of 3
AuctionBytes Interviews Former eBay Executive Jeff Jordan - December 29, 2008
eBay's Jim Griffith Talks about Changes Impacting Sellers - December 15, 2008
'PayPal Damon' Weighs in on the New eBay - December 03, 2008