Delicious: The Missed Opportunity

Yahoo have made some pretty shocking decisions of late, but the one that really stands out was the way in which they managed to totally squander the opportunity that Delicious gave them over a 5 year period.

When Yahoo bought Delicious in late 2005, they paid something like $20 million for it.  This wasn’t a huge amount considering News Corporation had bought MySpace for about $800 million a couple of months before, and it was totally dwarfed by the $1.6 billion that Google paid for Youtube a year or so later.

Back when Delicious got sold to Yahoo, I remember having a conversation with my boss about how Delicious was going to be a huge asset for Yahoo because we thought that Yahoo would eventually integrate a social layer into their search results based on the recommendations within Delicious.

Of course this never happened as it should.  Yahoo may have integrated some of the Delicious data into their core search results, and they may have used some of the user information to discover new content, but I doubt it ever happened to the extent that it should have done.  Although its on a bit of a decline, Delicious still gets more than half a million unique visitors a month according to Compete, and there is a big user community knocking around who will still add bookmarks regularly.

Even though socially curated search results would probably have put us out of work as in house SEOs, it was still incredibly exciting to look at the possibilities of how mass voting could create a more relevant and useful set of results than Google had at the time.

How good is Delicious as a Search Engine?

I decided to compare the top results in Google and Bing (US versions) with Delicious for a few popular terms to see whether there were any similarities.  These are the results I got:

Car Insurance

• Google #1 Geico.com
• Bing #1 Carinsurancerates.com
• Delicious #1 Geico.com

On this one, it’s worth noting that the #3 result in Bing was Geico.com, so there was a close correlation between the results.

Hotels

• Google #1 Travelocity.com
• Bing #1 Travelocity.com
• Delicious #1 Tripadvisor.com

There was clear agreement between the search engines for the number one site here, but it is worth noting that Bing rated Tripadvisor as the second most relevant website in their results for the term.

Mobile Phones

• Google #1 Wikipedia
• Bing #1 Wikipedia
• Delicious #1 GSMArena.com

There was plenty of agreement between the search engines – they both had the same two websites at position 1 and 2 (T-Mobile), but in this case there was very little similarity between the results on Delicious and the search engines – although there is a certain amount of ambiguity in how the websites were being tagged.

Maps

• Google #1 Google Maps
• Bing #1 Yahoo Maps
• Delicious #1 Google Maps

Not a great surprise as both the major search engines featured their own/partners product at number one, but again there was a close agreement between Delicious Users and the search algorithms about what was the best result.

What does this mean?

Ultimately, this is a very small sample – only 4 terms out of the millions that people use each day, and it is skewed towards high volume terms that are very competitive, however it’s interesting to see that for these types of keyword, Delicious certainly holds its own in terms of the results that its users are able to collate.

The limiting factor for any socially driven website is the user base.  Delicious peaked at around 5-6 million regular active users, and has a few hundred million bookmarked URLS, which means that at best, their index is around 0.2% as big as Google’s.  This would be a big problem for any attempt to provide representative and useful results at a deeper level.

This could have been so different.  If Delicious’ technology had been explicitly integrated into the Yahoo search results back in 2005, it would have had access to a giant user base, and could have been used to supplement the results with a social layer.  Votes and tags could have been built into Yahoo’s ranking algorithm, and users could have been encouraged to participate.

The big two search engines of 2011, Bing and Google are both pushing towards the use of increased use of social signals and personalisation within their search results.  They are having to do this through the integration of data from 3rd parties that is being tacked onto already complex technology.  It’s like nailing a dog’s head onto a lion and calling it a bear.

Dog + Lion != Bear

Yahoo had a period of 5 years where they could have created a true social search engine based on technology that they owned, and data that they had control over.  Rather than running into the arms of Bing, things could have been different: Yahoo might have been sufficiently different in 2005/6 to have fought off Google more effectively.