What will ultimately replace search engines?
This is a question I am asked frequently, and therefore have given considerable thought to. Ultimately, whatever it is, it has to compensate for the weaknesses of search though right? So, what are the weaknesses of search relative to the ideal search solution (he says knowing that we'll never stop searching for things we need, so search itself will never die, only transform into something more efficient)?
- a) the potential to be gamed by people ... and we all know this is a tremendous issue for search
b) poorer conversion rates than word of mouth, meaning a better solution is possible
c) too many irrelevant results ... people need a means to cut through the 'crap' to reveal the most relevant results
d) local search results are questionable ... too many small local business have no representation, or do not appear at all in localized search results
This then begs the question; what if a solution could be found that would minimize all these weaknesses?
As it happens, I think I have found the solution, or at least what I believe Yahoo is working towards (and if Yahoo is, I'm assuming Google is too) ... integrating word-of-mouth (WOM) into search. Let me elaborate.
Intuitively I've known for sometime that search engines would use what they could glean from social media to help improve their algorithms. For eg.; when Yahoo bought Del.icio.us, I gave them full marks for the purchase because it made perfect sense to me that Yahoo could then integrate bookmarks into their algorithms to reduce reliance on the much abused linking elements. In fact, both Dave Harry (Google Bookmarks the New Ranking Signal?) and Bill Slawski (Google Bookmarks and Personalization) have written about this method of personalization extensively.
What I couldn't fathom though ... was how search engines could possibly integrate word of mouth into their results, to dramatically improve results. Everything found so far in Google, Yahoo, and MSN patents seems to suggest personalization based on your own behaviour. Our behaviour though is often based on the knowledge of friends. As individuals, we are experts at very little, and tend to rely greatly on our networks of friends and acquaintances for advice aka word-of-mouth. This WOM element is completely abscent from present day search results ... and as searchers we're currently forced to gather friends' thoughts and recommendations from elsewhere, and they're certainly not integrated or woven into the algorithms.
Then ... came the AHA moment!
Stumbleupon, Del.icio.us, Google Bookmarks (when shared), and a few others, are forms of word-of-mouth. They enable the passing of information amongst friends ... online.
How Can Search and WOM Be Joined?:
Its simple really! If you begin to look at each of the social media for their strengths, you'll start to get glimpses of the jigsaw puzzle Yahoo is constructing (my understanding of Google's vision is less clear). Stumbleupon and Del.icio.us are essentially opinion sites. You voice your opinions (word of mouth knowledge) about web sites, pages of specific web sites, products, and even services. Thumbs ups and associated comments on Stumbleupon and saves on Del.icio.us could be viewed as recommendations. This is a form of word of mouth!
Now, lets take it to the next level, and start to reorganize the search results based on the cummulative opinions (thumbs ups and downs, and saves) of your "friends" ... as identified by their (ie. the search companies) summarizing your friends on various networks (including Yahoo mail I might add). Relevant pages (as determined by in-page and off-page factors) can then be reordered to show those liked by the most friends first, followed by those favoured by sequentially fewer friends. Those with a negative friend score ... completely buried. In the middle ... all those results not yet scored by friends. Where insufficient friend input exists (eg. how many of your friends will have given a thumbs up to a 2007 Nissan X-Terra?), possibilities include:
- a) searchers seeking out more knowledgeable friends in the space
b) the engines themselves could layer feedback from the masses on top of standard search results
c) the engines can merely revert to the algorithm as it would otherwise exist at that time
This means "your friends opinions have a tremendous impact on the results you see"!
Interesting eh? Might even be a solution to the local results problem search engines face. Will you be happier with the search results? ABSOLUTELY, and if not, merely modify the 'trusted friends' in your networks, and your results will change. In essence;
"you are responsible for the quality of your own search results"!
Now, I can hear a collective "whooooooah" from all readers. How you ask can the major search engines do this if all searchers aren't on Stumble or Del.icio.us? You first need user critical mass on Stumbleupon, Delicious, Google Bookmarks et al for this to work! Well, consider; if you're not on Stumbleupon or Del.icio.us, then the search engines would merely revert to their algorithms they would otherwise be using. How convenient!
It also means much less gaming of search engines, and social media sites. Wow ... a win-win all around!
In the end, its the ultimate solution and a win-win all around:
- 1) Searchers are given information they can make much better decisions with ... when they need it. Searchers become responsible for their own results, and have the power to constantly improve them. Noone else to blame but themselves if they don't like the results.
2) Search engines themselves are happy, because well, searchers are happy. Also, with the integration of WOM into search, search plays an even bigger role in the purchase decision ... meaning more searches, and higher revenues.
While integrating WOM into search would greatly improve search results, it comes with issues as well. So, let me ask:
- a) what do you think the probabilities are of Google, Yahoo, and MSN incorporating WOM into search (as described) in the next 5 years?b) which search company do you personally think is in the best situation to be "first to market" with this type of solution?
c) does it concern you that the search engines will know your 'hot buttons' in terms of understanding which influencers influence you the most?
d) do you think the general search population would be willing to change their bookmarking, commenting, and tagging behaviours sufficiently, to enable this type of user driven search to succeed and prosper?
Special thanks for Dave Harry for bantering around some of these ideas with me.
I'll continue to explore specific aspects of this possibility in future posts, but am really looking for feedback from others in the community.