Archive for April, 2007

The first principle of search relevance

Wednesday, April 11th, 2007

Last month, Matthew Berk discussed first principles in an article in Search Engine Watch. Among the examples he gave was this gem:

To deliver value to advertisers, focus first on being relevant to the consumer.

Matthew Berk is a smart man, in good company. He’s agreeing with titans like Dale Carnegie, who, as far back as 1937, wrote about the necessity of being genuinely interested in others:

You can make more friends in two months by becoming interested in other people than you can in two years by trying to get other people interested in you.

Carnegie’s comments fit perfectly with Berk’s principle. If you want consumers to care about you, or your advertisers, you must first care about them. And, as Carnegie stresses in his must-read relationship guidebook ‘How to Win Friends and Influence People’, you must be genuine. If you are faking, people will know.

Human nature is such a funny thing. I recently had the privilege of helping a friend who is just starting out in business.

“Let me pay you,” she said, “I value your time and I don’t expect you to give it away.”

What would your reaction have been? Mine was this: because she valued my time, I was happy to give it to her. Had she been unappreciative, I would have charged her or not gotten involved at all.

Carnegie and Berk are reflecting a near-universal truth of the human condition: when we feel valued, we are more inclined to value others.

So what does all this have to do with search? Berk said it: focus first on being relevant to the consumer.

What is relevance, if not caring? What is relevance, if not a reflection of the user’s needs, wants, and values? Without caring about the user, our search for relevance would fall dramatically short.

When people praise Google, they do so because Google seems to know what they want. When they criticize it, it’s because they’re getting stuff they don’t want. In other words, people praise Google when it’s relevant, and condemn it when it’s not.

Think about this from the user’s perspective. What the user is telling us is this: “I want you, the search engine, to understand me. I want you to know that just because I check my spam folder doesn’t mean I’m interested in recipes for Spam.”

At the same time, we get another message over and over: as much as people want the search engine to know who they are and what matters to them, they don’t want technology to figure it out by spying on them. What is ’spying’? Tracking search history, for example.

These two parameters, ‘understand me personally’ and ‘don’t track my personal behavior’, may seem to be at odds. We propose, however, that truly understanding someone doesn’t come from knowing whether they visited a certain site in the past. The premise of VortexDNA is that we need to get to know you before we can even begin to serve you. We want to comprehend your core purpose and values, your deepest drivers. If we do our job properly, there will never be any need to track search history, because that’s not the basis for our understanding of who you are.

On that note, I’ll invite contact from anyone who cares about personalized search, whether you agree with the VortexDNA premise or not. The truth is, I genuinely care about what you have to say, and it’s not just because it’s a first principle.

It’s because, without you, there’s not much point in what we do.

Google VP evangelizes for relevance

Wednesday, April 4th, 2007

New Zealand recently had the honor of hosting Internet legend Vint Cerf, currently Vice President and Chief Internet Evangelist of Google. While he was here, had a chat with my friend Steve Ballantyne, who in turn obligingly published large chunks of it in the National Business Review.

As the guy who led the development of TCP/IP protocols, and someone without whom the Internet as we know it would not exist, Cerf tends to have a few wise comments about the direction of the web. Here’s one of them:

Now I think Google is very definitely the place to be. Google swept its competitors aside very quickly – I was a big user of AltaVista until the Google page-rank algorithm showed up.

There’s still a problem with finding the information that is relevant. My guess is that the next step in search will require making things more relevant, which may require things like the semantic web that Tim Berners-Lee has been working on.

That was great to hear. It’s great that he’s got such a clear understanding of how the Internet is not yet satisfying all of the needs of its users. Obviously, it’s also great that Cerf is saying the same thing that we’ve been saying here. VortexDNA’s purpose in life is to be a Universal Measure of Relevance (and, yes, fellow Lynne Truss devotees, that title deserves proper noun status).

But it’s particularly great because Cerf is a Google guy, and what he’s really saying is this:

“Google’s page-rank algorithm is good, but it’s not enough.”

Assuming you know a bit about Google’s page-rank algorithm, you could also interpret his comments as follows:

‘Popularity’ does not necessarily equal ‘relevance’.

Ranking pages based on popularity was a phenomenal start. Google set the bar for search—and, in doing so, provided a launch platform to the moon. Now the technology is available to combine Google’s immensely powerful statistical algorithms with a means of ensuring those results are more relevant to the particular individual conducting the query. This collaboration will take us a step closer to Cerf’s vision of relevant search.

We don’t need to reinvent the wheel, but there’s nothing wrong with improving it.