Microsoft CEO Steve Ballmer has been making funny noises again. He’s been talking about the future of search, and the issues Google faces.

He said“Google does have to be all things to all people. Our search does not need to be all things to all people.”

Maybe so, but it’s hard to know exactly what Microsoft is trying to be, in terms of search. Nevertheless, Ballmer says that the company is experimenting with new business models for search, as well as new ways of presenting search results.

He indicated that Google plays it safe, and is scared of taking risks, or making changes to the way its pages are displayed. I’m sure the Google execs will be sniggering quietly in building 43 at the GooglePlex. From where I’m sitting that is one company I associate with innovation.

Search guru Danny Sullivan concurs: “Google can and does experiment and Microsoft’s search problem ain’t that you don’t take enough risks.”

About five years ago somebody told me that Microsoft wasn’t worried about Google because it was “going long”. The computing giant’s search strategy straddled a 20-year period, by which time it would be a real heavyweight. Maybe so, but what’s happened in the past five years, since I heard that statement? Not very much.

Microsoft’s problem is that it isn’t agile enough. Not that it isn’t smart enough, or big enough. 

But Ballmer has a point. There is a real need for experimentation in the search industry, if it is to evolve and continue to grow. There are some signs that it is becoming tougher for the search engines to maintain growth, and advertisers are finding it difficult to improve ROI from paid search.

What to do?

I think it’s pretty straightforward (although maybe not when it comes to execution!). The next leap the search engineers need to make is to transform search from a 100% keyphrase-driven activity, as it currently is. 

Search has always been about intent. That’s essentially what a search query is: an indicator of intent. You want something, you need something, you mean to purchase something, you’re going to do something. 

But most queries do not reveal the exact nature of intent. There is plenty of grey area, at least until certain trigger words are added to the search term.

When bidding on paid search keywords, retailers look for words that indicate conversion intent, which suggest that consumers are near as dammit ready to buy. For example, we know that adding brand terms to keyphrases indicates that the search journey is nearing the checkout. 

The searcher may start off searching for a ‘digital camera’, then for ‘digital camera reviews’, followed by ‘sony digital cameras’. Ultimately, having browsed around, a product may have been identified, and if it includes a model number then so much the better, eg: ‘sony cybershot black DCS W120’. If words like ‘free shipping’ or ‘special offer’ or ‘next day delivery’ are added to that term, then it is time to ramp up the bidding.

In that instance the search queries have done all the hard work. 

More often than not a consumer will go through various stages in the purchase cycle. There are three distinct phases in my example: subject research, product research, and brand research. That is followed by a further phase: merchant research. After that, the consumer may be ready to buy.

Would it not be better for a search engine to find out which phase the consumer is in, in order to present the most relevant results? This whole purchase cycle might happen over a period of hours, days or weeks, but should a search engine make more effort in tailoring the results and the navigation during the latter part of the journey? Do we always need links to Google News, Images and Videos? Is there a better way of finessing the search experience?

Gathering the facts

The search engines might be able to determine intent automatically (for logged-in or cookied users). The more you know about somebody’s behaviour the easier it is to make a few judgements (this can be gleaned from watching online activity, or even by looking at Gmail / Yahoo Mail / Hotmail). But for me nothing works as well as asking the question, and seeking out some explicit data. Ask the question!

You could argue that search navigation has to some degree tried to do this in the past (Product Search, etc), but maybe needs-based navigation is the way to go? Is it possible to make search a lot more user-centric, rather than purely keyphrase-centric?

Of course it’s a hell of a challenge. We at Econsultancy are faced with a similar issue: you know, what is the best way of presenting all of this information? And if that’s an issue for us then it must be an issue multiplied by a billion for the search engines.

We’ve decided that as well as the structured, topic- and subject-driven approach to navigation, we need a needs-based approach. This might be ‘I’m looking for a supplier’ or it might be ‘I’m an e-commerce manager, please show me what other e-commerce managers have been reading!’. I think it will work well, and we’re focused on metadata to be able to make that happen. 

Essentially we’re looking at a different way of capturing intent, and one that allows us to focus on presenting information in a more personalised way. We’ll see if we can make it work, before releasing it into the wild.

So let’s see what the future for Microsoft is, and search more broadly. All I know is this: if search is to evolve, it has to be about more than keyphrases. Maybe Kumo will provide the answers?

[Image by Nick, ProgrammerMan via Flickr. Various rights reserved.]