What is Microsoft up to in search?

I spent yesterday afternoon at Microsoft's shiny London offices near Victoria (xboxes in reception - gotta love that). I was there for their focus on travel advertising. Since we represent a couple of search clients in both paid and natural search, it was very important for me to learn more about that from Microsoft's perspective (along with most people in the industry, I think, we see some of the highest levels of conversions from Microsoft search - if only there was a little more volume).

Microsoft’s London office

As well as learning lots of things that will help our clients (some of which I've summarised in my update - coming soon), however, I was very interested in the advances in Live search (this is pretty relevant to the advertisers as well because these kind of improvements are going to be crucial in increasing the take-up and market share of the search engine).

The session on what Live is up to was delivered by Cynthia Crossey, the UK director of Live search. She started by telling us why they think there is still an opportunity for Microsoft to catch up and do well in the search space:

- 40% of queries remain unanswered - 50% of queries require refinement

She then made the analogy between where Microsoft was 10 years ago with respect to the database market (with v3 of SQL Server) and where Live Search is now relative to the larger players (Oracle etc. in the database market - Google and Yahoo! in the search engine market). Although it is the head of UK live search saying this, it does seem pretty convincing when they say that they are heavily focussed on search (with basic stuff such as increasing the index size, and targeting their #1 problem of relevancy as well as advanced algorithm updates and cool new stuff in the verticals and video areas).

One thing that I found particularly interesting (given that we only see the traffic numbers from our analytics) is the market share that Live have in terms of users (which, if they can improve the quality of the results, means that they have a much better chance to ensure that these people stick around, rather than just searching for 'google'):

- Live: 9 million users (5.4% queries) - Yahoo!: 7 million users (6.1% queries) - Google: 28 million users (81.2% queries)

The user numbers are much closer than you might expect, and Live is actually ahead of Yahoo!


The areas that Microsoft is investing heavily are:

- Quadrupled the size of the index (to over 20 billion). - Query intent - advanced query analysis - Query refinement (autospell, query suggestions) - More intelligent algorithm (called RankNet) - Extraction of structured information (e.g. product reviews / related people) - 'Rich answers' (e.g. multi-media embedded in the results) - coming soon


The MSN algorithm was always pretty simplistic - and for me, this would have been one of the biggest hurdles to thinking about using Microsoft's search engine in the past. Renowned for their R&D, it isn't that surprising Microsoft is investing heavily in this area. When I was at Cambridge, the science park was just being finished up (with the Gates building, Gates scholarships, etc.). Apparently a lot of what they have been working over there is improvements to search algorithm. There has been a lot of research into neural networks and the way that the human brain works. By incorporating this back into the algorithm, they hope to be able to answer your questions almost before you know you have questions...

There are also a number of basic things that have been improved - such as better stemming, better handling of stop words, and better understanding of query refinements.


They say that mobile Live search is already a good experience - and with Microsoft powering so many of the mobile devices, it seems that there is a very good opportunity for them to at least get Live search in front of a lot of new mobile searchers. If they can deliver as well as they think they can, there is a great opportunity to own the mobile space.

Our customers are starting to talk about mobile as a genuine business platform, so it's pretty crucial for us to stay ahead of the game here.


Already launched in the US, and coming soon in the UK is far more advanced verticals - including rich answers for each of their target verticals, which are:

- local and mapping (they see 32% of queries having local intent) - entertainment - health - shopping


One of the biggest areas of 'R&D'-type investment that Microsoft is making is around video - one thing that sounds particularly cool is some great previewing functionality. One thing they mentioned is automatically pulling out the bits where the audience 'roars' during a sports event in order to do automated highlights. By combining this with functionality whereby mousing-over a video thumbnail plays the preview, it will become orders of magnitude easier to find the video you want to watch.


One of the coolest things that was demoed was the maps functionality (by their own admission, they are far less good at branding than they are at technology - they have a whole load of names for their maps - I think the main name is virtual earth). I have read a little online about how good this is, but I was utterly amazed.

The thing that completely blew me away was what they call "bird's eye". This is high-resolution imagery taken from aeroplane at about 10,000 feet - which means you get multiple angles of real photo imagery at incredibly high zoom levels. I have seen a lot of cool technology, but it really is pretty inspiring when you get up close inside places like Microsoft and see what they are working on. Here is the scale of image they have for tower bridge - near our office.

They also have a similar product to Google Earth at 3dmaptour.com which integrates with user-generated content and lets people publish 3d tours which they can then integrate with some personal imagery and then share with their friends via Youtube.

##The future (something a little different)

Just before the wine tasting part of the evening, we relaxed with a session on the future of digital advertising (which was really more about the future of everything):

By 2020, Microsoft's predictions (some of which are trends, and pretty obvious, but some of which are thought-provoking) are:

- computers with processing power of human brain - broadband as a utility - global network - plug in anywhere you are - autonomous technology (could be a problem e.g. surveillance etc.) - privacy and transparency issues (redefine privacy?) - reduce barriers - both geographical and social - access to network for all - rise of technological refuseniks - people living 'off the grid' - services like postage disappear from high street as no longer profitable - rise of mobile device - continuing globalisation - rise of the huge metropolises - Edinburgh-Glasgow, Leeds-Manchester - at the same time, the nation state becomes less important, and we will become 'more local' - daily newspapers (as physical paper products during the week) will die out - news brands will remain (particularly online) and commentary will become more important - weekend newspapers are more experiential and represent 'me-time' - these have a healthy future - news magazines have uncertain futures as the web becomes the primary place to receive all information - glossy magazines keep the experiential nature and have a better future - TV has a good future - the power to build brands and excite people remain powerful - clips viewed on mobile devices - scheduling pretty much entirely gone - cinema will segment into the really big ticket blockbusters which are great events and the very small, art-house cinemas. The middle will move to home cinema / the web - rise of quality content vs. user-generated content - Youtube will still be around, but many people will seek out high-quality content - teenagers will still trust unknown peers more than authority - grown-ups will trust authority more than peers - demand for open-ness of media - take anything to any device - new teenagers in 2020 will never own a landline or read a daily paper newspaper

Get blog posts via email

About the author
Will Critchlow

Will Critchlow

Will founded Distilled with Duncan in 2005. Since then, he has consulted with some of the world’s largest organisations and most famous websites, spoken at most major industry events and regularly appeared in local and national press. For the...   read more