The Microsoft-Google spat explained
It took me a while to understand the Google-Microsoft spat regarding Bing incorporating Google results into Bing searches. Its important for us because we own a small position in Microsoft and a seven times bigger position in Google.
My first reaction was that Microsoft has engaged in theft. That reaction has softened slightly – but only slightly. I need to walk you through it.
Google's suspicion came from a very strange search: tarsorrhaphy. As Google notes tarsorrhaphy is a surgical procedure on eyelids. They started with an unusual misspelled query [torsorophy]. Google returned the correct spelling—tarsorrhaphy. At that time, Bing had no results for the misspelling. Later Bing started returning Google's first result (a Wikipedia page) to their users without offering the spell correction (Google posts screenshots). This was very strange. How could they return Google's first result to their users without the correct spelling? Had they known the correct spelling, they could have returned several more relevant results for the corrected query.
Google then puts out a veritable honey-pot of strange searches – for example “hiybbprqag” - and they rig their search engine to relate those searches to completely spurious pages. Lo and behond – three weeks later the same searches get linked to the same spurious pages in Bing.
Lets explain what has happened here. Lots of people (including 20 Google engineers) install Bing toolbars on their computers. The Bing toolbars report click-streams to Microsoft. Microsoft thus learns that when people search Google for the odd term “torsorophy” their next click is the Wikipedia article on tarsorrhaphy. Bing does not correct the spelling (it has not got that trick right) – it is just copying the Google users clickstream.
If it copies enough clickstream of course – and especially for rare searches such as torsorophy then it will copy the results.
Microsoft – in their response – claimed that Google used “clickfraud” to rig the results of the test – getting enough Google engineers to install a toolbar that they could rig the Bing results. I want to analyse that response.
Firstly the term "clickfraud" here is (a) emotive and (b) (I suspect knowingly) misused. If I were to put google adverts on this blog and then click those adverts myself so that I got revenue that would be clickfraud. Google has ways of stopping this (with considerable but not total success). But there is no fraud by Google in their click-stream. Microsoft is just saying that clicks resulting from a Google search page are perfectly valid to incorporate in a Bing result.
Bluntly: It's one thing to collect click data, it's another to look at what one search engine returns as a result for a query and add that to your index.
So, what Bing is effectively doing is saying: "we use Google ranking algorithm as one of our signals" but only via the mechanism of the customer's click-stream.
Do they really think this is an honest way of doing business? (If they directly stole the data by entering a search into Google themselves that would be criminal theft. But they get the results anyway via the click-stream.)
The antitrust case
The Microsoft antitrust case argued that Microsoft extended their monopoly in one thing (operating systems) by bundling other things (browsers) into the operating system (and hence killing netscape). I always thought this was a weak case because you could always download a netscape browser if you thought it was better. You can still download a browser based on Firefox from Netscape – and the monopoly in software hardly stopped Firefox and Chrome being the browsers I use (even in those rare times I use a Windows computer).
But the anti-trust case here is absolutely solid. Microsoft bundles the Bing toolbar with some versions of Windows (certainly if you click all the options once you open the included browser you will wind up with a Bing driven machine). It then takes streams that say if someone searches Google for X and then clicks Y then we should copy this and make a Bing search for X present result Y.
They are thus using their power in operating systems to copy (I would say steal) Google's results and hence weaken Google's business.
This is precisely what the antitust case failed (in my view) to prove with browsers.
Google has just asked Microsoft to cease using Google search results to populate Bing searches. Microsoft is pussyfooting around on this.
If they don't cease pussyfooting then methinks it is time to reopen the antitrust case.
John
Postscript: A lot of people seem to have a different view. They argue that Microsoft received consent from the users of the toobar - end of story.
Microsoft did not receive consent from Google.
Its Google's results that Microsoft is copying.
They are not copying them by inserting search terms themselves and copying them. (That would be criminal.)
They are copying them by watching ordinary Americans enter search terms and watching where those people then click.
It is still copying Google's search results.
The consent from the users of the toolbar is a complete red herring.
John