Matt Cutts, well known Google spokesperson, says “One big point of discussion is whether Bing copies Google’s search results … I didn’t expect that Microsoft would deny the claims so strongly”.
Here’s why Microsoft is right to deny them.
There’s no copying
First, Bing isn’t copying Google’s results. It’s not looking at Google’s results for particular searches and then changing the order of its own results to reflect these. This really would be “stealing” Google’s results.
Instead, Bing is using, as one of the signals to decide the order of its own results, which URLs people click on when they search for things on Google. If anything, Bing is in our search result pages stealing our click through data.
Yes, Google has therefore managed to influence Bing’s results by getting its engineers to mass click on odd / “wrong” results for some searches. But Bing’s doing this in line with its T&Cs. And this factor is only influencing its search results in a small number of cases where there are no other signals.
Also, out of 100 tests for made up words (like “delhipublicschool40 chdjob” and “ygyuuttuu hjhhiihhhu”), Google has found just nine instances where its fake-clicking experiments have affected Bing’s results.
In these cases, yes, Bing is clearly using, via its toolbar, click through data – what people click on when they search – as one of its signals. To me, that’s sensible – human behaviour is a better judge of relevancy than stuff like keyword density.
All that Google has shown, therefore, is that, in a small proportion of cases, with made up search terms that no one actually searches for in reality, and when there are no other signals (because the search term is made up), Bing is being influenced by Google’s fake click-through experiments.
There is no sign of cheating here.
(Matt does say in his post that “we thought this practice was happening for lots and lots of different queries, not simply rare queries. To verify our hypothesis, rare queries were the easiest to test.” It would help Google’s case if they showed some real cases of this happening).
Google’s experiment is flawed
Even worse, Matt Cutts is now claiming, in the light of Bing’s response that clickthrough rates are a factor but a tiny one, that:
If clicks on Google really account for only 1/1000th (or some other trivial fraction) of Microsoft’s relevancy, why not just stop using those clicks and reduce the negative coverage and perception of this? And if Microsoft is unwilling to stop incorporating Google’s clicks in Bing’s rankings, doesn’t that argue that Google’s clicks account for much more than 1/1000th of Bing’s rankings?
This is highly misleading.
For a start, there’s only negative perception because Google started this row.
Second, the fact that a factor has a weighting of 1/1000th in its algorithm is irrelevant.
If the other 999 factors are missing (as they are likely to be for made up search terms like “delhipublicschool40 chdjob” that no one actually searches for) then Bing is only left with the 1/1000 one – which becomes all important.
Third, searches are not all the same.
Let’s assume Google is clever enough to distinguish informational searches (looking for facts), navigational searches (looking for an entity like a site) and transactional searches (looking to buy). This means that Google gives different weights to different factors in different circumstances, depending on what’s being searched for.
It’s entirely likely therefore that Google itself treats searches for made-up terms that are never actually searched for differently to other ones.
This means Google would do the same as Bing – in the absence of all other signals for an informational search, it would fall back on the only signals it has.
Maybe it’s a bit like when Google gives too much weight to exact match domains (where the words in a domain exactly match the search terms someone typed in). In Matt’s terms, if this factor only accounts for some trivial fraction of Google’s relevancy, why not stop using this factor? Or if Google’s unwilling to stop incorporating the presence of keywords in domains in its rankings, doesn’t this argue that Google places too much weight on keywords in the domain?
Obviously, no (well, actually yes in this Google case – but in Bing’s case, no). it’s just horses for courses.
To sum up
Bing isn’t stealing Google’s results. It’s just taking account of something sensible in working out its results. And what it takes account of – what we click on when we search – belongs to us as users, not Google.
Even when Google has found this makes a difference, it has demonstrated this only in cases involving search terms that no one actually searches for – and even then in only 9 per cent of these made-up searches.
Overall, Google is making a massive fuss over something for its own reasons. Bing has done nothing wrong.