Sayu Labs
17Feb/110

Do changes in the number of results signify a change in google’s algorithm?

Can changes in the number of results returned following a search on Google tell us anything relevant or be of any practical use to Search Engine Optimisers?

Surprisingly for us we are beginning to think "Yes they can".

The name of the SEO game is improving positions in the SERPS to bring traffic and conversions to a clients site.  As such our SEO team are in the forefront (some might call it the firing line!) if positions move in the wrong direction.

When there is a change the first thing we need to ascertain is whether this change is down to our activities (possibly changes in page content, overly aggressive link building, etc.) or is it some other factor initially out of our control i.e. the rules of the game, as dictated by a search engine algorithm, have changed.

As part of our ongoing research we decided to look at the number of results returned for any given search to see if there is any relationship between this and movements in position.

It was a relatively small scale study where we collected data for about 5,000 searches over a three month period, November 2010 to beginning of February 2011.  We matched the number of results against position and page returned for a number of domains.

We found significant changes in the number of results returned for some keywords.  These happened mainly around 10th December and 20th January (for some searches the change happened a bit earlier, for others a bit later than these dates).

The question is, what has Google changed that gives rise to these sudden significant changes in the number of results returned?

We have seen examples of searches where the number of results increased dramatically as well as searches where the number decreased dramatically.

Maybe there has been a change in when a page is triggered to show in the results or maybe Google has updated the status of pages, removing no longer existing pages, while adding new pages.  If it was the latter however, we would expect a more gradual change as new pages are found and indexed.

When more results are found, you might expect because of the increase in competing pages, that the rankings would drop as new more relevant pages push older ones down, but you wouldn't expect the rankings to improve!  We have seen examples where the position has improved when the number of results increased.

As an example, for the phrase 'ceiling light fittings' the number of results increased from around 50,000 in November to around 65,000 in December.  Despite the 15,000 extra results the page we were tracking saw an improvement from position 100 to 30.  These changes are statistically significant as is clearly illustrated in the charts below.

ceiling lights position vs results

In another case for the phrase 'persian rugs' the number of results dropped from nearly 4,000,000 in late November to 900,000 in December.  Despite this drop in the number of results a page we are tracking has seen a drop in the rankings from position 13 in late November, to position 55 in December.

So if the number of results has changed and despite more results being added you see an improvement in the ranking (or you see a drop in the ranking despite a reduction in the number of results), does that mean there has been a change in the Google Ranking algorithm?

Whilst this doesn't tell us why the pages we are tracking have moved and it cannot be used as a predictor of change (because by the time we see the change it has already happened), it does let us know where to start looking and as such, even at this early stage, has been of benefit to our SEO team.

When caught in the firing line following a movement in position they now have, at least in some cases, a rational explanation as to why change has occurred.  If the movement in position can be linked in a statistically significant way to a change in the number of results returned then this will tell them where they should be looking for reasons why.  A change in the number of results matched by a change in position is more likely to be linked to an algorithm change.

So far we only have looked at the change in positions for a limited number of pages.  To find out more about how changes in the number of results is affecting the rankings we will extend our analysis to the top 30 listings for all searches where we have seen a significant change in the number of results.

Armed with the information as to which url's have improved their position and which have lost out following such changes, we can begin the process of analysing the winners and the losers to ascertain the differences between the two.

We'd love to hear your views so if you've got any comments please let us know.  We'll report back here as soon as we have completed the next stage of our investigations.  Sign up to our mailing list to get notification as soon as we publish.