How Search History Influences Google AdWords |

How Search History Influences Google AdWords

By Richard Fergie / April 27, 2009

Last week I asked on Twitter if anyone had any good ideas for a PPC related blog post. @bhartzer replied suggesting I do one about PPC mistakes using his post as inspiration. In the post he shows how many businesses that shouldn’t be were bidding on the keyword “keyword”

I did my own search on the word “keyword” to see if this also occurred in the UK. It does, as you can see from the adverts that were returned. So far, so much a duplication of Bill Hartzer’s post. Then I clicked the Renault advert and was taken straight to the homepage without even being sent to a specific landing page; given the ad text I’d expected to find more information about the £2000 scrap scheme and how to use it to buy a Renault.

Now I was thinking that I’d write a post entitled something like “Renault: Worst… PPC… Ever”. The interesting thing happened when I then tested queries from the SpyFu report. Here are the top three ads for the query “antique dealers”, largely powered by my previous search:

The next query I tried was “campus computer shop” where the results looked much more normal indicating that it might only be the previous search that is used in messing up everyone’s referrer statistics, instead of multiple queries this time.

I ran another test, in fact I ran several; here is one that displays a similar pattern:

  • 1st Search: “Mortgages”
  • 2nd Search: “Cheap mortgages”
  • 3rd Search: “Book shop”

This one is even better:

  1. “Car insurance”
  2. “Cheap car insurance”
  3. “Men’s trousers”

Google Ads for "Men's Trousers"

There were only two clothing retailers on the front page, both below the fold (unless you have a really big screen).

So What Does This Tell Us?

Apart from being a fun way to mess with people’s heads when they look at their analytics this also gives a bit of insight into the Google Money Quality Score/Expanded Broad Match algorithm. The two examples that I’ve found above both have the first search being made in a highly competitive vertical with the followup being in an area with much lower CPCs.

I haven’t told you the details of all the test searches I did to try and find the examples above (I know, I know, the science/study bias police are on their way) but this only seems to occur where there is a big difference in CPC between the two searches.

Is it Good or Bad?

It is good for Google as it increases their expected earnings for what would otherwise have been a low earning search.

I think it is bad for searchers because, apart from displaying irrelevant adverts, it is not adding anything new; they have already had the irrelevant ads displayed to them on their previous searches.

I’m not sure if this is all good or all bad for advertisers. It is bad for the book sellers, clothing retailers and their equivalents because valuable ad space is being taken up by adverts that they cannot compete with without losing money. For advertisers whose ads are being shown for irrelevant keywords I am undecided. It is true that the keyword is not at all relevant to their services but it is important to remember that in PPC advertising keywords are only an approximation that we use; actually we are targeting people. The people the ads are being shown to have expressed an interest in the service so in that respect the advertising in targeted. On the other hand they are not quite so interested right now. Different advertisers will have different ideas about whether this is good or bad.

Google taking previous searches into account when displaying adverts helps explain why travel ppc got a click from the search “shooting holidays USA.” What is the most crazy, off the wall or embarassing example of this that you can find?

More posts from this author

blog comments powered by Disqus
01865 703000