May 242012

Sebastian Socha shared some interesting images of heat maps for a few different local search result pages. He uses Attention Wizzard for the purposes of the study. It is conducted on and all the keywords used are in German.

The local search results that he analyzes are the following:

1) [Finanzberater Berlin] (financial advisor Berlin)

The SERP includes 3 AdWords ads at the top, (at least) 8 at the sidebar, 2 organic web results (one of which has both ratings bar and a Plus Box associated with it) above the Google Places pack, and a 7-pack of Google Places results.

2) [Kosmetikstudio Frankfurt] (beauty salon Frankfurt)

This SERP is similar to the one above - 3 AdWords ads at the top, 6 at the sidebar, 1 organic web result above the pack, and 7 Google Places results. However, there are two major differences: the upper AdWords results have location extensions, and the web result doesn’t have any specific”items” associated with it.

3) [Optiker Frankfurt] (optician Frankfurt)

There are no paid results associated with this search. 2 organic web results above and (at least) 3 below the 7 Google Places results are the features of this SERP.

4) [Restaurant Kreuzberg] (I think it’s obvious what it means)

Not only there aren’t any AdWords results above the pack, but also there are no organic web results. Interestingly, one of the web results has the bullet snippets associated with it.

5) [Schloberg & Reich] (this is the name of a company)

This is a typical branded SERP. There are 2 sitelinks under the main result and 3 web results with which there are ratings snippets associated.

What do these heat maps tell us?

There are a few very notable trends:

- the map, and especially the part where the red pins are, is very attractive for the users
- the bold keywords also catch the attention easily
- users are interested in the “local” information - address, phone number

Some other things I noticed:

- AdWords results with location extension get more attention than the normal ads
- ratings snippets are not that important if there are many results on the SERP that have them; however, if it is only one result that has them, this result is a clear winner
- Google Places search results that have keywords in the title tag get more attention than these that have just the business name as a title tag
- it doesn’t matter that much if you rank A or G in Google Places

For the one-box branded search the trend is pretty obvious - pictures are the most prominent part of this SERP, followed by the geo-directions (address, transportation). However, these results might differ from business type to business type. In the example above the business is a beauty salon, so it is natural that people would be seeking for the address. If the business was a plumber, or electrician, it is very possible that the phone number would have been the most attractive part of the SERP.

What do these conclusions tell us about how to do local SEO?

- title tag of the landing page associated with the Google Places listing should include location words and keywords as close to the beginning as possible
- meta description of the landing page should also include location keywords
- you should not sweat too much over rankings in Google Places; as long as you are on page 1 that is pretty much enough
- if you really want to “beat” Google Places only with website SEO (if you do not have address in the area you are targeting for instance), make sure that you get ratings snippet, as well as local Plus Box (whoops, this link will go live tomorrow) associated with you web result
- if you are doing AdWords, targeting locally, location extension is a must; location keywords in the title and the text are also very important

  7 Responses to “Heat Map of the Local Search Results on Google”

Comments (7)
  1. I find automatically generated heat maps to be suspect. Here is what Attention Wizzard says about their own results:

    AttentionWizard results are 75%+ correlated with eye tracking and mouse tracking approaches. The algorithm is optimized for fast computation. Slightly higher accuracy could be achieved by more compute-intensive algorithms, but these can take several hours to produce a result.

    While 75% might be fine for a correlation study, since we don’t know which items correlate accurately and which do not, it makes its use for this purpose very suspect.

    • Mike, I agree with you. However, I don’t think such studies should be taken at face value even if the results are delivered by more “compute-intensive algorithms”. The study sample itself is too small to draw any final conclusions. That is why I divided my takeaways in three parts, only the first of which provides more confident points. Everything else is just based on my observation of these 5 heat maps.

    • Hi Guys,

      I do admit, that i’m also somewhat sceptical towards algorithmic heatmaps such as the one i posted today. Real Eye-Tracking is far better.

      Nevertheless i would also draw those three main conclusions Nyagoslav mentioned in his article:

      * The Map is relevant, users do choose businesses by location
      * Wiseley chosen and therefore bolded keywords in page title draw much attention of the user
      * Contact Information is very relevant (especially the phone number)

      That’s not rocket science and one does not necesserally need heatmaps to draw such conclusions, but it’s nice that heatmaps do support this. Local SMBs do still too often forget about basice like that on their websites…


      • Thanks for stepping in, Sebastian.

        I agree with you that these are some trends that are generally true, and one more study, even not perfectly scientifically proven, that supports them is always welcome as a side reference.

  2. Great post and while I agree the data may be hard to take at face value it at least gives some insight and something to think about. In regards to your, “- it doesn’t matter that much if you rank A or G in Google Places” I’d be curious to see how that plays out in more complex studies. And does it depend on the query? Optician and Restaurant seemed to get less attention than the first couple examples. Might be nice to compare some heatmap data if its out there and see what the common behaviors are.

    • Mike, I am not sure about other studies, but I have one which could come to sort of prove the point that it is not that much the rankings that matter. It is a case study and you can see it here.

  3. This is an interesting take at explaining how to perform local SEO, altho I don’t use heat maps. Good old eye tracking methods will do for me. What’s good about heat maps is that you have empirical data.

Sorry, the comment form is closed at this time.