Learn more about your visitors with Internal Site Search

By 21st May 2015December 18th, 2017Digital Marketing
site search detective

Segmentation is awesome. Without segmentation, all you would have is a big pile of aggregate data and there are few data-driven decisions to be made from aggregate data. Segmentation can be applied to any report in Google Analytics and will, most probably, change your world. Using segmentation is like using 4D glasses, it opens up new dimensions! OK, I might be exaggerating a little… but seriously, segmentation will completely alter your approach to web statistics.

Site search is, well, also awesome. If you haven’t turned it on in Google Analytics, then stop everything you’re doing – including reading this – and go switch it on immediately! Done it yet? Good. In a nutshell, the site search informs you of things like what your visitors are searching for, which page(s) they are searching from, if they refine their search, when they leave the site after searching and much more.

Segmentation and Site Search… Two awesomes make for, well a lot of awesomeness!

Site search + segmentation

Two of the most powerful segments that you can create are “Visitors that made a purchase” and “Visitors that didn’t make a purchase”. Using these two together will show you all of your site traffic, segmented between purchasers and non-purchasers. I don’t think I have to explain why this is important…

Using these two segments in the Site Search Report can be a real eye-opener in terms of what the different segments search for and how they behave after they’ve made a search. It provides you that granular insight you need to make data-driven decisions for your site.

Visitors that did and didn’t make a purchase

The below pie charts illustrate a useful and broad overview of search usage. I’ve segmented site searchers into those that “made a purchase” and those that “didn’t make a purchase”. In this specific example we can see from the left hand-side pie chart that out of the visitors that “made a purchase”, 15% searched the site. In the pie chart to the right, out of the visiors that “didn’t make a purchase”, 11% made a site search. These segments are really useful in other Google Analytics report as well e.g. the Cohort Report.

So, what does that tell us? Does it indicate that people who makes a site search are more likely to purchase? Well, in a sense – yes. In general, visitors who make a site search are typically looking for something in specific, and they are also more likely to be further down in the conversion funnel. These visitors would therefore want to find the product they’re looking for as quickly as possible, and also be the most likely to make a purchase. Something to bear in mind while your analysing your site searchers.


Two pie charts illustrating traffic that did a site search


Percent of total unique searches

In the list below, we can see that Query 10 and Query 5 have higher percentage of Total Unique Searches from those that didn’t make a purchase, compared to those that did make a purchase. This indicates that people are searching for this product but, for one reason or another, don’t end up converting. For purposes of this blog, and to further help simplify the example, let’s say that Query 10 relates to search for ‘Shampoo’.


Site search queries


Analyse site search on a specific page

In Google Analytics Site Search Report there is a section called ‘Pages’. From this report, I can see that both the visitors that did and didn’t make a purchase most commonly use the site search from the home page (/).

Google Analytics site search

By clicking through to the report for the home page (/) you will get data specific for that page. From this report it looks like the majority of those that didn’t make a purchase have searched for one specific item – ‘Shampoo’. Interesting… let’s do a little more investigation!

The table below displays the results of applying a filter to only show the ‘Shampoo’ related queries (filter the table by using the Advanced Search). Before further examining that table though, here’s a quick summary of how we got here:

  • I clicked through to specifically look at the search terms used on the homepage. This page is of interest because it is where most visitors start their search.
  • From this view, I noticed that the visitors that didn’t make a purchase seemed to have searched more frequently for ‘shampoo’ or ‘shampoo’ related queries, than those that actually made a purchase.
  • I then filtered the table to only show ‘Shampoo’ related queries… And here it is:

Filtered search quries in Google Analytics site search report

From this table we can see that over 70% of those that didn’t make a purchase, searched for ‘Shampoo’ related queries. Without worrying too much about the other sections of the table, we can also see that the percentage of search exits and percentage of search refinements are relatively high for those that didn’t make a purchase… A cause for concern.

Data-driven decisions

OK, so what are the data-driven decision to be made from this? And what does this all mean? Well, good question – I’ll give you an example in a few moments. However before I get to that, just bear with me through one more paragraph!

Your data-driven decisions should be made in relation to what your objectives are. You can read further about objectives, strategies and KPIs in this post. E.g. if you don’t want to be even remotely related to ‘Shampoo’, you might as well ignore the above insight. If ‘Shampoo’ is something that you actually provide and want your visitors to buy, here’s an example of what you could do.


From the insight we’ve seen that ‘Shampoo’ is by far the most common search query made by those that didn’t make a purchase. Clearly, visitors are searching for shampoo but for some reason aren’t converting. We also see that a big percentage of those that searched for shampoo either exited the site after their search or refined their search. So, what to do?

Search status – Based on the above, the first thing you should be doing is testing your own site. If you were a visitor to the site looking to buy ‘Shampoo’, ask yourself the following: Are you happy with the search result you get? Do you feel it gives you results that are relevant? Are you able to quickly find the relevant product?

Navigation – Some visitors will search your site because they can’t be bothered going through your navigation. Others will use search because they can’t find what they are looking for. If ‘Shampoo’ isn’t found via the navigation, you should consider adding it under a relevant menu. If ‘Shampoo’ is already part of your navigation, there might be something fishy going on with your search and in which case you should be asking your developer to look into it, immediately!

Wording – Do you use a different name for Shampoo? Head-Soap..? Hair-purifier..? You’ll be surprised how many companies name regular products with a totally strange name because it’s part of their brand identity. Google might understand that queries like ‘head-soap’ is related to shampoo but internal site search engines are, well, not Google.

Visibility – If you use banners on your home page, consider adding a new banner with clear action points. For example, a banner that screams “See our latest Shampoo”, makes is very clear to your visitors that you do actually provide shampoo.

Communication – Do you use newsletters, email campaigns, blog posts, or something of that nature? Make sure to write about ‘Shampoo’, ensuring that you make it clear to your visitors that you stock it.

Take away

Segment your site search and you will gain a whole new world of insight. To simply understand what your visitors are searching for on your site can be powerful. However, to understand the site search behaviour of different segments can be extremely useful… not only work as an idea for blog posts (like mine!) but it can also be used as the basis for site optimisation.

If you have any question about site search, segmentation or any odd Google Analytics question, don’t hesitate to contact us.

Want to learn more about segmentation? This guy has some more insights.