Search metrics

Metrics are data points that we use to measure the performance of traffic.

In an attempt to be platform agnostic, the metrics we’ll discuss here talk about what each one is, how you use it on its own and how you use it with other metrics. We’ll also talk about how measuring metrics successfully means you need to take scenarios into consideration.

Click through rate (CTR)

Click through rate is the volume and percentage of people clicking through to a page. A high percentage means that either people have confidence in a search result or popularity of an item in a menu.

CTR can be interpreted in two different ways depending on what you’re measuring.

Search results

CTR in search results gives us an indication of success in matching search term with page. The higher percentage the higher the confidence the visitor has that you have matched their intent with a search result.

If the search result the visitor is clicking on is the one you want them to click on for that specific type of query then a high percentage CTR is a measure of success in matching intent.

However, if the page is not one you want them to be clicking on, a high percentage click through is measuring a failure.

These are the boundaries I use to identify how well a search term is performing in relation to CTR.

100 – 70% – This term is performing really well, you shouldn’t need to do anything with this search term.

70 – 50% – This is performing OK, could definitely do better. However if there is a low volume of search queries being carried out for this term then you can probably just monitor it.

Less than 50% – These searches are performing poorly. If you see a 50% CTR or lower on searches that are in the top 25% of your searches, you should take this as a problem and investigate.

Conversion Rate (CVR)

Conversion rate metric is used to measure the volume and percentage of people who complete an action you have determined to be a conversion compared against the traffic to where that conversion occurs.

A conversion in a lot of cases is when someone completes a transaction. A transaction could be any one of the following:

  • Buying something on an ecommerce site
  • Signing up for a newsletter
  • Downloading something

It is essentially converting an action from a possibility to a reality.

CVG is primarily used as a measure of sales.

It’s quite tricky to use as diagnostic measure as all it does is tell you how many people are actually completing a task against the volume of traffic.

It would be incredibly rare that you have a 100% conversion rate (I’d say impossible, but just because I haven’t experienced it, doesn’t mean it hasn’t happened but it would be a unicorn event).

An additional consideration in conversion is the amount of time it takes to convert someone. This information is a lot more useful for purposes of diagnosing problems.

Measuring conversion on an information only site

If your site is purely information and has very little in the way of conversion events it’s still possible to measure them.

Types of information conversion events are:

  • Clicking on advertising links
  • Clicking on sharing buttons
  • Sign-ing up to get more information
  • Leaving comments on blog posts (although this is more of an engagement metric rather than conversion)
  • Registering for an account or logging in (although logging in is more of an event than conversion)
  • Reading to the bottom of a page
  • Clicking on pagination in an article
  • Clicking on a related link

Search results

A conversion rate metric as part of a search result is usually used to tell you the following:

  • How your search compares to navigation, SEO and organic traffic at converting your visitors
  • If you have problems with the merchandising of your product or service

In some instances, it can also be used to re-rank your search results (this is dependent on what search platform you are using). In this case conversion rate becomes a measure of popularity. For example, if visitors are buying something at a high rate after searching for it, it could be used to boost the result higher.

More like this


Benchmarking is a bit like putting a stake in the