How Ignoring Inactive Time Is Skewing Your Analytics Data

When was the last time you opened a website in a background tab and totally forgot about it, or switched to it checking out some other similar websites first? If you are like me, you probably have 50+ tabs currently opened and you have no idea what some of them are.

As a user, it’s usually not a big deal, but when it comes to analytics tools, the majority of them (including Google Analytics) calculate Time on Site/Page from the point when you opened the website, regardless of whether it was in the background or not.

Look at the following chart to understand why that results in incorrect metrics.

This chart is made using actual data which we recorded from CodeTheory.

Inactive Time indicates either the tab was in the background or the browser window was minimized.

Now let’s take a look at how Google Analytics calculates the actual time spent on site, whereas how it should have been calculated.

How Google Analytics sees this data

Since GA does not know about the inactive time, it sees that as the actual time spent on site. So, it will just add the total time spent and average it out to calculate the time spent on site.

We used 40 latest samples from the same data to calculate the average time and it turned out to be ~201 seconds or 3.35 minutes, using the GA way.

Now let’s see how it compares with the result when we take inactive time into account.

How average time on site should be calculated

When we calculated the average time on site taking inactive time into account, the results were astonishing. Using the same 40 samples, the average time spent on site resulted in ~90 seconds or 1.5 minutes!

This result is less than half of what we got from GA’s approach, which tells us how effective inactive time can be!

Why is this so important?

Now you must be asking yourself why calculating inactive time is so important. The first reason is that many, many of us are now used to tabs when compared to old times (6 years ago?).

According to a 2010 study on tabbed browsing, about 80% of the participants said that they used tabs as a reminder to do something.

Having the tab open is a reminder to me. […] if it’s at the end of the day or lunch time while I am cracking a sandwich or something and I’ll say, ‘Oh yeah, I want to go back and look at that link’ because I see the tab sitting there.

Another example could be searching for something on Google. We normally open several links at once and then visit them one by one until we get what we were looking for.

Consider this case where Amy (an imaginary person) opens the first four links from a Google search results page. She spends 2 minutes on the first one, then switches on to the second one and spends another 3 minutes and so on.

Now according to Google Analytics, the time spent on the first site would be 2 minutes. On the second site, it would be 2 + 3, i.e. 5 minutes. And for the last one, total time spent would be 11 minutes!

Let’s take a look at the third tab. Amy spent only 1 minute on it but according to Google Analytics, she seems to have spent 6 minutes on the same page.

While Amy didn’t find the site much useful, the data in GA would tell otherwise. As the data analyst for the third website, if I looked at that data, I would assume that the page was very engaging, which would obviously be inaccurate.

How does this affect analytics?

I’ll explain this by showing you one more example. Imagine a person, Harvey, who’s a product manager. Harvey uses Google Analytics extensively to analyse how his new products do on launch, and post-launch.

Harvey releases a new product and uses Adsense, Facebook and Twitter to advertise about this new product.

As an average internet user, I was reading a blog post and in the middle of it, the ad of Harvey’s product appeared. Getting excited after reading the pitch, I decide to click on the ad, which takes me to a new tab. Since I was in the middle of the blog post, I leave the tab open and return to finish reading it.

After about 3 minutes, I take a look at that product and within 10 seconds, I leave the page because the actual value proposition didn’t excite me.

Summarising the events happened:

  • A user sees an ad about a product and clicks on it.
  • A new tab opens up taking to the product landing page, but the user decides to look at it later and return to what he was doing.
  • After 3 minutes, the user decides to take a look at the product page, but the low-quality value proposition makes him leave the page within 10 seconds.

Effective time spent on the page: 10 seconds
Total time spent on the page: 3 minutes, 10 seconds.

Now according to this research on web browsing behaviours, 99% of the web pages have a negative ageing effect. It means that the users either leave within 10–20 seconds of visiting a web page or they stay for more than 2 minutes. You can read more about it here.

Coming back to the example, Harvey decides to check GA. He notices that the average time spent on the product’s landing page is more than 2 minutes but the conversation rate is less than 1%.

Looking at both of these metrics, he starts thinking about what could be wrong. If the users are spending that much time, then there might be a problem with either the signup form or the users had a hard time finding the correct information.

If he had known about the inactive time then, he would be able to infer that visitors are leaving within half a minute which might mean that there’s something wrong with the value proposition or maybe the problem is something else, but that would be enough to drive him in the correct direction and take a better decision.

Conclusion

With the rising usage of tabs, the inactive time is becoming more and more significant. Hence, it’s an important metric which should not be ignored because it would be enough to skew your analytics data by a large scale.