How do you test and measure, if you can’t measure?


As you may know one of my mantras is ‘test and measure‘. That is, always be testing things/ideas/tactics and measuring the results.

Which is all fine and good if you can actually measure the results. But what should you do in situations where you can’t reliably measure? And how do you even know?

Take for example the recent confirmations about Facebook inflating video metrics. This all started in a lawsuit against Facebook a few years ago, however it is only recently that the extent of the deception has been confirmed.

Summary of the issue: how they calculated time spent watching (total time divided by ‘view’ instead of total time divided by ‘started watching’)

Eg if total time watched is 3 mins, and 3 people start watching, but only 1 actually views for more than 3 seconds, then there is a big difference

By one calc the average watch time is 3 mins, with the other it is 1 min, ie a 300% difference

How to accurately test and measure

So what can you do in this instance?

The key here is to push your measurement to be more than just engagement – aim to get visits and conversions as well. Engagement has always been a bit of a fluid metric, especially when it means different things on different platforms (eg a video view means something different on Facebook to YouTube to Twitter), so instead aim to measure reliable metrics: visits to your site, signups on forms, purchases in your store.

It’s also a reminder that using multiple analytics packages is fine – they won’t agree exactly, but they should be within 10% of each other. Eg using both HubSpot analytics and Google analytics on your site.

Add comment

By Craig Bailey