Last week I logged into one of my favourite applications, I was presented with a notification that asked me if I wanted to leave a review of the app in the Google Play Store. Without even thinking, I quickly dismissed it by hitting the maybe later button. Immediately afterwards, I started thinking, which is often a dangerous thing, but in this occasion I think we’re OK. I'm a fan of the company, a fan of what they are trying to build and a fan of the application itself. And yet I quickly dismissed the notification. Why? And more importantly, would there be a way to increase the likelihood of my engagement there? If we could increase engagement in this small area, what would this mean for their referrals and their business?
I started thinking through the problem from a behavioural science perspective: what are the factors involved in me forming an intention (to either give a review or not) and then what are the factors involved in me executing on that intention. Then, what product or messaging intervention can we put in place to close the gap between desired and observed behaviour?
Unsurprisingly there are a number of factors involved and more than a few interventions that could be designed and tested - however one in particular seemed to stick out. Humans tend to prefer to act in a way that is consistent with past behaviour - this is a relatively well known behavioural science principle. Rather than immediately asking for a review as part of the notification, could we break it down into two parts, to leverage this principle?
A test would be relatively straightforward to set up too. Let’s imagine two conditions:
Condition 1: the app asks users for a review, as it does today.
Condition 2: break down the request into two constituent parts.
In part 1, we ask a straightforward question:
We're interested in your feedback. How likely would you be to recommend Withings to a friend, from 1-5? (1 being not at all, 5 being highly likely).
Importantly, there is no mention of leaving a review at this stage - this is purely for the app to understand how they are doing. Then, once completed, part 2 is triggered: A follow up message is presented (perhaps only if the first result was positive):
"Thanks for the feedback! Would you like to leave us a review?"
The hypothesis would be that overall, more people would leave a review under condition 2 than condition 1, as a result of a desire to maintain a consistent view of their behaviour. Certainly there is a behavioural science rationale that would indicate this might be a possibility. As with any project like this, it has to be tested in this way - but it is perfectly testable.