One of the most common reasons to analyze users is to figure out why they do what they do. However, it is also common to see their behaviour as the result of your company's actions, when the truth is: it's the other way around.
I recently saw a graphic explaining that customers are much more likely to buy from companies they follow on Twitter.
I've also heard that users are more likely to continue using a music app if they have created more playlists.
Seems logical, doesn't it?
That's because we're humans, and humans have a hard time, intuitively, with cause and effect. If we're looking for a cause, we will find one, even if it's actually the effect. Even if it's completely unrelated.
In fact, it's just as likely that customers follow a company because they already buy from that company.
It is also just as likely that a user will make more playlists because they've used the music app longer.
"What's the difference," you ask?
If the analysis is correct, and Twitter followers are more likely to buy, it means your product promotions are your most valuable tweets. And your social media messaging is worth a serious investment, because it's working!
However, if people that already buy your product are more likely to be followers, you should be promoting your social media presence on your products, and using Twitter as a way to extend their behaviour in other ways.
Those are very different ways to use your time/money.
If users are staying longer because they have more playlists, you should push them to make more. If your analysis is right, more playlists equals more loyalty.
However, if they are making more playlists because they have been a member longer, it might mean they are bored of the old ones. In that case, the old ones "don't count" in their minds, and pushing them to make more playlists actually "burns" their experience faster (the playlists are lower quality).
One of those analyses says you should push for more playlists and ignore other features. The other says that loyalty is caused by other features, and you should encourage people to be more thoughtful with playlists.
It is very appealing to believe that our work on Twitter caused non-customers to become customers. So simple, yet so effective!
It is also very appealing to believe that the feature we created is so good that the more people use it, the more they love us!
The problem comes from the classic problem of treating correlation as a cause & effect. Just because two things happen together, doesn't mean one effects the other.
Our intuition doesn't like that, but it's true.
It is depressing to think that everything we do on Twitter is preaching to the choir. Are we just tweeting a lot of noise to people who would buy our shit anyway?
And if playlists are just a side-effect of existing loyalty, then what causes the loyalty? Does that mean we have no idea what people like?! Maybe.
Correlations need to be deconstructed and tested. Until you have proof that Twitter causes purchases or playlists cause loyalty, you don't know. You merely believe that is the case, and you have to admit to yourself that you have an incentive to believe it, even if it's false.
If you're wrong, it will be a very expensive mistake.