If you're mining data for your company you need to know about the toddler effect.
So what is the toddler effect?
It is the fact that someone else, other than the intended user, is generating your data part of the time which results in outliers which are masked due to the sheer frequency of outliers.
Still not with me?
Well, the effect is called the Toddler Effect (TE) for a reason.
A perfect example is Ryan's Toys Reviews. It's a youtube chanel that has 5 million subscribers, but on avarage it has 500 million views per video! How do those numbers combine?!
Toddlers! Unsigned, unsubscribed toddlers who watch and rewatch his videos over and over again on a global scale.. In fact, 12 month old toddlers are just as avid and tech savvy as their parents, when it comes to digital media!
Why does this matter to you?
Let's say you are a tech company, for fun let's say you have a data scientist on board that helps you make descisions such as which services and apps to launch based on your current users. And you have data that looks something like this.
You can take in to account an outlier in the data and still do statistics and create a descision model that makes sense for your buisness, all is well.
But what if your data looked like this. You can't rule out toddlers though restriction in time (e.g. exclude data between time x and y) because toddlers are tiny vampires, they are up at all sorts of hours of the day and they use their tech accordingly.
Parents often lend their children their mobile devices, computers and gaming id's because they are too young to have one on their own, and it's too much of a hassle to create a dummy account. A lot of parents do (the sheer number of video comments saying lasdliuehkeuy on youtube is a witness of it).
The fact that parent lend their accounts to their children creates steady outliers, which most data scientists and analysts can descipher, most of the time, but not always. According to my Spotify data I listen to one song over 100 times in a row. Is it me or my infant? How does Spotify decide? How could anyone; witout doing some form of survey asking the user to specify to what extent they give their toddlers their accounts? So how do you answer if your 500 million views come from parents eager to use your service, or toddlers who are using it as dinner entertainment while their parents are desperatley trying to finish a meal in peace and quiet?
So how do you circumvent this problem? A new generation of children are creating a problem for data scientists that needs to be adressed?
An easy way is to create a toddler firendly version of your service! That allows toddlers to use it, without the additional problem of loosing all their parent's preferences and leaving a trace of strange comments on the internet.
Youtube Kids was a great initiative. But the problem is global. Parents all over the world need similar services, not just parents across the pond! Sure, the toddler effect will still persist in some cases, but rest assure, you'd be solving a problem for both parents and data scientists all over the world-