When not to use analytics

No, it’s not opposite day.

Although we obviously love and advocate for analytics and all of the things they can do for your brand, there are some times where it’s better not to tap into analytics. Let’s break it down.

And as always, if you’ve got questions, comments or concerns you can find us on Twitter @UnionMetrics.

Let’s start with a sports analogy

How do teams do their draft picks? Analytics, of course. But sometimes someone great doesn’t fit the model— like Jeremy Lin, for example.

The Freakonomics podcast asked Daryl Morey, General Manager of the Houston Rockets and one of the first NBA executives to make extensive use of analytics to choose players, about this.

Here’s what he had to say:

“Well, one thing that was tough about Jeremy — because he did actually, produce in college at a level that looked insanely well, meaning if he had played at say Kentucky or Duke or whatever for sure he would have been a top pick in the draft. I have no doubt of that. The problem was he played at Harvard, and actually most of the models that are used from an analytics perspective to forecast draft picks, they’re built on people who are drafted. And Jeremy didn’t look like anyone who was drafted. The number of Ivy League players that have become NBA players is extremely small. So one of the things you have to be careful about with analytics is when to not use things. And I incorrectly chose to not weight his time in the Ivy League high enough, and he ended up going undrafted.”

So what’s the brand takeaway here? If you have an analytic model you’ve worked out for some aspect of your strategy but you come across something great that doesn’t fit it, consider that the model might not perfect and that good things that fall outside of it can be worth pursuing.

The importance of data privacy

With GDPR going into effect across the pond and more GDPR-like laws in the works in the US, data privacy should be top-of-mind for every brand. Brands also need to keep in mind how this affects the analytics that are available— like on Instagram, for example.

Facebook has been through several data privacy scandals and Instagram is cracking down on the amount and type of information available through their API, severely limiting anything considered personally identifiable information. This means some analytics that were previously available no longer are.

If you want to learn more about the specifics around this, you can check out our ungated webinar from earlier this year.

The bottom line

Analytics are great and they can absolutely help you shape a better strategy across social and particularly with your content strategy.

But you shouldn’t dogmatically adhere to a particular model- or even a specific metric- at the expense of the bigger picture. Make sure you have systems in place to check your best measurement intentions.

And as always, we can help with that.

Photo by rawpixel on Unsplash