Over the weekend, an interesting story appeared first on  Twitter. That is, apparently, where this sort of story would begin, but first, the context. The Trump campaign had been bragging that a million people had requested tickets to its first rally in months, this one in Tulsa, on Saturday. So many people were expected that the campaign had set up overflow areas, and made plans for the president and vice president to make personal appearances there. 

The actual turnout, as you've probably heard by now, was quite less. There was no need for overflow areas, as only 6,200 people showed up, according to the local fire department. You can debate the reasons why that is (there is still, you remember, a global pandemic), but that's not nearly as interesting to me as the fact the campaign was sure it would have a far larger crowd based on the data.

Except, it turns out, data can be a tricky thing. That's an important lesson because it's also true for your business. We'll get to that in a minute, but first, let's look at what apparently happened in this case. Apparently  TikTok-using teenagers, many of them K-pop fans, registered in droves, creating the impression that far more people were planning to attend. 

It can be hard to write a story about how a presidential campaign got trolled by teenagers on a social media network without getting into politics, but we're going to try. I'm interested mostly in the data because that's what matters to your business. 

Here are three things this example should teach every business about the importance of getting the data right.

Confirmation Bias

No doubt, data tells a story. The problem is that we often look for the story we want to tell, and then find the data that supports that story. This is known as confirmation bias, and it's such a powerful phenomenon that it can happen without our even knowing. In this case, clearly the Trump campaign expected a massive turnout. That's one of its core brand promises--that it draws large crowds. 

Even if the numbers seemed a little out of whack, the campaign expected that a lot of people would want to show up, and the data appeared to support that expectation. As a result, there was no motivation to look more deeply at whether there might be something else going on. I'd argue that a smart leader should always be skeptical of the data, and ask whether there could be another explanation for the story it appears to be telling.

Basic Data Hygiene

Of course, you could argue that, for the campaign, the number of registrations is a win regardless of how many people actually attended. For data-driven campaigns, an email address or phone number is almost as valuable as money for mobilizing supporters, and collecting data on a million people sounds like a huge success.

The problem is, 16-year-olds don't vote. They generally don't make contributions to political parties. The Trump campaign didn't even ask people if they were at least 18. In addition, an individual could easily sign up using a free temporary number from a service like Google Voice.  

What that means is that the data gathered from many of those fake registrations is likely useless to the campaign. Never mind that people shared that if you simply reply to the confirmation text with the word STOP, the campaign has to remove your number anyway.

The lesson here is that data is only as valuable as the process you use to collect it. It might seem counterintuitive, but making it too easy for people to sign up for something doesn't always help because even if you don't get trolled, you could still end up with a lot of people who aren't a good fit, or with bad information.

Unrealistic Expectation

The final, and probably the most important lesson is that the Trump campaign relied on the data to create expectations. In this case, they were completely unrealistic expectations that were based on bad data. Telling people that you expect a million, or even a hundred thousand people, to attend only to fill a few more than 6,000 seats costs you in terms of credibility. 

By the way, try to set aside any feeling you might have about the current president or his credibility because this lesson actually applies to everyone. When you draw conclusions and create expectations based on the data, you should probably be sure that the data is real, and the expectations are realistic. Otherwise, what you thought you might gain in data, you'll find you quickly lost in trust.