AnalyticsVerifying Business ValueIf (You Write a Good Hypothesis), Then (Good Things Happen)

If (You Write a Good Hypothesis), Then (Good Things Happen)

In order to formulate valuable insights, analysts need to be constantly coming up with hypotheses and testing them to the best of their ability.

While analysts spend enormous amounts of time collecting, cleaning, and managing data, the goal is to come up with insights.

Insights however, never, ever present themselves. One has to dig for them. In mining terms, you don’t just grab a shovel and stick it in the dirt. Instead, you have to have a hypothesis.

If we dig here, then we can find (gold, diamonds, rare earth elements).

So What Makes a Good Hypothesis?

It starts with a question.

How do we increase sales?
How do we get more customers tweeting about us?
How do we increase customer satisfaction?
How do we improve the quality of our leads?

That leads to a guess.

I think this is the reason we’re having trouble.
I think this might be the way to get out of this mess.

That comes from the gut.

An educated guess is a very good thing. If you are a fan of machine learning, imagine 86 billion neurons (your brain) all voting for the most likely/logical/feels right answer. If those 86 billion neurons have years of experience, then you know where that “sixth sense” comes from.

The power of been-there-done-that is huge…and so is thinking outside of the box. But the two together are very powerful. Put experience together with lateral thinking, sprinkle it with data, and you have an unstoppable opportunity to improve things.

That’s easy to understand.

You understand it well enough to explain it to a 10-year-old or a CEO, Same thing.

“If you can’t explain it simply, you don’t understand it well enough.”
– Albert Einstein

That leads to a test.

If we try this we should be able to prove what I’m saying.

That’s straightforward to test.

I didn’t say easy.

That has clear variables and constants.

You know what you are testing and you’ll be able to tell if the test proves true or false. The specific things you are testing are quantifiable and the outcome verifiable by others.

So, validate your data.

Nothing is worse than coming up with a great idea, testing the bejeepers out of it and proving that you are a genius, only to have somebody look at your results and say, “Yeah, but your month-over-month figures are off by a week,” and you have to start over with much less credibility than you had before.

Formalize your hypothesis and your test.

I think that this is a problem. I think this is the reason we’re having this problem. I think this is a good way to test if I am right because of this. This last bit is critical.

“What makes you think that?”
“Oh, I don’t know.”

Ask others for their opinion of your hypothesis and test.

Your thoughts, guesses, insights, and gut feelings are very valuable, but not to the exclusion of others’ thoughts, guesses, insights, and gut feelings.

Getting others’ opinions is not only a great way to check yourself and get great input, it’s also a great way to get buy-in so that your results will be respected and will much more likely lead to positive changes.

Prove yourself wrong.

Many hypotheses are easy to validate, but not prove. Yes, in this example, it all worked out. But what if you had tried it that way? Is it true of all instances?

“If the result confirms the hypothesis, you’ve made a measurement.
If the result is contrary, then you’ve made a discovery.”
-Enrico Fermi

If you can prove yourself wrong, then you can drop that test altogether and get on with your life. Try something else. And test it.

Image via Shutterstock. 


US Mobile Streaming Behavior

Whitepaper | Mobile US Mobile Streaming Behavior

Winning the Data Game: Digital Analytics Tactics for Media Groups

Whitepaper | Actionable Analysis Winning the Data Game: Digital Analytics Tactics for Media Groups

Giving a Voice to Your Brand

Whitepaper | AI & Automation Giving a Voice to Your Brand

Mobile Messaging Masters

Whitepaper | Mobile Mobile Messaging Masters