Here's the first part of that code in a copyable format:

How Do You Make Your Data Talkor Sing?

by | Jun 23, 2023 | Uncategorized

Of late, real-life experiences have inspired me to write about business issues. One such was a recent comment made by a client. “I’m sure we can find some data to support that approach,” he said. On face value, this seemed like a harmless comment, until I thought about it a bit more.

We use data in many ways. We use it to evaluate, validate, understand, and we use it to make confident, fact-based decisions. All thanks to Edward Deming, the father of Six Sigma and business process improvement. Twenty-five years ago, he made a statement that’s just as true today: In God we trust; all others must bring data. This is what has made data synonymous in the business world with “evidence-based” and “fact-based” decision making.

How this data is used and what the data can inform are becoming critical drivers to success or huge accelerators of failure. To get to the right answers, we need to make the data talk. I’ve seen folks do it one of three ways:

Marketing and data expert Terry Hunt, author of Scoring Points, once said that the more data you have, the more general the answers will be. If you have data on three very different people, you can see their uniqueness. Say you have three people in a room: one older south-Asian guy, a millennial white female, and another woman that’s somewhere between the other two in age. These three represent three distinct groups. However, when you add a million more people on top of these three, the differences vanish. If you look at a cohort that includes everyone who buys your brand and ask, “Hey data, who’s my customer?” It will tell you, “They’re all somewhere between thirty and fifty-five. They make anywhere from $80,000 to $700,000. The majority of them live in suburbia. Etc. Etc.”

Confession is the most fundamental way data is used. It’s also called ‘blue-skying” the data. This is caused by the firm belief in the data world that more is better. Unfortunately, the more data you have, the more general the answers.

Nobel Prize-winning economist Ronald Coase once said, “If you torture the data long enough, it will confess to anything.” If you are like me, you have been in countless meetings where you’ve seen how the same data is used to support different points-of-view. This is why I love the torture analogy. It is a well-established fact in law enforcement that answers produced under torture are not reliable.

Tortured data, in my experience, is frequently used to support decisions and actions that need quick action. Like in the comment that inspired this post.

The best way to get to the right answer from your data is to interrogate it. Interrogative approaches allow you to begin with an outcome in mind. They allow you to find the narrative from the data. Here is the approach that we use: Start by asking the questions where you already know the answers (this helps level-set for baseline credibility). Ask new questions in incremental steps. Use each answer to refine each area of inquiry. Use my favorite – the W4H (why, who, what, when, how) questioning approach. And keep evolving the questioning tree.

The effectiveness of this approach is enhanced significantly if you have deep, fundamental knowledge of and empathy with your customer. Interrogations are valuable when you want your data to be generative…inspire ideas and new directions.

As Alvin Toffler, futurist and prognosticator, once said: you can collect all the quantitative data you like, but you still need to deconstruct it and use your own judgment to make it make sense. He had a very good reason for this caution – when we ask more specific questions, consumers will be more specific in their answers. To the point where their answers will have more to do with the questions asked than about their actual behavior. The data no longer provides a window into people’s true motivations and desires, the true purpose of any investigation.

So, what’s the net?

Data confessions will deliver generic answers. Tortured data will produce unreliable answers. Careful interrogation of your data, however, will provide actionable and insightful answers.

Might have you singing all the way to the bank.