The Difference Between Truth and Facts Is Vocabulary

A year and a half ago, I wrote “It’s Not an Exact Science” for ClickZ. For those of you who want the “too long; didn’t read” version, it goes like this: Get comfortable with fuzzy numbers.

The Internet may be the most measured, but not the most measurable.

  • Direct mail is much more measurable 
  • Direct sales is much more measurable
  • Telemarketing is much more measurable 
  • Infomercials are much more measurable

We thought:

  • Advertising clicks delineate impression impact
  • Search keywords reveal actual intent
  • On-site behavior exposes individual desires
  • Social media sentiment delivers exact brand sentiment

forecastingWe were wrong.

Advertising and marketing are not exact sciences. Humans are not rational. Analysts are not accountants; we are statisticians. We deal in probabilities and likelihoods. We create models and “All models are wrong, but some are useful.” 

Knowledge, cognition, ingenuity, creativity, and inspiration will always be in demand.

This article came back to me as I was contemplating the next step: So what can an analyst do about it? How do you change your behavior in order to make other people more understanding that we do not deal in facts? We actually do not know how many people actually showed up on our website or Facebook page and never will.

Our job is to get others to understand that we deal in probabilities and not facts. We are not Sergeant Joe Friday from Dragnet. (“Just the facts, ma’am.”)

never-say-certainTo do that, we need to change how we present insights to insights consumers. We should not offer proof. We should not try to win them over to the obvious by displaying the awesomeness of our logic. Instead, bring them along gently.

Insights consumers are pressed for time. They want answers. When we start to explain the unreliability of the data and the importance of the sample size, we lose them. That’s when they tell you to just give them the data and they’ll figure it out.

But we know these decision-makers are so thirsty for knowledge they will disregard our educated approach.

“Conducting data analysis is like drinking a fine wine. It is important to swirl and sniff the wine, to unpack the complex bouquet and to appreciate the experience. Gulping the wine doesn’t work.” – Daniel B. Wright

So here’s my advice: Shift your vocabulary and draw them into the conversation. This second step is critical because they know more about their domain than the analyst. The analyst may know more about the data, but this is where correlation and statistical anomalies meet experience, gut feel, and common sense. Analysts do sometimes come to a conclusion based on the data that is simply a coincidence using a model that was not as useful as hoped.

The domain expert can look at a perfectly logical revelation and counter it with:

  • “Of course movies starting with the letter A are more popular – we list them alphabetically.”
  • “Of course online sales took a jump the week in that region – there was a five-day blizzard.”
  • “Of course we sold more low-end laptops that day – our competitor’s website was down.”

Once you have a clearer understanding of their expertise, shift your vocabulary. Mimic the argot of the weather prognosticator who talks about a chance of showers. Use the vernacular or the gambler running the odds. Follow the lead of doctors who talk about relative health risks.

probability-lineDon’t let the business side of the house even think that you are giving them the absolute correct answer to a formula. Frame your insights in the language of probability:

  • The data suggests…
  • It seems more likely…
  • One could conclude …
  • Based on the data, it feels like…
  • If I were placing bets after seeing this …

Image Source: Math Is Fun

And then, draw them into the supposition process.

  • Doesn’t that seem logical?
  • Does that meet or challenge your thoughts?
  • Do you think it means this or that?

It shouldn’t take long to get them to see you as an advisor and not a report writer. And remember, the longer they refuse to listen to your informed opinion based on the data, the more likely they probably are to face a higher risk of losing their jobs.

Related reading