How to Measure Anything

“We can’t measure that.” – You probably know the excuse.

In times of Gen AI hype, it’s more popular than ever: “Productivity growth thanks to Gen AI is there – but not measurable.”

Yeah, right, I’m supposed to believe that now.

As Mr. Demming said:

“In God we trust. All others must bring data.”

Many things are measurable if you invest enough time/money/energy/brain to make them measurable. Sometimes you have to change business processes or even business models. Sometimes rough estimates or proxy variables do the job.

I remember a client, a broadcaster for young people, that published its content exclusively on platforms such as Facebook, Instagram, YouTube and TikTok. However, its shareholders, traditional TV stations, wanted to know the “audience share”, i.e. what percentage of the target group was aware of the youth channel.

My proposed solution: classic market research, i.e. customer surveys. Plus classic branding i.e., a message at the end of each clip: “This is a format from XYZ”. From one (quarterly) survey to the next, the “audience share” quadrupled and their budget grew.

In short: a data strategy is also primarily concerned with the question of how to obtain missing data. A data-driven company is not one that is driven by data, but one that is driven by collecting the right data.

This book, “How to Measure Anything”, focuses on one question: how to get the right data. Despite having been in the data business for more than 20 years, I found this book very inspiring and informative. It also confirmed that we at Datentreiber are on the right track: rather than starting with the question “What data do we have?”, we ask “What data do we need?”