Giles Edwards interviewed Andrew Willshire. 1 hour. Listen here. The discussion included data analysis and continuous learning, recency theory, systems thinking, ROI and diminishing returns, and more.
Rough notes jotted while listening. Might not be accurate.
After earning a PhD in Engineering, Andrew Willshire (AW) began working on econometrics and marketing mix modeling. (Logical.)
Pet peeve is “Data are…” Should be “Data is…” (I agree.)
Giles asked about The Long and the Short of It. AW suspects most advertising falls under the category of “boring and effective” and is not entered into consideration for awards, and is thus not in the IPA Databank. So the 60/40 (brand/activation) conclusion is based on incomplete data.
Giles asked about the Facebook / Google duopoly. He noted the two can’t agree on definitions. Semantics may be a root cause of many problems. Question about the murkiness, fraud, and lack of transparency in digital marketing.
AW: Facebook is an egregious offender. They’ll tell you that you can reach twice as many millennials as exist. [Ad Contrarian. Techcrunch.] If one pixel is shown for one second, that counts as a view. [Augustine Fou.] Most Facebook ads are just scrolled past.
Giles: Each device is counted as a separate person.
Discussion about bounce rates. When a person finds what they need and lingers on the page for 5 five minutes, it is a good thing. Someone who bounces in a few seconds is not. The idea that a bounce is bad is simplistic.
AW: Data doesn’t just fall from the sky. It’s an artifact. The objective should be to understand the process. What’s going on? Data is a means to that end. How what we do affects that process.
Giles: Data is no substitute for critical thinking.
Giles: Why do you think long-term effect are overemphasized?
AW: Brand and Activation have lots of definitional issues. Recency theory: Erwin Ephron.
Analytics. Most common mistake is to focus on summary statistics. More interesting is the distribution over time.
Systems thinking. There’s an underlying process that has generated this data. Helps you understand or sends you looking. If I do this, where will that show up in the data. Continuous learning. Try this and go back and see if you’re right. Let’s understand why this failed. We made an underlying assumption that is wrong. Let’s go looking for that.
ROI is a stinking metric. Diminishing returns. It’s an efficiency metric, not an effectiveness metric. More investment will be less efficient after you capture the low-hanging fruit. Tim Ambler paper, “ROI is dead, now bury it.”
YouTube. “If Russell Ackoff gave a TedTalk” (Discusses systems thinking and quality)
How to Measure Anything by Douglas W. Hubbard
Management F-Laws by Russell Ackoff and Herbert Addison
Everything is Obvious by Duncan Watts
The Mind is Flat by Nick Chater