Latest / MBA Training Data / Data literacy: measure decision quality, not training
Transcript
- 0:00Host: Welcome to the Leaders Insights podcast.
- 0:07Host: Today's episode, Data Literacy Programs that Change Behavior.
- 0:11Host: The concept CDOs get wrong.
- 0:13Host: Most data literacy programs are a waste of money.
- 0:16Host: You've said that out loud in a room full of CDOs.
- 0:20Host: Defend it.
- 0:21Expert: Happy to.
- 0:22Expert: Walk into 90% of these programs and what are they doing?
- 0:25Expert: Teaching people what a median is, how to filter a dashboard, the difference between
- 0:29Expert: a bar chart and a line chart.
- 0:31Expert: Then everyone gets a certificate and goes back to running the business on gut feeling exactly
- 0:36Expert: like they did before.
- 0:38Host: But surely people need to understand the tools before they can use them?
- 0:42Expert: Understanding the tool isn't the goal.
- 0:44Expert: Changing the decision is the goal.
- 0:46Expert: I can teach you what a confidence interval is, the range where the true number probably
- 0:50Expert: sits.
- 0:51Expert: And you'll still approve a marketing budget because someone in the room sounded certain.
- 0:56Expert: Knowledge and behavior are different animals.
- 0:59Expert: Nobody measures the second one.
- 1:01Host: So what does a program that actually changes behavior look like?
- 1:04Expert: It's built around decisions, not concepts.
- 1:07Expert: You take a real recurring decision—pricing, hiring, inventory—and you rewire how it
- 1:13Expert: gets made.
- 1:14Expert: You require the person proposing it to show what evidence would change their mind before
- 1:18Expert: they walk in.
- 1:20Expert: That one question does more than a semester of Excel training.
- 1:23Host: Give me a company that did it right.
- 1:26Expert: Booking.com, comma, years back.
- 1:29Expert: They didn't run a literacy course.
- 1:31Expert: They made experimentation the default.
- 1:33Expert: Every product change had to go through a controlled test, and thousands of people were empowered
- 1:38Expert: to run those tests.
- 1:40Expert: The literacy came from doing, not from a slide deck.
- 1:44Expert: People learned what a P-value was because they needed it to ship, not because HR mandated
- 1:49Expert: it.
- 1:50Host: That's a tech company, though.
- 1:51Host: Easy for them.
- 1:52Host: What about a bank?
- 1:53Host: A manufacturer?
- 1:54Expert: Fair jab.
- 1:56Expert: But the principal travels.
- 1:57Expert: MIT Sloan Management Review has been tracking this.
- 2:00Expert: The organizations that see returns tie learning to the actual workflow, not to a training calendar.
- 2:07Expert: The manufacturer doesn't need everyone to understand regression, fitting a line through
- 2:11Expert: messy data.
- 2:12Expert: They need the plant manager to change one reorder decision based on a forecast and
- 2:17Expert: see it pay off.
- 2:19Expert: Small, concrete, tied to their day.
- 2:22Host: You're describing something slow and unglamorous.
- 2:25Host: Teachers want a number.
- 2:26Host: What number do you give them?
- 2:27Expert: You give them a decision quality number, and yes, it's harder to fake.
- 2:31Expert: Pick 10 recurring decisions before the program.
- 2:34Expert: Track how they were made.
- 2:36Expert: Ciao.
- 2:37Expert: Gut.
- 2:38Expert: Politics.
- 2:39Expert: Or evidence.
- 2:40Expert: Six months later, track again.
- 2:42Expert: DBT Labs published data suggesting teams with mature data practices ship decisions faster,
- 2:48Expert: though they sell data tooling.
- 2:50Expert: So I'd cross-check that against O'Reilly's annual survey work, which is independent and
- 2:55Host: less flattering.
- 2:56Expert: Less flattering how?
- 2:57Expert: O'Reilly's radar research keeps finding the same thing.
- 3:01Expert: The bottleneck isn't skills.
- 3:03Expert: It's culture and trust in the data.
- 3:05Expert: You can train every analyst in the building and it changes nothing if the CFO, the finance
- 3:10Expert: chief, still overrides the model whenever it disagrees with him.
- 3:15Host: The failure is at the top, and no literacy program touches the top.
- 3:19Expert: So the CDO's own boss is the problem.
- 3:21Expert: Constantly.
- 3:22Expert: I've watched chief data officers roll out gorgeous training to 5,000 employees, while
- 3:28Expert: the executive committee makes every real call in a hallway conversation with zero data.
- 3:33Expert: You want to change the culture.
- 3:35Expert: Start the program with the 12 people who actually decide things.
- 3:39Host: Uncomfortable, but that's where the ROI, the return, the money back, lives.
- 3:44Expert: Isn't that just politics dressed up as data?
- 3:47Expert: All of it is politics.
- 3:49Expert: Anyone who tells you data culture is a technical problem is selling you a platform.
- 3:53Expert: The technology's been fine for a decade.
- 3:56Expert: What's broken is that showing evidence and being told you're wrong is socially expensive,
- 4:01Expert: and people avoid it.
- 4:02Host: Fix the social cost, and the literacy follows for free.
- 4:06Expert: One thing a listener can do Monday morning, go.
- 4:09Expert: Pick a decision you make every month.
- 4:11Expert: Before you make the next one, write down what evidence would change your mind.
- 4:16Expert: If the honest answer is, nothing would.
- 4:18Expert: You've just learned you don't need data.
- 4:21Expert: You need to admit you've already decided.
- 4:23Host: That's more literacy than most programs deliver in a year.
- 4:26Host: Sources for today's episode.
- 4:28Host: The new stack, DBT Labs, vendor, data tooling, KD Nuggets, O'Reilly Radar, MIT Sloan Management
- 4:36Host: Review.
- 4:37Host: That's your five.
- 4:38Host: The full CDO track, structure, no fluff, is at MBA-training.com.