The Story Your Data Lets You Tell

Key takeaways

• We are built to wrap events in cause and effect, and we will do it even when the events were mostly noise.
• A clean explanation for why a number moved tends to feel like understanding, when often it is just a story that goes down easily.
• Hindsight sharpens the illusion: once you know how things turned out, the causes look obvious and the outcome looks like it was always coming.
• Your data can tell you what happened. The "why" is usually something you supplied rather than something the numbers contained.
• A quick check is to ask whether you could tell an equally convincing opposite story from the same figures.

We convert randomness into cause and effect on contact

Nassim Taleb gave this a name worth keeping: the narrative fallacy, the habit of building a satisfying story on the logic of "it came after, so it was caused by." We do it because we love a cause and will accept a loose association when no real one is on offer. A tossed coin has no memory, yet a run of heads makes us reach for a reason. A flat sales week has no agenda either, and we will not leave it alone.

A tidy explanation feels like knowledge even when it is not

We always tell stories. The trouble is that the good ones feel like analysis. Sales jump eighteen percent in March. By Monday there is a confident account: the new creative landed, the audience matured, the pricing test worked. Each is plausible. Each is repeatable in the meeting. None of them has been tested. And because the story is fluent, and we have now said it out loud a few times, it hardens into the thing we know, in just the way the first post in this series described.

I have done this myself. A homeware client had a wonderful November, and we built a lovely narrative about brand investment finally compounding. It was only in January, looking at the wider category, that we noticed every comparable brand had enjoyed the same November. An early cold snap had pulled the whole market forward. Our brilliant strategy was explaining a bump that would have happened if we had sat on our hands. We were not lying. We were pattern-matching, which is what we are for.

Hindsight rewrites the odds after the result is in

Once you know the outcome, the causes line up obediently behind it. This is why we are quick to blame people for sensible decisions that turned out badly, and quick to credit lucky calls that only look obvious now. The decision and the result are different things, which is the subject of a later post in the series on judging decisions rather than outcomes. For the moment it is enough to notice that the story you tell about a number depends heavily on whether you already know how it ended.

Try telling the opposite story from the same numbers

The test I trust most is also the most irritating to run. Take the explanation you feel sure of and try to argue its reverse from the same data. If conversion rose, build the case that it rose despite your changes, not because of them. If you genuinely cannot construct the opposite story, you may simply be right. More often you will find the data is quieter than your confidence, and that the numbers told you what happened while you supplied the "because".

What I still cannot work out is how to keep this honesty up when a good story is also useful, when the team needs a reason to believe and the tidy version is the one that gets everyone moving. A true "we got lucky" rallies nobody. Perhaps the craft is telling the motivating story out loud and the accurate one to yourself.