The hard part is the discipline.

In a world of data abundance limiting yourself, and deciding which data matters is the hard part.

You need to be ruthless in deciding which metrics matter, and collectively build a shared causal model of your business.

xmrit.com/articles/con...

But constraints also help analysts guide their investigations

We saw this benefit recently - Cedric noticed a published article triggered a rise in visitors and unsubscribers, but not subscribers, leading to the discovery of a missing signup box on the site.

xmrit.com/articles/bui...

Leadership teams should focus on 1 and 2, as they are the ones in the end who decide which metrics truly matter, and should be acted on to improve.

Data constraints:

1/ Force you to state which metrics matter and which aren’t important.

2/ Build a causal model of the org, trying to understand which metrics influence another - and by how much

3/ Standardises where to start investigations when something exceptional happens

Businesses are complex, with every tool having a new dashboard to look at, and a new set of metrics to analyse.

You end up drowning under irrelevant data. Constraining how much data you look at paradoxically makes you better at using data.

The trick was although the list seems long it's actually short for two reasons.

1. You spend time on the metrics that show something exceptional worth going deeper on.

2. Every business has an almost infinite number of metrics to look at. Paring down to 400 is the challenge

When I first started joining the Commoncog WBR with @cedricchin.bsky.social the thing that surprised me was how many metrics there were. (>75 at the moment.)

I gave a nervous laugh when Cedric told me Amazon goes through 400+ in under 1h!