A 90-day data literacy programme: from zero to something that works

Before you start

You need three things: a programme lead (not necessarily full-time), a sponsor in senior management who explains why it matters, and two or three pilot teams willing to try.

Days 1–30: measure and prioritise

1. Week 1: define levels by role with a role-based framework.
2. Week 2: prepare a short survey per role, with practical questions. How to do it.
3. Week 3: run the survey and collect results by team.
4. Week 4: pick the pilot teams where the gap is largest and mistakes cost most, and set a measurable goal for each.

Days 31–60: train with real cases

1. Short sessions: 45–60 minutes, once a week, using the team’s own data and reports.
2. One topic per session: reading an indicator, spotting a misleading chart, checking data quality, explaining a conclusion.
3. Practice between sessions: a 15-minute exercise on their real work.
4. Include AI: what data must not go into an AI tool and how to check its answers. This helps you meet Article 4 of the AI Act.

Days 61–90: consolidate

1. Reference tools: a business glossary with definitions of key indicators and, if you have one, the data catalogue.
2. Go-to people: one person per team to turn to with data questions.
3. A second measurement with questions equivalent to the first.
4. A report to management: what improved, what didn’t and how to extend it to other teams.

How to tell if it worked

• The second measurement improves in the capabilities you worked on.
• Repeated questions to the data team about what an indicator means go down.
• Fewer misreadings show up in reports and committees.

Common mistakes

• Starting with the whole company at once, without pilots.
• Theory without your own data: people forget within a week what they don’t practise.
• Not measuring again: without a second measurement you don’t know if it worked.

Frequently asked questions

Do we need to buy a training platform?

Not to start. Internal sessions, your own examples and a simple survey are enough for the pilot. A platform makes sense when you roll it out company-wide.

Who should deliver the training?

Someone who knows the team’s data: an analyst or a data steward, with outside support if needed. Real examples matter more than training experience.

Informational content.