← Back to the problem

The builder's path

Five steps from a problem to its solution.

For: Metabase admins can't tell whether an alert was skipped on purpose or broke.

“Some things are under our control, while others are not under our control.”
Epictetus · Enchiridion 1 (tr. W. A. Oldfather)

Routes to a solution

Six ways to take this forward

Each route is a different kind of journey, with real people and organisations to contact at each step, starting with those in every country.

  1. Step 1

    Build in the open

    Publish the code and invite contributors from the community.

  2. Step 2

    Fund the maintainers

    Sponsors, collectives and open source grants.

  3. Step 3

    Get it adopted

    Civic tech groups and public bodies take it up.

What this problem needs

Ideas, sources and tools for this problem, from first interviews to funding.

  • Open data
  • Alerts
  • Analytics

Ideas to explore

  • Idea 1New status valueRecord a ‘skipped: no results’ status where the send is cancelled, and show it in the admin history with a short explanation.
  • Idea 2Skipped-count summaryAdd a column showing how many times in the last 30 days each alert was skipped for no results.

Specific to this problem

Talk to first

  • Metabase ↗

    the team that tagged the issue and reviews contributions

Starter prompt

Built from this problem's brief. Paste it into Claude or any AI assistant.

Help me design and build a first working prototype for this problem in Anywhere.

Problem: Metabase admins can't tell whether an alert was skipped on purpose or broke.
Who it hurts: Admins of Metabase, an open-source business intelligence tool, when someone asks why their alert never arrived.
The moment it hurts: When a user reports a missing alert and the admin sees only ‘Success’ in the history.
What people do today: Admins check Slack, the alert settings or the dashboard behind it, even when nothing is broken.
Why it is still unsolved: The status is technically correct, since the alert did run, so it has not been treated as a bug.
What solved looks like: The admin task history shows three distinct states, delivered, skipped because there were no results, and failed, with an automated test for each.
Where: Anywhere

First, list what you would need to learn from the people affected before building, as interview questions.
Then propose the simplest version that would help, the screens or messages it needs, and the data it would use.
Keep it cheap to run, accessible, and respectful of people's privacy. Flag anything you are unsure about instead of guessing.
  1. Understand it

    Talk to the people who live it before you build.

    “Arise, awake; seek out the wise and learn.”

    उत्तिष्ठत जाग्रत प्राप्य वरान्निबोधत

    uttiṣṭhata jāgrata prāpya varān nibodhata

    Kaṭha Upaniṣad · 1.3.14
  2. Prototype with AI

    Get a working version in front of people fast.

    Using AI is not cheating. Every tool carries the work of the people who made it, and using it well is how knowledge has always moved forward. The thinking, the care for the people who live the problem, and the judgement are still yours.

    “The Analytical Engine weaves algebraical patterns just as the Jacquard-loom weaves flowers and leaves.”
    Ada Lovelace · Notes on Menabrea's Sketch of the Analytical Engine, Note A (1843)
  3. Find the data

    Open data that fits this problem.

    “You cannot step twice into the same river.”
    Heraclitus · Fragment B91, via Plato's Cratylus
  4. Build it

    Tools matched to what this problem needs.

    “That which is an obstacle on the road helps us on this road.”
    Marcus Aurelius · Meditations 5.20 (tr. G. Long)
  5. Launch and get support

    Ship it responsibly, and find help and funding.

    “While we teach, we learn.”

    Homines dum docent discunt.

    Seneca · Letters to Lucilius 7.8

Kit written for this problem from its brief; every link was checked. Links go to each provider's own site. Zetema is not paid for any listing.