demetrious.co
← All work
Case study · Nonprofit

A membership nonprofit’s scattered data, in one place the staff can use.

A small team, thousands of members across the country, and four years of data nobody could read in one view. I fixed that, and taught Claude what the data means.

Coworkers crowded around a computer monitor studying a dashboard
50states of members
7–8Kmembers
4 yrsof data pulled together
5+data sources, one place
The situation

Too many moving parts for a small staff.

This nonprofit has members in every state, and membership has been moving in the opposite direction it used to as the culture around it changes. They have a lot happening at once:

  • Events in different cities and states
  • Several kinds of membership, with talk of consolidating them
  • Annual and monthly members, with renewals and cancellations
  • Social media campaigns and Meta ads
  • Website traffic and monthly users
  • An aging CRM sitting in the middle of all of it

The goal was to free the staff to do the human work: enrolling people and reaching out to them.

What I did

Three steps.

1

Pull it all into one place

Website, social, ads, membership and event data, going back four years, brought together instead of living in separate tools.

2

Teach Claude what it means

What does each column mean? What counts as a member? How do these things connect? I wrote that down, so the AI reading the data understands the organization.

3

Turn it into answers

Dashboards for the team, plus the ability to ask questions and get answers on demand, and to see what the data predicts.

How it fits together

From scattered sources to answers.

Their data
Scattered across tools
CRMWebsiteSocialAdsEvents
What it means
The context Claude reads: columns, members, how it all connects
What they get
Answers when they ask, and where to focus
DashboardsOn-demand answersPredictions
An illustration of the approach. It is not the client's system or data.
What changed

Intelligence on demand, for a nonprofit.

  • Data on demand. The team can ask a question and get an answer instead of building a report.
  • New angles. Claude can correlate across systems and show overlaps nobody had connected before.
  • Predictions. It points at outcomes to expect and the places the staff should be looking.
  • Time back. Less time inside spreadsheets, more time with members.

It’s the same idea Palantir uses for governments and giant companies, scaled to a nonprofit’s budget. Read how it works.

A note on privacy

I don’t name clients or show their data.

Nothing on this page identifies the organization. Want something like this for yours? The first five clients pay $250.

Book a founding sessionMore work