Talk to us

LaunchGood: a data team that grew with the platform.

LaunchGood is a global Muslim crowdfunding platform. It had no data team, so data work kept slipping to tomorrow. Here is how that changed, and how donations grew over four Ramadans.

Talk to us

Global Muslim crowdfunding platform

Samad Husain, Datstra’s CEO & Co-founder, was LaunchGood’s first data hire.

Build the foundation first.

  1. 1

    The job was for an analyst. The team needed an engineer.

    Samad applied for a data analyst job. In the interview, he asked about LaunchGood’s data setup. Every answer was no.

    He told his future boss the company needed someone to build the systems first. An analyst would mostly build spreadsheet reports. She agreed.

  2. 2

    Replace spreadsheets with one warehouse.

    Every team worked out of spreadsheets. There was no single view of campaigns, donors or money. The team built the warehouse, the analytics strategy the company never had, and the change management that came with it.

  3. 3

    Grow to five, with a dashboard for every function.

    Samad built the data team. It grew from one hire to five people.

    Finance, marketing, program management and business development each got dashboards. Teams started making decisions from the numbers.

  4. 4

    Anyone can answer a data question in Slack.

    An AI-powered Slack bot lets people get data answers on their own.

Donations grew over four Ramadans.

At the start $254M
After four Ramadans $664M
22% Peak first-time donor retention, 48% above the industry average.
60 seconds Time to answer a data request, down from one to two hours.

Working with Datstra was life changing. They joined our team when we had no data team and data used to be a later ‘to-do’ item for engineering, often getting moved to tomorrow. This led to costly delays in decision making. No one had ever envisioned, designed and implemented an end-to-end data infrastructure at the company even though we already had 1M users. Datstra did that with mastery.

Maria ArshadChief Product Officer, LaunchGood

What an agent would add now.

A clean data foundation makes an agent trustworthy, and LaunchGood already has one.

  1. Finish data requests

    Today the bot answers a question. An agent could also follow up when someone needs more.

  2. Send each team its numbers

    An agent could send each team its figures on a set day, then ask about anything that moved.

  3. Chase missing numbers

    If a team misses its deadline, an agent could send a reminder and tell the data team what is still missing.

Read the Human Development Fund case study AI agents for nonprofits