Alpha This is a new service – your feedback will help us to improve it.

Contributed by


CAST is a charity that helps nonprofits use digital, data and design for social good, through practical learning opportunities and support.

In 2018, CAST did some user research with charities that were starting out with digital. It showed that the individuals involved in this work (‘digital leads’) often felt lonely in their organisations and wanted to bounce ideas off others who were doing similar things. They often struggled to attend traditional networking events and wanted more flexible ways of meeting people.

A trial was set up to connect people in the charity sector working in digital for a chat (inspired by a Nesta experiment). The only thing they needed in common was an interest in digital. Every two months, CAST matched two peers on email and asked them to meet for a coffee: virtually or in real life. (Since the start of the pandemic, it’s virtual-only and will stay this way until further notice.)

The Coffee Connections website

Recipe status

This recipe has been in use since May 2018.

We are not sharing this recipe as the perfect solution to a problem, but we believe CAST’s learnings could be very useful to other organisations.

Users and needs served

  • As a nonprofit digital lead, I need a flexible way to meet others who are in a similar situation to me.
  • As a nonprofit digital lead, I need to learn how others are making progress with digital in their organisations.
  • As a nonprofit digital lead, I need to get external feedback on my ideas about digital.
  • As a nonprofit digital lead, I need to know if others are working on similar projects.

Software and tools used


Typeform is an online tool to collect data, using forms and surveys. CAST uses Typeform for participants to sign up and take part in Coffee Connections.


~£21.50 per month, which becomes cheaper when you pay for a year upfront. They also have a 25-40% charity discount for nonprofit organisations.


Very intuitive tool to set up and get live quickly. Easily customisable to suit your organisation’s branding, with templates you can copy and edit.

It’s worth testing to make sure your Typeform design and set up is accessible, as their forms haven’t been fully tested in some cases (for example, for Typeforms embedded in a web page or chat).


A tool to store and collaboratively edit software and data. CAST uses it to store the user data when people sign up as well as the code that creates the matches between participants.


There is a free option for teams and developers, which includes unlimited repositories and contributors.

If you need more space, there are paid for options:

Team ($4/month/user); Enterprise ($21/month/user).


GitHub can be used for lots of different things, ranging from very technical (writing and merging code) to less technical (it can be used as a project management tool and to store project or product documentation). It can be a bit intimidating to begin with but there are lots of tutorials out there if you’re interested in learning more. That said, it’s worth getting help from a developer to get it up and running.


Directus is a database management tool. It allows non-technical users to manage data without having to access the raw database.


There is a free option for teams and developers, which includes unlimited database tables and records, as well as users - but doesn’t include hosting.

If you need hosting and other features, there are paid for options:

On-demand Cloud ($49 for first six months, then $99 after that); Enterprise Cloud (custom price).


Before using Directus, only the developer at CAST could access the data. To make it easier for other team members to monitor new sign ups and trigger the matching process every two months, they started to use Directus.

Recipe steps

1. Understand your users’ needs

Do some research with your users to understand what they want to get out of connections within the network. It’s worth trying to find out more about:

  • What makes a good match - do they like random matches or to be matched according to specific criteria?
  • Frequency of matches
  • Where the people you’re trying to connect are currently gathering. This is important as there are off-the-shelf matching tools that you might be able to use instead. For example, if everyone in your network is using Slack, you could explore Donut.

2. Collect participant data

CAST used Typeform to collect participant data because it’s intuitive for users. It also has an application programming interface (API), which made it straightforward to send the data to GitHub.

You can start much smaller though: for the first few months of Coffee Connections, CAST put names in a hat and sent match emails manually.

3. Set up the database and matching mechanism

Using the Typeform API, you can make sure data flows from Typeform to GitHub to be stored.

Then you need to write a script to trigger matches.

CAST runs the matches every two months, but you can run them more or less frequently.

People are paired up randomly (avoiding anyone they've previously matched with or who is in the same organisation as them). Each pair gets an email addressed to both participants, introducing them and asking them to get in touch.

Make sure an unsubscribe link is included in the emails.

4. Set up Directus

This allows non-technical staff to edit the data and trigger the matches themselves, without having to go into GitHub.

For example, if someone emails and asks to be removed, this can be done in Directus, by editing the database directly.

It’s worth doing some research to work out if this is the right tool for what you need. Understand what features you want out of a platform so you can select one that is fit for purpose.

5. Train staff

Train relevant team members in how to use Directus and manage users’ data.

6. Monitor progress

Look at sign ups and unsubscribes to monitor success of the initiative.

CAST also has a ‘follow up’ email that can be sent to ask participants whether they’ve managed to meet up with their match.


Be patient - it can take a while to grow numbers.

Consider whether you want the matches to be random or based on particular criteria. Depending on your network, you might match on:

  • location
  • participants’ role
  • type of organisation they work for or any other relevant criteria.

CAST found that matching at random works well, although occasionally heard that some participants would prefer matching to be more specific. It’s worth gathering feedback to monitor this.

Consider the best match frequency for matches to suit your users. CAST initially started off with monthly matches but this was too frequent for users, so it was reduced to every two months.

In the match email, provide guidance on how long to meet for and suggestions for what people might want to talk about during their coffee. Some people get nervous about having a chat with someone they don’t know, so it’s important to make them feel comfortable.

You might need some initial developer support to get this set up on GitHub and Directus.


The connection between Typeform and GitHub could break and go unnoticed. This can be mitigated by ensuring your software is frequently checked by a developer.

Participant data becomes out of date when people move roles. Ask participants to let you know if they get an out of office from their match, so you can remove them from the database and offer a new match if possible.

Points of contact

For further information about this recipe, you can contact:

Tori Head of Digital Practice


Many thanks to CAST for contributing this recipe.


This recipe is licensed under a Attribution-NonCommercial-ShareAlike 4.0 International Licence.

That means you are free to copy, redistribute, and build on the text of this recipe, but only for non-commercial purposes (if you want to use it for commercial purposes, get in touch with us at [email protected]). You must give credit to both Catalyst and CAST and link back to this page. If you build on this recipe then you must share your version under this same licence.

Recipe published on August 24th, 2021. Last updated September 1st, 2021