Weeknotes: 01/02/19
We’re trying something new: weeknotes. We’re doing a lot of strategic thinking about our organisational Outcomes, and one of the first tasks is how we count people. Here’s some notes from a little while back.
Counting 3 million people
Good Things Foundation are ambitious, and we make our ambition quite plain in our strategy. In 2017, we set ourselves the goal of supporting three million socially excluded people to improve their lives through digital from 2010 to 2020.
This sounds simple, but counting those 3 million people is actually quite complicated. For starters, the number has to meet three conditions. We are talking about individual people who we have supported who are socially excluded.
Like many charities, we run a range of different projects and services, including a number of learning platforms. Most of these are in the UK, but we also run a large programme in Australia through our sister charity in the southern hemisphere. Over the last few years, we have also worked in Ireland, the Philippines and Kenya.
Counting is really the process of understanding and standardising our data. This creates challenges because the data comes from different sources.
The first thing we needed to address was the risk of double counting. This could happen in two main ways.
Double counting 1 — overlap
Some people may have been counted twice because they have been supported through two Good Things Foundation services which collect data in different ways. The most common example of this could be users who engage with the English My Way platform, and later progress on to using the Learn My Way platform for other skills.
Another common overlap could be between:
- Project data we collect through our sites such as Learn My Way registrations
- Wider data, sent to us by external partners we are collaborating with, or from the survey responses we receive from our Online Centres
For the former, we can choose to consider any English My Way learners whose names and Online Centre IDs match a learner on Learn My Way as duplicates, and thus not count them. The latter is much more difficult to avoid, and is something we must continue to think about.
Double counting 2 — duplication
Some people may have been counted twice because they have registered as a beneficiary twice. This is not simple to solve, but a first approach is to deduplicate based on names in our database. I.e, if there are multiple people with the same name registered for Learn My Way at the same Online Centre, we could assume these are duplicates and only count them once. The downside is that we might discount genuinely different people from the same centre who happen to have the same name.
This week
This week we developed the code that brings our different databases together and deduplicates any names that appear more than once at a given Online Centre. As of now, a new distinct user added to any of the six databases will appear as a single entity in the big count. This count will soon be automated and will update on a monthly basis.
We have also connected this count to the other numbers of beneficiaries, who are counted outside of our databases, such as via the Basic Digital Skills Survey, and the equivalents in Australia.
So.
We have made great steps, but there are three key aspects we need to consider throughout the process and discuss with the rest of Good Things Foundation.
Double Counting
How do we address ‘double counting 1’ — specifically, how can we ensure beneficiaries counted outside of our systems aren’t also recorded by our systems?
What kind of risk does it pose and how do we address that risk?
Support
What does ‘supported’ mean. If someone has registered on Learn My Way, used it once and never returned, can we say that they have been supported? If someone in a centre uses our resources but isn’t funded, can we say we have supported them?
Socially excluded
How do we work out who is socially excluded?
