
In the social sector the collection of data has traditionally been about compliance with funder requirements. Often staff and clients are collecting lots of data but they don’t get to see or analyse it - it all goes into a central database seen only by the data experts, the funders or the senior managers, and not by the people actually delivering the services or the clients themselves.
At its best, data should be used to help people across an organisation to make better decisions. This is the world of ‘business intelligence’ where data is used and analysed as it is generated without needing to wait for long term program evaluations. It is all about processing data to create reliable, accurate and timely visualisations that are used for practical decisions. At the frontline it can be powerful to have data and visualisations that help a case worker or intake worker to better understand needs in context, and to plan services or manage waitlists, among other things.
While data dashboards often start with needs from the Executive team or senior managers, they shouldn't stop there. Funders in the social sector have so much power, their data needs often dominate the data systems of social organisations. But even if you collect data for funders, you can still turn it into useful tools for the frontline. Data should be about empowerment of your staff, helping them understand the big picture and the small nuanced picture of individual clients, and enabling better decisions. Here are 5 actions you can take to better empower the frontline with data:
1. Conference with staff about their needs and frustrations
Good data system design requires a clear view of how data will be used at the end of the process. Get buy-in by trimming the number of metrics collected to just what’s useful, and streamline the process of collection. Don’t just rely on the quality managers or ‘experts’ - find out what staff need at their fingertips.
2. Provide live dashboards that focus on relevant client visualisations -
At intake, teams need live data (even if it was entered only moments ago into a system) that brings various threads together in single charts and dashboards. Enable filters down to service groups or individual clients, and allow for multiple clients or groups to be visualised. Think about the decisions made at each step of a service model and what information is needed to inform that decision.
3. Produce aggregated views of needs-based data
Many programs deliver to groups of people - so data needs to show the aggregate needs and risks across groups of people - this is only possible in a well-designed visualisation. If you are planning for a service to a group of 10 people, the data should highlight the average levels of need, special context or risks across that group, so a worker can understand need and outliers at a glance.
4. Compare clients to averages, historical baselines and performance targets
Dashboards are powerful tools to look at data across time and individuals. It may be useful to understand, say, a level of need from a survey tool as against the average for all clients so that a case worker can understand needs relative to others they have served. Data helps understand what the overall group is like but also if one or two people don't fit the group. This isn’t possible just looking at the individual data in the database.
5. Use data for service planning for each client
Good data can help identify risks and need levels - not every client needs the same intensity of service. Well designed triage rules can help focus time and effort to the right levels, enabling staff to serve more people in the same amount of time. With enough data you can set expected levels and targets to make it easier to manage performance.
Building live dashboards does require some technical expertise especially to link legacy data to new client survey data, and to ensure the backend works seamlessly. But it also requires a good understanding of how metrics can be used, calculated and visualised, and how data is applied in a practical environment. For most social organisations, the help of a data specialist can get the system designed and empower staff to deliver good results.
Once these initial steps are underway, and you have a good flow of the right data, you may be ready for a next step which is to do ‘deep dive’ analytics to answer more complex questions and even venture into the use of machine learning to predict risk or program outcome, but that’s a topic for another time.
Comentários