Return to site

Why all good data analytics should start with people

Emilie Leckie - January 2021

I work with a lot of different organisations helping them to understand, clean, organise and optimise their data. While the output of data, clean or dirty, is a standalone entity, treating it as such will invariably result in poorer outcomes for you clients. There’s a tendency to regard data as something other than human. While it’s true that the human experience is nigh-on impossible to plot on a graph or represent as a set of data points, it’s equally true that organisational data is fundamentally human. People decide how, when and where to collect, store, aggregate and present data. As such, you need to dig down into the human processes and understand the ‘why’ behind data decisions before you open Excel and clean your first cell.

While the process involves a lot of questions around data, these questions are really trying to understand the needs of the organisation. There’s a brilliant opportunity in this type of project to help organisations run more efficiently overall by creating processes that automate as much as possible, while improving the quality of analysis that can be used at the other end.

My starting point is always to ask a client questions like “How do you use this data?" and “How would you like to use this data?”. I also ask clients to describe the story, step-by-step, of how they interact with the data from collection to reporting. My UX compatriots will recognise this as the user journey.

Understanding this process involves asking questions like “is this how you would like to be accessing/storing your data” and “tell me what your ideal ‘dream’ data journey would be”. It will create the basis for how you approach cleaning their data and optimising their processes. I would argue that you shouldn’t begin cleaning a single spreadsheet before you’ve understood the ideal process inputs and data outputs of the whole organisation.