Extracting marketing insights from n0n-marketing data

Business perspective

Our client developed an exercise device which is monitored when in use via a smartphone app. The app gives real time feedback and also sends session data to a cloud-based database. This generated a large amount of data but it was not structured to allow any marketing or business analysis. The client was looking for a licensing or investment partner and needed to unlock the information they had, in order to build a compelling business investment case to the market.

How we helped

We conducted a stock take of available data working with the IT company that developed the client’s app and underlying data base. This began with identifying what variables already existed or could be created from existing data. The next step identified missing information that was needed to achieve the company’s business objectives. The challenge was that the existing data capture was based around the technical needs of the device and providing user feedback during exercise sessions – it required considerable restructuring to be able to analyse user trends.

Real time exercise data (captured multiple times per second) was aggregated based on time stamps into exercise sessions, and sessions were associated with dates. This allowed exercise patterns to be analysed across time. The session data was also restructured to line up all users by day of use (eg sessions on day 1 of using the device, week 1, etc) allowing considerable insights into use.


The report showed patterns of use, from drop off rates (and number of days to drop off) to who was still using the device.  We determined which patterns in initial weeks were most likely to translate into long term consistent use, as well as the most appropriate intervention points to prevent user drop off.

We developed a set of variables to monitor and report through Power BI, along with recommendations for additional data capture. This enabled the business to both measure and influence their customers’ behaviour individually and in aggregate.