Data projects: Bridging the gap from theory to practice
Updated: May 30
Many data and reporting projects we encounter have sophisticated strategies and designs for using data in the business. But these have big underlying assumptions. The assumptions are that the data required to fulfil the project is available, clean and fit for purpose.
It can often be easier to shoot for the stars when scoping out a desired and often complex future data-led or reporting state, not constrained by the limitations of the real-world environment. But then coming in with a pragmatic lens on what is achievable can be likened to the impact of popping all the balloons at a kids party.
This can be overcome with investigation and feasibility as part of the design process. Identifying what can be achieved today, what needs some development and what needs significant investment and development to support and deliver the vision.
These are three areas we know help businesses as part of that investigation:
1) Understand your data ecosystem
Start with the data ecosystem - e.g. what data exists today, how can it be accessed, what does it contain, how do the disparate tables link to each other?
In many instances, there is no clear view of the data ecosystem much less any documentation around what is available. You cannot design a data output or report without knowing what sources can be used to populate it. Some of this is a consequence of agile projects where developments move so quickly and there is no time or budget allocated to creating documentation. However, having access to relevant database documentation can save you a lot of time and misunderstanding about the data available in your business. If there is nothing available, spend a bit of time upfront sketching out a conceptual view of the data available as it will save you pains down the road.
2) Understand and validate your data content
Audit what the data contains - e.g. can you identify individuals, what volumes are available, what content is in each field, do you need a lookup to make sense of it, what is the minimum / maximum variable?
Only by understanding the content and distribution of the data you can access can you identify opportunities for how the data can be used to support a data project, a reporting vision, and also identify any risks with the data that will have a negative impact on the project. We identified on a recent project that order data was only held for 6 months and had unusual peaks linked to data load timing, not linked to the purchase date. Only by auditing this did we identify an issue with the source that had to be rectified to support the project or the outputs would have been incorrect.
3) Understand patterns in behaviour to inform the project design
Analyse the patterns in the data to build a baseline understanding of behaviours - design the analysis or report to support specific hypotheses for your vision e.g. how do customers engage with your business, what are the different purchase patterns that may impact the business growth, what makes a customer valuable?
Achieving an insight into the patterns of behaviours within your customer base enables you to design metrics that are most valuable to your business decisions. For instance, if you can identify and report consistently on what impacts a customers value you can put metrics in place to monitor those attributes and design strategies to optimise them.
Starting with these three areas to assess the feasibility of your ambition ensures your sophisticated strategies can be grounded in reality, but with a pragmatic roadmap to support development needs.
We Are MoJo supports businesses on their journey to be data-driven, including investigation and feasibility of data-led initiatives against these three steps outlined above. Get in touch with us today if we can assist your business in its data journey.