Today, being data driven will provide your organisation with an competitive advantage by optimizing its relevant business processes. Moreover, being data driven is actually mandatory in a world where more and more competitors master data themselves. In particular, this not really about whether your organisation has or has not found optimal processes yet, but it is rather a question of how fast it can adapt to a rapidly changing economic environment.
Data Thinking is the process of making an organisation data driven. It is an holistic approach to build and exploit the organisation’s data potential. The outcome of Data Thinking is a data strategy that
Data Thinking can not only be applied once. It can very well be used to improve upon existing data infrastructures. In fact it is best applied as part of an interative lean data process where the biggest “data wins” are identified in a first data thinking step, these wins are then implemented in an MVP (minimum viable product), the resulting KPI are used to improve the measured processes and the feedback from this MVP is used to continue with the next data thinking iteration.
Unlike traditional big data initiatives within organizations, Data thinking will identify actionable data/KPI first before anything is being implemented. Together with a lean data approach it will avoid building complete and extensive data infrastructures that eventually do not provide the right data or are not being used because the thinking within the organisation is not yet data driven.
In order to achieve this, data thinking is modelled after the design thinking process. It is trying to build a data strategy by taking the perspective of the (internal) customer - in this case the stakeholders that are involved in decisions along the processes - and defining KPI which support these decisions. It is a team effort where an internal or external data expert is steering the process and is inviting ideas and feedback from the stakeholders.
The detailed steps of Data Thinking follow the design thinking process very closely and are described in the following.
1. Understand the business model
Which Processes exists in the organisation? How business relevant are they (in terms of success/revenue impact or levers)? Which data is used by which departments. What is the feedback on the current situation by various stakeholders.
This could be a “pull” task, i.e. the expert is requesting feedback from stakeholders, but might also be a first “Analysis Workshop”, in particular if there isn’t a clear picture of business relevant processes in the organisation.
Analyse the available data and compare it to the feedback given by the stakeholders in the previous step.
This is done by the data expert.
3. Define data model
Define a first data model with relevant and/or missing data sources, business relevant KPI, split up by process/department if applicable.
Done by the data expert.
4. Defining and prioritizing first iteration
Presentation of the results from previous steps (preliminary data model). Matching of previous stakeholder feedback to the current and possibly later capabilities of the data stack.
Generate Ideas: What data will enable the organisation to trigger what actions
Feasibility evaluation of ideas by the data expert and prioritization of ideas by the whole team
Done in a “Data Thinking Workshop” by stakeholders and experts
5. Data MVP
As part of a overarching lean data approach a data MVP implementing the prioritized KPI is created. This should be done as lean as possible, still in case of data this isn’t something you can do with pen and paper within the workshop
6. Decisions based on data
Also as part of an overarching lean data approach, before refining and extending the data MVP, the data must be used to steer processes. Therefore the data expert should function as an analyst to the organisation’s team and represent the data side in team decisions. This will be necessary until the teams have adopted data driven thinking and pull data themselves from the analysts or through self-analytics tools.
Please note that the question of whether the steps “prototyping” and “testing” are part of the data thinking itself or part of an overarching lean data approach, is only a matter of taste. I personally prefer the latter where data thinking is only the conceptional part. Also, some people define data thinking as a design thinking process to build data products. In my opinion this would just be the classical design thinking approach. Data Thinking is a process to build a data strategy, not a product.
Data Thinking is related to digitalization and process mining. Digitalization may be a prerequisite for implementing a data stack that measures a process, hence Data Thinking may drive the digitalization of business relevant processes in order to evolve and optimize these processes based on data. Process mining on the other hand is the actual measuring of all necessary data along the process. As such it will be part of a lean data approach if little or no data is available for a process yet.
Data Thinking will help your organization to become data driven and unlock hidden data potentials. At 9 friendly white rabbits we employ this method regularly in our projects. While of course data thinking can be applied to an organization as a whole, it very often is used in the context of a single department or even a single process, for example a sales process or and website customer journey. For smaller projects it may be more “lean” to apply an abbreviated method where the analysis part is done in a “pull” mode by the expert who requests the necessary information from the stakeholders via email or phone calls. If the complexity of the process is small, also the ideation part can be done by the expert and only the the MVP is then tested with the stakeholders, which is then iterated.
The 9 friendly white rabbits help startups and companies with data-driven online marketing & MVP testing for their digital projects.