WTF is a Data Strategy?

Posted on 3 mins read

I have seen alot of different documents that have been labeled “Data Strategy” and I have attended a lot of different meetings billed as “Data Strategy Sessions”. Unfortunately I’ve seen much less impact from these documents and session than I would like as it is common for Data Strategies to be:

  • Disconnected from established business strategies & priorities
  • Defined at high levels that are not actionable
  • Incorporate unobtainable, “wish-for” data sets

What is the point of a Data Strategy?

Data Strategies should be relatively boring documents that provide specific direction on what data assets need to be collected in order to achieve desired business outcomes. It is important to note that business strategies dictate the data you need - not the other way around! It is often tempting to say “We have data X - what can we do with it?” but this bottom-up approach has been proven to focus effort on the convenient and not the important.

What is the structure of a Data Strategy?

Generally a Data Strategy should include the following components:

  • Desired Business Outcome/Objective - Your Data Strategy should start with this in order to make sure you are focusing on actual business needs.
  • Use-Case Definition - This provides an outline of how you will deliver your Business Outcome via the use of data.
  • Required Data - This provides details on the type of data, level of granularity and coverage believed necessary to deliver use-case. The executional outcome of the Data Strategy will be to close gaps on acquisition of Required Data.
  • Owner - The person who has accountability for closing Required Data gaps.

Tactics for closing Data Gaps

Once you have identified the Required Data needed to deliver your Business Objectives there may be additional decisions that need to be made on how to source data. I would view this as more of a tactical decision vs. a strategic decision in most cases and should be delegated to the execution team that owns the strategy plank.

Of course, there are some instances where Corporate Data Strategy may impact your specific Business Unit/Team Data Strategy and should thus be incorproated as execution guidance. Specifically, if the company has determined that X data should be sourced as 1st party to drive Y use-case then execution team should at least start with assumption that 1st party data aquisition is the preferred aquisition tactic.

Balancing Transformation vs. Reality

One of the hardest parts of creating a Data Strategy (and any Strategy) is balancing a need to push for constructive disruption vs. pushing so far that developed strategies are un-realistic. If you are limiting Strategy only to data-sets that you have or data-sets that are commoditized it is likely that your strategy will have limited value in driving competitive advantage. However, if your Data Strategy asks for impossible data-sets with zero chance of acquisition you will not achieve anything beyond creating pretty documentation.

To this end I would recommend 80% of your Strategy being “reasonably attainable” and 20% being more “moon-shot” oriented. The Data Strategy should be updated on a Quarterly basis to review progress vs. data aquisition goals and use-case activation and to adjust Strategies if found to be set at too low a bar or if objectives have become determined as impossible.