4 Trends Influencing Data Strategy

The need for a data strategy is not new, however the urgency of having one that is clear and actionable feels like it has increased. Why is this? There are 4 changes that have been happening gradually, and now suddenly (paraphrase credit to Hemingway)

Democratization

Jobs are increasingly based on the interpretation of data. Most importantly, based on the novel interpretation of data. Where the human element isn’t needed, that interpretation can be replaced by an algorithm. New tools allow data to be explored and manipulated in creative ways. Consumption patterns shift continuously as knowledge-workers use these tools to find, explain, and understand anomalies. The workforce is increasingly data literate and capable of working with data without ‘expert’ support. They can connect to different datasets and interpret them without the help of the IT department or dedicated data specialists.

An increasingly data-literate workforce has increasingly open access to data. A successful data strategy accounts for this by addressing the ‘transparency’, ‘accessibility’, and ‘literacy’ aspects of democratization.

Transparency means that tools are in place making easy to find data. There is a catalogue and glossary, along with a certification and/or authorization process for datasets. It is clear which data has its quality actively managed and which does not.

Accessibility means that tools are in place making it easy for knowledge workers to connect to new data environments. If permissions are needed, there are fast-track processes in place both to gain approval and to activate those permissions.

Literacy covers the 3 dimensions of data literacy covered in my post on this topic (link below). Knowledge: Do people knowhow to interpret data? Skill: Do people know how to explore data? Culture: Do people support decisions with data? Change programs tend to focus on ‘skill’ since that is most easily taught, however a successful data strategy that democratizes data needs to consider all 3.

Orchestration

There was a time, a time that many colleagues recall, when the data associated with an application sat in a single database associated with that application. As it became clearer that there was benefit from breaking down these siloes, subsets of data began to be moved around overnight. A knowledge worker had the application associated with their process or function, but now data from other teams. No more. Today, knowledge workers use multiple data tools (a visualization tool for this; a modelling tool for that; Excel is still here).?They connect these tools to different databases at will based on what they are trying to do. On-prem, cloud, SaaS. They don’t care, and similar data elements will reside in multiple places with different context and quality.

In the data multi-verse, data needs to be easily available just in time and fit-for-purpose in a demand-driven environment. A data strategy should understand the use cases for data across the organization and be capable of provisioning data to support them.

This is easier said than done. There is a perception that as long as data is brought together into ‘one place’ that the demand side of data will take care of itself. Aside from this not being true (see: Data Democratization), in the modern world of data engineering this isn’t the goal. The fastest speed-to-value comes from identifying discrete use cases that need to be supported by data, and bringing the data together for each of those use cases as needed. Imagine a customer journey map and identifying the specific touchpoints on that journey that can be enhanced with data and analytics: this represents the data that needs to be brought together “first.” The priority of those touch points represents the roadmap, and by the time “all” of the data you imagined consolidating at the start of your roadmap is brought together, there will be new sources from previously unimagined partners and capabilities that need to be woven into the data fabric as well.

Drive to Market Strategy

Organizations are typically good at managing data that is on the “critical path” to how they make money. Trust me, they are, even if it’s informal: they wouldn’t be good at making money if they weren’t. However, the digitization of everything has led to an accelerated rate of change. The critical path shifts quickly, and the data ecosystem needs to be able to adapt quickly as it does. This means modern, flexible, interconnected data platforms and an agile governance capability that is able to quickly shift its attention as priorities shift.

Data strategy needs to adapt as market strategies change, and they are changing faster all the time. The data strategy needs to have built-in adaptability.

There are two aspects of a data strategy that need this adaptability. One is data management and is about the tools that are in place for provisioning data. It is important that the tools at the foundation of data management be extensible for both adding new data sets and supporting new consumption-layer applications. The data organization itself also needs to be flexible and adaptable. A governance team is typically organized around identified sets of data domains. The relevance of those domains will evolve over time as market strategies change. The data governance organization model needs to provide for leadership looking ahead to what data will be important in the near-term future, and for data stewards to shift their attention (or expand) accordingly.

Risk

Data Governance has moved beyond ‘compliance’ as data has become more relevant to business strategies. A prioritization process is needed to support the allocation of limited resources. This means having a formal data governance structure, with appropriate full and part-time roles clearly delineated in a RACI. It means having defined data domains that are topical rather than based on the physical environment where the data resides (ie, “Customer” is a data domain, one which will have subdomains; the “CRM Database” is not a data domain, but a storage environment that includes some, unlikely all, of the customer data). Emerging privacy laws make it all the more important that personal data be actively managed, here with an eye towards compliance.

Data is an asset that has associated risks that need to inform strategy. Partner with the general counsel’s office and compliance to assess and manage risks.

It’s important that data engineering teams have a solid understanding of how the company views and assesses risk. The legal and compliance groups can help articulate how risk management contributes to the business case for a data strategy.\

Summary

These four trends are increasing the need for organizations to be methodical and flexible in how they think about data management. Understanding how each is relevant to your context can provide valuable inputs to your data strategy.

Related Articles

3 Dimensions of Data Literacy that Support Transformation

Outcomes Driven Data Governance: Focusing DG on Growth

5 Fundamentals for Managing Data as and Enterprise Asset