How to Foster a Data-driven Culture

One of the core roadblocks to an effective, data-driven workplace is organisational culture. Teams need to shift from traditional ways of working to adopting a ‘data mindset’.

Thusoo and Sarma describe this in their book ‘Creating a data-driven enterprise with DataOps’. [1]

This shift entails identifying and building a cultural framework that enables all the people involved in a data initiative – from the producers of the data, to the people who build the models, to the people who analyze it, to the employees who use it in their jobs – to collaborate on making data the heart of organizational decision-making. (Ch. 4, para. 4)[1]

If an organisation has a top-down, data-driven culture, supported by technological infrastructure, teams are more able to adopt the data mindset in favour of ‘intuitive’ decision-making.

To embed an organisational-wide data-driven culture, everyone in the organisation must have a clear understanding of the role played by data, and have confidence in its ability to enhance decision-making. People need to feel safe posing solutions based on data, regardless of their position in the organisation. This enables them to use data to start, continue, or conclude every business decision, no matter how major or minor.

Let’s have a closer look at some of the key components that support and foster a data-driven culture.

Have a single source of truth

Bringing your data together and having a single source for metrics and reporting is critical when an organisation is trying to foster a data-driven culture.

Having a central source of truth:

• reduces the time spent searching for data
• reduces the effort required to keep data organised and documented
• makes it easier to automate processes or reporting from a central source
• reduces the risk of bad decisions being made on out-dated, low-quality, or incorrect data
• aligns all parts of the organisation on key metrics, and reduces friction between teams when inconsistent metrics are pulled from multiple systems.

Create a data dictionary

A data dictionary contains a clear set of definitions for the data and metrics used by an organisation. A central list creates common understanding and prevents confusion, especially when under pressure. You don’t need to worry if your ‘sales made’ metric removes refunded items – with a data dictionary it’s easy to check data and metric definitions as you need them.

When you create a data dictionary, you’ll highlight some of the scenarios where confusion can arise. You can condense metrics into a single, valuable measure, or split metrics to capture multiple perspectives.

Make data accessible

If people in your organisation haven’t got access to data, you can’t expect them to make data-driven decisions. Assess the needs of individuals throughout the organisation and ensure the right data is available to each role and business function.

Use dashboards, visualisations, or regular reporting to help teams access and interpret data to perform specific functions – from managing customer support teams to making executive-level investment decisions.

Develop data literacy

Developing data literacy involves training staff in key data skills. You don’t need to create a workforce of data-scientists, but you can lift everybody’s skills to be more data-focused.

Depending on the current literacy in an organisation, this training can cover anything from introductory data analytics (e.g. plotting and interpreting data to avoid making assumptions), to supporting those with existing statistical skills to use advanced analytics tools or programming languages. Conversely, training existing programmers on statistics can be a ‘quick win’ for some organisations.

Data visualisation is another common area for uplifting data literacy. Understanding how to select the right chart type and how to present information in a clear, concise manner helps individuals, and those around them, to interpret data and make better decisions.

Encourage experimentation

The concept of experimentation was popularised in books such as ‘The lean startup’,[2] and the design sprint process from Google Ventures.[3] As an organisation’s data use grows, you can help to cement the importance of data in the decision-making process by encouraging experimentation. This also counteracts the ‘highest paid person’s opinion’ (HiPPO) effect.

Experimentation forces organisations to clearly articulate goals for an initiative or experiment, and to evaluate the outcome against the original goal to determine success. Stating the goals up front, and evaluating performance against them, prevents the cherry-picking of results. Using data to validate experiments depersonalises the decision-making process and helps to create an environment where it is ‘safe to fail’ for the sake of learning and making progress. A / B testing is a common method for experimentation, whether it be for a change to website design or marketing messaging. (In A / B testing you present two options for a single variable and measure their performance with real customers.)

Consider this . . .

We’ve touched on the need to include everyone in the organisation – from front-line staff to senior executives – as part of a data-driven culture.

• Why do you think this is important?
• Where do you sit in your current organisation and how could your role contribute to developing a data-driven culture?

References

1. Thusoo A, Sarma J. Creating a data-driven enterprise with DataOps. O’Reilly; 2017 Mar. Available from: https://www.oreilly.com/library/view/creating-a-data-driven/9781492049227/
2. Ries E. The lean startup. USA: Crown Business; 2011. 336 p.
3. The design sprint [Internet]. Available from: https://www.gv.com/sprint/