Conclusion:

Too often, information communications technology (ICT) and business analytics groups focus on business intelligence and analytics architectures and do not explore the organisational behaviours that are required to take full advantage of such solutions. There is a growing recognition that data literacy (a subset of digital workforce maturity1) is just as important, if not more important, than the solutions being deployed. This is especially true for organisations embracing self-service analytics2.

The trend is to give self-service analytics platforms to management that are making critical business decisions. However, this trend also requires managers to be trained in not just the tools and platforms, but in understanding how to ask meaningful questions, select appropriate data (avoiding bias and cherry-picking), and how to apply the principles of scientific thinking to analysis.

Observations:

With recent advances in Cloud computing for big data, organisations are able to open up business intelligence (BI) and analytics to a broader audience. The terms democratised analytics, citizen analyst and self-service analytics are becoming a fixture of how BI solution vendors promote their solutions and influence business stakeholders. There is now great interest in enabling BI and analytics for non-BI specialists.

However, while the technology is available, data literacy remains a stumbling block to enabling managers to do their open data exploration and make evidence-based business decisions.

IBRS defined data literacy as: “The ability to read, work with, and analyse information in a rigorous and scientific manner, and argue and communicate findings from the exploration of data.

The term argue requires specific attention in the above definition. It requires staff to challenge data and findings, rather than being passive recipients. This is done not to disprove a finding in a black and white manner, but rather to refine and clarify ideas. It requires staff to explore the validity of data, question the causality and correlation of information, and consider if supplemental data may be needed. Of all the aspects of data literacy, the ability to argue in a constructive manner is the most challenging to instil, yet one of the most important to achieving high-quality decision-making based on data analysis.

It is also important to consider data literacy within the context of the Four operating models of Business Intelligence, since these models define the broad intent of using data. The operating models are:

Data literacy and the four operating models of business intelligence

  1. Traditional business reporting: The first characteristic of data literacy is the ability to read data. This is possible with traditional business reporting, which has curated information. At this level, staff can comprehend reports to understand what happened.
  2. Data visualisation: In this operating model, staff can work with information using data visualisation. Staff move from reading and describing historic information to exploring why it may have happened.
  3. Key performance indicators: In this operating model, executives are responsible for not only exploring information, but in determining key performance indicators (KPIs) that will drive change. The creation of KPIs needs to be rigorous, and require the identification of what activities will drive desired business outcomes. This requires a maturing of data literacy, especially in relation to communication, since subordinates will have need for confidence in their leaders’ decisions on what makes a difference to the business. Executives must be able to prepare a compelling case and argue for the KPIs and the design changes in behaviour.
  4. Decision support: In this operating model, decision-making is automated, often by applying machine learning. To be successful, executives must have a high level of trust in information and be capable of identifying potential areas for business improvement using data, forming hypotheses on what is possible, and then testing changes and reviewing the results. At this level, executives must approach data and business improvement in a scientific manner.

Recommendations to build data literacy

1: Start with outcomes, expectations, strategies

To address the data literacy gap, organisations should set clear outcomes and detail how each organisation will achieve those outcomes.

  • Create a roadmap starting with your expected outcomes (the vision).
    • What are the expected benefits of developing data literacy?
    • How will progress towards data literacy be measured?
    • What tools will be provided to staff to allow them to consume and explore data?
    • What training must be put in place to encourage staff to think beyond the tools, to the purpose of why they are exploring data?
    • How will staff understand the underlying semantics in the data available to them?
  • Establish rules to staff for them to make appropriate decisions on the data they use.
  • Put in place technologies (such as enterprise data catalogues) and processes to allow staff to find data to inform their business needs decisions.
  • Allow teams to communicate and collaborate about the data they are using and how they use it.
  • Select a cohort of staff from outside of ICT and the BI groups to learn the basics of data science.

2: Empower fluency in data: Assess, learn, adopt

The first steps in developing data literacy require staff to be familiar with the data available to them and then have skills and knowledge to begin exploring data.

  • Staff learn where the organisation’s data is coming from and are alert to changes in data sources.
  • Staff learn that data is the start of a process, not the end itself.
  • Staff working on data learn that data quality varies and that quality metrics should be considered. They should understand that different levels of data quality are acceptable for different analytical tasks: meaningful business decisions do not have to be based on perfect data, but rather appropriate data.
  • Staff are empowered by new tools and have the confidence and the basic skills to assess, explore, analyse and learn from data.

3: Encourage the ability to challenge assumptions

Once staff develop basic skills and confidence, it is vital that they quickly develop a mindset where they question data analysis – especially their own. Without habitually questioning analysis, organisations run the risk of having staff that use data to simply support a preferred story. It is important to recognise that this way of thinking is not just a work skill, but a life skill. To be effective, people will need to be able to apply this critical way of thinking to their work.

  • Staff learn how to seek insights within data that challenges existing wisdom.
  • Build confidence among staff by eliminating fear and vulnerability among teams. Cultivate a culture where everybody is encouraged to challenge assumptions.
  • Encourage and publicly recognise curiosity, creativity, and critical thinking.
  • People learn how to frame questions (hypothesis) and identify data requirements to test those questions.
  • Demystify data and analytics as intimidating topics through hands-on training and potentially hackathon like events to give staff time to explore what is possible.
  • Senior executives should openly demonstrate that challenging assumptions is not only acceptable, but beneficial.
  • Look for opportunities where previous decisions led to undesirable results and rather than assigning blame, explore the data and interpretations that led to the results, repositioning the event from a failure to an experiment from which we can learn.

Finally, embedding data literacy within an organisation’s culture means embedding continual learning into that culture.

Next Steps:

  • Before investing in new BI and analytics tools, investigate the level of data literacy within the organisation. There is little value in investing in new data analysis solutions if the organisation is unable to leverage benefits from them.
  • Where data literacy is lacking, start by setting a strategy with clear outcomes and expectations as detailed in this paper.
  • Recognise that growing data literacy can be disruptive, especially as staff learn the art of challenging assumptions. This can lead to disputes, where staff bring in opinions and need to be right in the data conversation, as opposed to the ability to argue and communicate ideas. It is vital that senior executives lead by example in this regard.
  • Recognise and reward behaviours of data literacy, not just the outcomes of data analysis.

Footnotes:

  1. Digital Ready Workforce Maturity Model, IBRS, 2020.
  2. Power BI is driving data democratisation: Prepare now, IBRS, 2020.


Joseph Sweeney

About The Advisor

Joseph Sweeney

Dr. Joseph Sweeney is an IBRS advisor specialising in the areas of workforce transformation and the future of work, including; workplace strategies, end-user computing, collaboration, workflow and low code development, data-driven strategies, policy, and organisational cultural change. He is the author of IBRS’s Digital Workspaces methodology. Dr Sweeney has a particular focus on Microsoft, Google, AWS, VMWare, and Citrix. He often assists organisations in rationalising their licensing spend while increasing workforce engagement. He is also deeply engaged in the education sector. Joseph was awarded the University of Newcastle Medal in 2007 for his studies in Education, and his doctorate, granted in 2015, was based on research into Australia’s educational ICT policies for student device deployments.