How can you improve your staff and partners' data literacy and quality skills?

Let’s face it business users hate having to explain acronyms and basic terminology. Getting common themes and glossaries of terms will pay dividends in getting buy-in from the business. Grab glossaries from the internet and pose the questions ‘is this right, do we agree with this’. You’ll be amazed how quickly business and technology can combine and drive out quick wins. Data quality isn’t something users often grasp, use the 5 whys approach to discussion and probe ‘what happens when this goes wrong, how do you fix it’

Key in getting started with data literacy and quality is identifying what data is most important to the organisation. Not on all levels will the terms "data literacy" and "data quality" resonate with staff and partners. Start by making it concrete: are we speaking of product codes? Supplier names? Financial projections? Use these concrete examples to illustrate the current state and how increased data literacy and quality can help improve existing challenges.

The term "data quality" is one of the most widely debated terms in the field. It means so many different things to so many different people ... all working in the same industry. I'm reminded about the parable of the blind people each touching a different part of the elephant and trying to explain to each other what an elephant is. Each person is correct and incorrect at the same time. Current state assessment of data literacy needs both a qualitative view as well as a defined quantitative view.

Begin by examining the existing data literacy and quality skills among your staff and partners. This can involve surveys, interviews, or assessments to gauge their current understanding, capabilities, and gaps in data handling, analysis, and decision-making. Understanding where they stand is crucial for tailoring the learning plan to meet their specific needs and starting from a point of shared understanding.

Information users often assume someone else has ensured the quality of the data before it gets to them, so a survey asking them questions like this is great for awareness and getting the right people asking questions. The most common issue is inconsistency, where information is named something similar (or occasionally the same) as data that means something else. Alternatively, a data engineer has implemented a different calculation in custom code, rather than using a pre-calculated source.

If you're trying to improve staff data literacy and quality skills, remember that not everyone needs to be at the same level. Just because you're in the weeds of data quality does not mean everyone needs to be in the weeds of data quality. Some people just need to know that someone is validating data quality so the data is ready for them to use when they need it. Some people need to know how to report a data quality problem when they encounter it. Not everyone needs training in how to find and fix data quality issues.

A data quality exercise is often spurred by an event or series of events. The goal could be to re-visit the activity or activities that spurred the data quality review and correction to see if the issue remains. It doesn't always have to be a specific number. The data is unique to the business and so the goals will be unique as well.

One of the reasons data governance initiatives often fail is the lack of a clear pathway for incorporating new data sources into the established practices. The primary data source is typically the users; however, if these users do not possess a well-defined metric to target when introducing new sources, and if that metric is not closely tied to the process of assimilating fresh information into the existing framework—a process that should be an integral aspect of the learning plan—then organizational silos are likely to resurface. This recurrence of silos can gradually erode the very foundations upon which the practice is built.

I agree with all of the above, but I feel its missing a key ingredient! What value does doing these things bring to an organisation? Identifying and realising these qualitative and quantitative benefits is central to any benefits realisation use case.

Vital to translate this into financial and/or customer terms, else the response will be 'So what?' - turn a '20% error reduction' into a '20% cost reduction in re-work' or '10% lift in customer satisfaction scores'

Based on the assessment and established objectives, establish a comprehensive learning plan that addresses the identified needs. This plan should include a combination of training sessions, workshops, and self-paced learning resources that cover topics such as data quality principles, data governance policies, analytics tools, and critical thinking with data. Ensure that the learning materials are engaging and relevant to the learners' roles and daily tasks.

Create systems to actualize data quality quality by checking quality issues based on project goals. These could be standards that need to be followed, how data is sent, time frame sent, what is expected.

Be sure to take include the why of data quality. People who enter/ create the data are often unaware of the consequences of bad data. Make compliance easy by enforcing business data quality rules at the point of entry.

Learning plans need to be relevant to the role of the individual/group. This is straight forward in small to medium organisations with a single/simple mission but will need to be tailored in larger organisations, such as government departments, that often represent multiple "businesses" or outcomes.

Develop the learning plan, ensuring that it is compatible with different learning styles and schedules. This may involve arranging live workshops, providing access to online courses, and creating internal resources such as guides or cheat sheets. Encourage participation by making learning opportunities accessible and demonstrating the value of improved data literacy and quality abilities.

Hopefully, with the right leadership and organisational strategy, implementing the learning plan will be the easy bit. If people are motivated correctly in context, by observing the behaviours of leaders for example, in ways that demonstrate the value of data in their data to day operations, then motivation will naturally follow. Communicating the value and benefits of the learning plan becomes much easier if the context is understood intrinsically.

It's important that as many stakeholders as possible are educated about the negative impacts of poor data quality. Beyond what marketing or sales is seeing and experiencing, the ripple effects need to be communicated to anyone who could contribute data to a database or data source. Instilling a culture of data quality with clearly communicated consequences to the firm and the individual, will help garner more buy in.

Sometimes, Data Quality Initiatives can be seen as a one-off project investment. Ensure it is never set up that way. Maintain the team. Keep sharing with the business. Make it an organisational KPI or OKR. Nudge everyone to keep focus. Embed it in development cycles.

A sustained learning culture is one where individuals compete, not only to excel in their jobs, but to help others. This a facet of high-performing teams. This behaviour must then be present in the context being motivated to achieve data literacy. Of course celebrating and highlighting success based on a data lead culture is part of this but it will come much more naturally in a data-lead organisation.

Start with simple things that help people unfamiliar with data to understand the language, concepts, opportunities, potential "gotchas", and tools available. 1) Begin a list of defined terms with definition, owner, and access site. 2) Gather examples of data or visuals and label related data quality issues (avoiding persons or groups involved). 3) Share examples of solutions developed in-house to avoid or resolve data issues. 4) Encourage an environment of questioning why things are done a certain way offering consistent explanations. 5) Create short demos or links to basic steps to help people be unafraid to use preferred tools. 6) Find non-monetary ways of rewarding contributors. These are just starters.

Consider the costs of not addressing data quality in the business or organization. What is the cost of: a) lost customers b) loss of trust / irrelevancy in your information

Something that worked very well for me was listening to people. Understand their point of view and decipher the impossibility of understanding and generating value from data as a consequence of a reduced vision of reality and lack of confident in data that is produced by different sources. Their vision is reduced to their individual objectives, making their scope smaller and smaller until they establish their silo. When the objectives are well known, we try to obtain the appropriate means to fulfill them and this is achieved by giving them the means such as training and understanding the concept and importance of data quality in order to obtain accurate results.