September 20, 2022

Where To Start With Data Quality?

When it comes to physical assets like equipment and offices, ensuring they are known, well maintained and fit for purpose is a core business function, so why do so many organisations not treat data in the same way?

Whether they realise it or not data is the lifeblood of any medium or large organisation. It drives efficiency, income and customer experience yet is often considered more of a necessary evil than a valuable asset.

The best organisations leverage their assets to fuel decision making at all levels and better understand the people they serve. By treating data strategically, forward-thinking organisations are able to provide both themselves and their customers with better outcomes.

For organisations which haven’t reached that level, now is the time to plan the journey from data rich/information poor, to data-driven. Data quality management is a specialism, but is rooted in common sense. The goal is to get reliable information, to the right people, at the right time; no more, no less.

This could be providing potential customers an up-to-date view of the available stock, recognising and accommodating a customer’s vulnerability, providing a maintenance engineer with the correct address and job details, or providing the senior leadership team with accurate financial information to allow them to set the business strategy. At every level, in every operation, better quality data adds tangible value.

Step 1 – Get the right team in place

Many organisations choose to operate with a Chief Data or Information Officer, others opt for a Head of Data position and some manage change from the COO’s office. Whichever the case, having senior sponsorship of transformational data activity is critical.

From there, experts from both the technical and operational sides of the association need to be brought together. Data quality transformation should be a business-led activity, supported by technologists along the way. By bringing together both sides early in the initiative, buy-in can be assured and this is essential to success.

Tackling complex data quality issues can be as much about building a consensus amongst peers as it is about technical investigation, so the leader of any data quality activity should have strong negotiation and communication skills. Resources for root cause analysis and data cleanse will need a solid understanding of the operations of the association and good technical competence to ensure they can see issues from all sides, leading to more robust fixes.

Step 2 – Think about potential data quality partners

The right partner will act as an accelerator, and provide the specialist, yet temporary skills required to get a data quality management initiative started. From planning to design and through to execution, an experienced partner with knowledge of the sector should provide best practice advice and help organisations avoid common pitfalls. They can also provide expertise on the software available to support data quality management, and how best they can be leveraged to maximise the value from the investment. Finally, where resources are scarce, many partners will be able to deliver turn-key solutions to get the ball-rolling as the internal capability is being established.

Step 3 – Decide your strategy

One of the secrets to a successful data quality exercise is focusing on where invention is truly necessary. Whilst the idea of measuring the quality of all data sources may sound appealing, the reality is starkly different. Identify key business processes and regulatory requirements and start with the data items that feed those. In most cases this will reduce the scope to something manageable and, usually, well understood.

Another key piece of advice would be to take a lead from software development, and deliver little and often. Choose one area to make a start and work it through, before starting on the next key area. This will build momentum, allow lessons to be learned, and spread the load on all stakeholders over a broader period, to allow parallel activities to continue unaffected, whilst still providing value and confidence in the potential of the project.

Step 4 – Prove the concept

That word ‘momentum’ is hugely important in data quality projects. Build it by quickly choosing and implementing a software package to support data quality measurement and starting to prove the value in doing the work. Here is where a partner will inject the most value, by getting rule sets developed quickly and accurately.

Bring people on the journey by starting a regular data quality forum with representation from all key areas, even those not initially directly involved. This will encourage people to contribute to how data quality could be improved in their area, and quickly highlight overlaps which may increase the priority of certain issues.

Step 5 – Prioritise the results

Prioritising the issues found is key. Many data quality projects fail because teams become over-faced with volume of issues and lose sight of the value of issues. Low volume issues in critical fields almost always represent more valuable cleanses than high volume issues in less critical areas. Think, for example, do we really need a four or five line address for our processes to be successful, or will the first line and post code be sufficient?

Step 6 – Cleanse and fix

Cleansing data could be an article on its own, however, as a starting point it’s important to create stretching yet achievable data quality targets. Aim for 97-98% quality in key fields, 100% is unlikely to be achievable or cost-effective. Use a prioritisation matrix to identify the short, medium and long term wins. Disregard any which cost more to fix than work around, but look at root causes to try to prevent more being created.

Root cause analysis is also a vital facet of these projects and is what prevents cleansing becoming regret spend. Many organisations overlook this part, and enjoy clean data only for a short time after investing in cleanse. Prevention is always preferable to cure, so look at the end-to-end process, from all angles; from technology issues to training, incentivisation, resource levels and process efficiency. You will find gains to be made.

Step 7 – Expand and repeat

Once a successful pilot has been completed, expand the coverage of the solution to more key areas and processes. Keep the quality forum alive, keep socialising progress and results, and encourage people to highlight workarounds that could be removed or persistent issues that are harming your customer experience.

Finally, enjoy the benefits of high-quality data. Better decisions, made faster; better customer experience, consistently; a more efficient business, delivering real value.