Data quality: how to avoid 'garbage in-garbage out'

Nearly every business leader we listen to has the same target outcome for their data: 'getting it right to enable better informed decisions'. But, this can be a big challenge. Poor inputs, processes and analyses all lead to poor outputs. Unsurprisingly, the data market is flooded with products and services to help you understand, control, and analyse data and present your insight.

Getting the service that's right for your organisation depends on understanding what you want to do with your data. An effective approach to planning how to manage your data quality is filtering it through different frameworks. Here's three that should work for you.

Four aspects of data quality

You need to step back and carefully consider what matters to you. This four-aspect framework helps you identify the full range of potential initiatives so you can make a best informed decision on how to allocate your resources.

For example, within value creation are you seeking clarity on where to focus 'cross selling efforts'? For business protection, are you looking to enhance the control around the storage and usage of personal information? Or within operational delivery do you have systems whereby data quality needs to improve to let analysis be executed or straight-through processing implemented?

Then you need to evaluate each potential initiative and identify exactly what needs to change.

Roadmap to data decisions

Obtaining agreement on what change to fund can be tough. To maximise impact you can use the three steps below to plan your roadmap.

• test the tone from the top – before you invest your time, is there really appetite?
• appoint an impartial champion to facilitate the business wide programme – who is best placed to span functions, and understands opportunities and risk?
• create a prioritised landscape of potential initiatives for evaluation and execution – is the consultation broad enough?

Management information pyramid

Organisations we’ve worked with have executed different initiatives depending on what matters to them. The management information (MI) pyramid, or iceberg, is a useful framework for visualising a hierarchy for each initiative:

At the top level, and typically business user interface, many initiatives are at the tip of the iceberg: implementing tooling such as PowerBI TM and Tableau TM to provide information in a visual format that can be manipulated by the user.

At the governance level, many organisations are trying to answer questions about their data such as:

• what is the data
• what is its source
• how was it generated
• what is included
• what is excluded
• when was it added?

They may also want to mature their data governance by developing 'go-to' people ,and formalising roles and responsibilities in ownership and stewardship. To further establish a data quality culture, they're also sending employees on data apprenticeships to build skills and awareness.

Larger corporates, especially those with PLC or banking reporting obligations, multiple systems and or extensive IT transformation plans are going even further into the detail. It's now common to let users build their understanding of data used in reporting and processes through the adoption of lineage tooling to see 'where data comes from and goes to,' as well as dictionaries to ensure users have the capability to understand 'what data is' and 'what data is available'.

For every company, data is another programme of work to understand and plan. You need to think carefully about what should be on your roadmap to ensure the relevant benefit is delivered efficiently.

From data quality to better MI and KPIs

Our next article in this series about making best informed decisions will build upon this baseline by helping you step up a level from data quality to ask 'what management information and KPIs should exist?'

For more guidance on data quality management to avoid garbage in, garbage out, or to discuss the specific needs of your organisation, get in touch with Simon Davidson.