Key steps the company took

  • Implemented a simple process to manage and continuously improve data

  • Defined the roles of stakeholder, data steward, and data owner

  • Used a smart data platform to centralise diverse data feeds

  • Trained data stewards in soft skills to engage stakeholders

Establishing new data management processes

The business needed simple, continuous processes to improve its data quality

Alongside a number of partners, we’ve been helping a large international seed breeding company to get more value from its data. The company has been growing quickly – both at home and abroad – thanks to popular products and a strong sense of entrepreneurship.

However, this rapid organic growth meant the company lacked clear data management practices. Different countries and subsidiaries used their own data definitions and variables – and the lack of an unambiguous common data language was starting to hamper operations. Getting good management information was difficult; worse, consolidated reports weren’t always trusted.

But the company saw great potential. It knew better data management could quickly give better insights, and so better control. It would benefit from enhanced insights into its customers and its market. And improved confidence in its data and reporting would enable it to make firm decisions without double-checking. We helped to develop, implement and execute simple data management processes that would help to define, standardise and continuously improve the company’s data. Based on our trusted continuous approach – but tailored to the business and its maturity level – we needed processes to:

  • Create business and technical definitions for data, to establish a unified understanding and language
  • Define, measure, and continuously improve data quality
  • Identify and resolve the cause of data quality problems

All these processes needed to be repeatable, and to apply equally across various data domains. Although we were starting small, the processes needed to be performed continuously, rather than being a one-off project where the data could become muddled again. This way, data quality improves structurally and permanently over time.

The business used concrete use cases to choose which processes to apply first. As departments requested a report, it meant the concepts must be unambiguously defined, and the quality must be good enough to produce reliable results. Linking data management improvements to specific reports in this way helped stakeholders to understand the value of the process. For example, when it became apparent that different regions were using different definitions, the reason for writing a standard version became clear: it’s a logical step to create a reliable international report.

Defining roles, technologies, and competencies

Implementation involved a smart data platform, data owners, data stewards, and skills

To deliver the data management processes in real life, we needed to define three separate roles: data steward, data owner, and business stakeholder.

Data stewards take a lead role in implementing the processes. They establish data definitions, propose quality controls, monitor outcomes, facilitate stakeholder discussions, and investigate the causes of poor quality data. 

Data owners hold the ultimate responsibility for the data. They enable the data steward’s work, and approve the proposed definition and quality rules. 

Business stakeholders are also known as process owners. They set the desired business outcomes for the process, and indicate what should be included in the report – discussing this with the data owner.

Initially, the company’s data was stored across disparate systems, and hidden in an array of spreadsheets. To bring these sources together, the business invested in a smart data platform. This central source helps to unlock the data, and makes it more accessible. The platform delivers the reports though Microsoft BI, giving fast, interactive access to the information. The company’s own data quality dashboards are delivered the same way. Further developments, connecting the data platform to other applications, is ongoing. But the technology can only do so much. In order for the reports’ meaning to be clear and reliable, the data steward needs to establish an unambiguous lexicon of definitions in advance.

This is one reason why training is so important, with data stewards at the forefront. The skills required go far beyond theoretical knowledge of data quality practice; they also need the real-world ability to help stakeholders overcome conflicts about data definitions, and demonstrate the need for a data management process. As the data stewards acquire and use these skills, the knowledge naturally spreads around the organisation – creating awareness and understanding in other parts of the business.

Figure 2. Important data capabilities of a data owner

Concrete use cases deliver results

Although its data management processes are still relatively new, the business is already seeing pleasing results from the work. Decisions are faster and more effective, because the reports are clear and unambiguous to understand. It also expects to achieve a reduction in operational costs by eliminating slow manual work.

One reason the added value has been seen so quickly is that the investments were made in small steps, each just enough to achieve the next clearly defined outcome. Costs are kept in check, and the benefits are clear. For example, improving a specific report used in a particular meeting could make that discussion measurably faster by making the terms clear and unambiguous. That’s a tangible, concrete outcome which shows the benefit of data management to stakeholders across the business – better than ambitious promises which take longer to deliver.

Concrete use cases also surface needs and priorities from the organisation. Data stewards quickly discover – and request – the training they need. And the added value stops people attempting to do data management alongside the day job, as the business case builds to hire dedicated data professionals.

This in turn builds the company’s own in-house expertise, spreading knowledge to colleagues in other departments. Eventually, the growing number of interconnected applications and roles expose a requirement for a data architect to organise all the activities into a coherent data flow. And that’s when the possibilities get really exciting.

To find out more about how we can help, get in touch.