Stop creating data strategies. Create strategies supported by data

To achieve success, data should always link to your organisation’s strategic objectives. When it is connected to your company’s goals, it provides immediate business benefit. When you actively try to solve a problem using data, your initiatives are well-defined with who, what, when, where and how, with the power and ability to deliver tangible, impactful and meaningful change.

Think about the KPIs or OKRs you track. These are the metrics you deem to be fundamental to the success of your organisation, and that data should support them. When you implement a strategy supported by data, you instantly gain a competitive edge through insights and continuous improvements in areas that truly matter to your business.

Getting the data right is essential as you modernise. However, it will likely be one of your biggest challenges too.

Data quality: avoid garbage in, garbage out

The impact of generative AI will live and die by the quality of your data. Data quality measures include attributes such as accuracy, completeness, validity, consistency, uniqueness and timeliness. When your data is of poor quality, it can reduce your revenue by up to 25%. Additionally, it can lead to greater downtime, with time-to-resolution for data quality issues having increased by 166%.

Typically, organisational data is siloed, exists in different formats, is recorded inconsistently, and, in some cases, is simply incorrect or outdated. Therefore, the first step to improving your data quality is to assess and understand its current state:

• How much data do you have?
• Where is it?
• What format(s) is it in?
• Who has access to it?
• What controls are in place to govern it?

It’s only once you can see your entire data estate laid bare that you can identify what is or isn’t working, and the steps needed to make improvements.

Data ethics: use data for good

Most data sets are inherently biased. Attitudes and ‘norms’ of the past are reflected in that data, which can then negatively influence decision-making about the future, perpetuating a dangerous cycle. One business area where this data bias is a major concern is human resources and recruitment. For example, Amazon had to stop using an AI-powered recruiting tool that favoured men for technical jobs. The historical data it was trained on was predominantly from male candidates and employees, making the AI believe that male gender was a desirable characteristic for future recruitment.

If left unaccounted for or unchallenged, AI models perpetuate the bias and continue to discriminate against certain groups. Alarmingly, 27% of data professionals will actively check for skewed or biased data during data ingestion.

Another issue is Retrieval Augmented Generation (RAG), where generative AI models seek out data. Without appropriate guardrails in place, AI tools can access data they shouldn’t and potentially present it to the wrong people. Take the Chevy chatbot, for instance, which offered to sell a customer a car for $1.

When training or using AI models, it’s crucial to ensure the tool is set up correctly. Every organisation will do this differently depending on the nature of their business, use cases and risk appetite. However, broadly speaking, you want to ensure that:

• The tool can only access data you’d like it to access
• All AI outputs are checked, and models are retrained when bias is found
• Guardrails are continually monitored to ensure they align with the changing needs of your business and data estate

Data privacy: address security and compliance

Your organisation has to manage and protect ever-increasing amounts of data. This comes in different structural types, presenting a range of privacy and security challenges. Traditional structured data formats (such as files and documents) will often be scattered across your organisation in various departmental or regional data siloes. But increasing amounts of your organisation’s data is now likely to be semi-structured or unstructured, creating additional management and privacy problems.

Finding information in this unstructured data is difficult as it isn’t easily searchable. It’s also hard to collate information for a comprehensive view of a person, product or situation. And if that data falls under GDPR or other industry regulations, you’re more likely to fall foul of fines.

According to Gartner, unstructured data accounts for 80-90% of all data. It can take several forms, including text, images, videos, audio recordings, social media posts, emails and document scans. To be able to use what’s valuable and govern what’s sensitive, you need to get your unstructured data under control.

Again, this comes down to identifying what data you have, but also classifying it. For example, data could be defined as personal, confidential or commercially sensitive. Once it is classified it’s easier to put appropriate controls in place to protect it, as well as demonstrate your ongoing regulatory compliance.

Data governance: retain control of data

Governance is seen as awkward because it requires you to look beyond the technology and consider the process of assigning responsibilities for certain jobs. When you suddenly need data stewards and data curators, it can involve structural changes to the business. These can be costly and hard to achieve. Additionally, governance is seen as boring, bureaucratic and a barrier to innovation.

It’s unfair that governance gets such a bad reputation. Innovation is, in fact, elevated by trustworthy, high-quality and accessible data, with 20-40% of IT budgets being swallowed up by fixing poor data governance issues.

Data governance doesn’t have to be hard. A well-designed data governance programme does involve people with different roles and responsibilities, but it’s a collective effort. Together, they create standards for governing how data is collected, stored and used, allowing you to:

• Understand your data estate, which is good for compliance
• Know data is protected at all times, which is good for security
• Harness accessible data, which is good for your people
• Leverage high-quality data, which is good for business outcomes
• Use data insights to improve your offering, which is good for customers

The need to build a data-driven culture

Not realising the potential of your data doesn’t mean you won’t experience success. However, creating a strong foundation based on reliable data will enable you to achieve much more, including making more informed business decisions, reducing repetitive tasks, elevating personalisation and unlocking previously inaccessible insight.

But good data isn’t just about aiding decision-making. The technology, processes, people and rules that surround data generate additional benefits, allowing you to:

• Make data more accessible and shareable while maintaining its security
• Enable the business to be more agile and react to changes quickly
• Deliver a better customer experience with joined-up data for continuity of service
• Ensure ongoing regulatory and contractual compliance

The emotional impact of high-quality data shouldn’t be overlooked either. It’s great to improve processes and automate manual tasks - everyone can see the benefit of compiling a report in an hour versus a week. It’s easy to calculate the time, money and resources saved. But what’s harder to quantify is the emotional impact of the updated process. The burden of tracking down and manually combining datasets has been lifted from someone’s shoulders. The fear of incorrect data entry and the pain of working late have also been removed. Also, when putting forward suggestions that can affect the business, data can work as a comforting anchor to help validate a claim. Without supporting data, an idea can not only fall flat, but make it generally feel less grounded. These outcomes are often intangible but all equally valuable. Therefore, the success of a data-driven strategy shouldn’t just be measured in numbers but assessed holistically across the business.

Unlock the value of your data

We tailor our cost-effective, secure and scalable data platforms to your business, helping you to keep pace with change, work smarter, operate more efficiently and achieve your goals.

If you’re worried about missing out on valuable insights and competitive opportunities as a result of ineffective or inaccessible data, speak to one of our data experts.