5 Steps to Data Modelling that Drives Business Value
Alexis is a Principle Consultant in Data & AI and has been working in the business intelligence domain for the past four years, leading projects end-to-end across multiple sectors, delivering true business value. His areas of specialty include data warehouse design and ETL processes. Recently, Alexis was also involved on projects hosting data in the Microsoft Cloud.
December 3rd, 2021
Did you know that 65% of the general population are “visual learners”? This means they need to see information in a visual format to help them fully understand it.
In the data world, almost everyone appreciates the value of representing results visually. Charts and graphs make it easy for users to interpret complex results at a glance, so they can make accurate and timely decisions.
But there is another case for using visuals during your business’s data journey, and that’s right at the onset of developing a data model. Data modelling is a visual representation of how and when data is used in a business, as well as by whom, which helps technical teams produce a highly relevant data model that meets business and user needs.
Let’s take a look at 5 key aspects of data modelling, and how to use it to accelerate the value of data in your business.
1. Hold a workshop and invite all necessary stakeholders
While business users may feel that they already know what they want from a new data solution, a workshop is the best way to dig deep into their requirements to truly understand what’s required.
For example, a department manager might say, “I need a dashboard to manage department performance.” While this sounds great, it doesn’t give technical teams much to go on in identifying which data the department manager relies on to judge how their team is performing.
Holding a workshop with technical teams and business users (such as department managers), allows for greater discussion about:
- What data business users consider valuable
- How they currently obtain the data they need
- How their needs tie in with other business units
One way to encourage input in a workshop is to ask: What is working well? and note down people’s responses. Then ask: What is not working well? and write down these responses. Then you might pose the question: “How might we…” which encourages users to start brainstorming different ideas to answer this question.
As you record responses, put them on a board for everyone to see. Once all the ideas are collated, you can then evaluate each one for their priority level in terms of business value vs the effort required to deliver it.
By doing this, you then have a clear roadmap to start development.
2. Engage with a UX designer
Ideally, invite a UX designer to be part of your initial workshop so they can help facilitate conversations between technical teams and end users. UX designers have a knack for immediately identifying usability issues, which they can call out and address before any development takes place.
If it’s not possible to have a UX designer in your workshop, make a point to engage with them after the requirements are established. They can help to define which reporting features are needed to best display information for maximum usability.
3. Encourage curiosity by playing out data linkages
Business data is frequently very siloed, with separate divisions (such as finance, marketing and sales) each working within their own unique systems and with their own specialised set of processes.
During a data modelling exercise, business users get the opportunity to see how their data links with other areas of the business – which also invites curiosity as to how to make data work better for everyone.
As I wrote about in an earlier article, curiosity is essential for data in business because it often results in business users finding new and innovative ways to engage with data for the benefit of the entire business.
4. Integrate business logic into your data model
When technical teams set out to develop a data model without engaging with users, the outcome is generally far from ideal.
This is because technical teams usually don’t have exposure to everyday business operations, and lack the understanding to integrate business logic into their data designs.
By carrying out data modelling, technical teams are empowered to gain a deeper appreciation for the business logic behind user requests – and use this new knowledge to develop a model with appropriate data rules that accurately meet business needs.
They can also name data entities based on the business’s natural vocabulary, so the outcome is naturally easier for people to use.
5. Sit back and enjoy the benefits
After completing an effective data modelling process, you can simply sit back and watch as the benefits of your early planning and stakeholder engagement efforts pay off.
Some of the benefits you might see include:
Faster development – as technical teams are more confident about what needs to be produced and why.
Less errors – because the majority of these will have been identified and planned for during the modelling phase.
Better collaboration – between technical teams and business users, who now understand each other and can work together in the future to further the business’s data initiatives.
Higher user uptake – having been involved in the development phase, business users are more likely to engage with the new data system once it’s deployed.
More business value – with faster development, less errors, and a model designed specifically for business users, you’ll be able to immediately start leveraging your new data solution for the business value it can provide.
To find out how Antares can facilitate a data modelling session in your business, simply contact us.