Build a data culture

Data utopia

In a sun-dappled data world, everyone in your organisation has a shared understanding of performance, and can get some of the data they need from the tools themselves. Data is a language that weaves through all elements of your work and conversations. There are no overworked data gatekeepers who have to respond to all the data needs. Instead everyone finds and shares insights, and builds on your audience understanding.

Sounds like a data utopia right? But it’s possible. My first experience building data cultures was at GOV.UK, where data was baked into the service and content design process. We saw the measurable effects of data-driven actions, and had the support to get the data we needed. 

Of course, it’s not an easy road…

Blockers to good data use

Let’s assume that Google Analytics is our central source of website performance data. The new version of GA4 is hopelessly confusing, making it much harder for dabblers to get what they need. Training can provide a good starting point, but you need time and patience to go beyond a basic use of the tool. 

Another challenge is that there are many ways to measure the same thing. In GA4, different reports will give you different results. There is no ‘right’ way to use the tool, just a way that’s right for your organisation, which makes it hard to get consistent results. 

Adding to this, data alone doesn’t indicate success or failure. If you don’t have clear conversions on your site (contact form submissions, purchases) it can be challenging to define success. Performance depends on your user needs, your organisational goals, and how you’ve designed your site and content. 

There are also many data sources to deal with, providing different results. This makes it difficult to get a holistic picture of your performance.

Steps to build a data culture

Dashboards

These can be a good starting point to get everyone on the same page. They allow you to monitor all the things you value in one place. You can see how performance changes over time, and keep an eye on your user needs in order to meet them better.

Here’s a blog post on how to create your own dashboard. However don’t stop here, you need to socialise the data.

Talk data

Make data a core part of your work conversations, whether you’re in a meeting or sprint planning. Share what data you have used to inform your decisions, and share product performance insights. Once a couple of people start to talk about data in this way others should see the value. Discuss which sources of data, and which measures are right for the job to keep everyone aligned and learning. 

Design with intent

Design with intent and measure that intent; identify problems using data, and assign concrete measures to understand whether changes have the desired effect. If things haven’t performed as planned, find out what needs to be fixed.

User insight sessions

Set up a regular session for people to gather and ask questions, share useful data-gathering techniques, and share performance insights and case studies. These sessions will: 

• help upskill your organisation
• prevent duplication of effort from people trying to find similar insights
• keep everyone on the same page in terms of data best practice, and which sources and measures to use
• help everyone to build a clear picture of your audience and performance in their heads, resulting in better work all round. 

Note the ‘user insight’ label – this goes wider than data. Cover everything from user surveys to user research.

If these are done well, you’ll build a shared picture of performance, which will evolve as more insights become available, and external factors change. 

Celebrate success

Data is the carrot. It’s a way to celebrate successful work, and identify problems with a view to fixing them. Used effectively, data will empower your team, help them make the right decisions, and show that their work is making a difference. Creating this environment can be as simple as shouting about successes in meetings, emails and reports.  

Avoid a blame culture

If someone identifies a problem through data, they need the support and resources to fix it. A visible problem is much better than an invisible problem, and it can become a success if fixed. Avoid a blame culture – people should feel like data is empowering them to do a better job, not opening up a potential hole for them to fall into.

Assign data cheerleaders and experts

Once your team has had training on relevant data tools, assign people in each of these roles. A person can be both a cheerleader and an expert.

Data cheerleaders

A data cheerleader could be a junior who is passionate about the user experience and improving the product. They’ll organise the user insight sessions and find people to speak. They’ll keep data in the conversation at meetings, asking questions, pointing to useful tools and connecting people who have related goals. They need to be confident and positive, to push the value of data and be a contact point for those with questions.

Data experts

A data expert will be someone who has good experience with the tools, and the capacity to help others with intermediate-advanced level data questions. Some organisations get stuck with having one data expert for dozens of people, and this leads to over-work. Your data experts need time carved out to provide support, do deep-dives when necessary, and continuous learning to keep up to speed with the tools. 

A data expert does not need to have a background in data. In fact data purists often focus on the wrong things, because they value data accuracy over practical UX implications. I’ve seen experts who block healthy data cultures by criticising other people’s use of data. Digital performance data isn’t exact, and we don’t need total accuracy – we want high-level trends that affect large volumes of users.

Data experts will ideally be approachable and generous with their knowledge. 

A functioning data culture needs ongoing work to keep it alive – the flame will go out without the right fuel. Good data use happens when everyone makes improvements with a shared understanding of what the data means. People are given the time and support to get the data they need. Insights are celebrated, regardless of whether they indicate success or problems. And people talk data, all the time.