In this age of big data buzzwords, increasingly complex acronyms and CEOs eagerly touting their newest machine-learning-enabled toasters, it’s easy to lose an audience that isn’t fluent in the lingo.
Building, embedding and sustaining a positive data culture as a company grows is challenging enough without having to decipher all of that nonsense. So at Cuvva, we use language that makes sense, and avoid data jargon.
Our data culture allows us to do some pretty powerful things. Here’s a quick look at how it operates for various teams within the business.
Before investing all of your budget hiring the greatest data minds of all time and kitting them out with cutting-edge tools, understand that any culture change starts with strong leadership.
At Cuvva, we’re lucky to be steered by leadership that values data in the decision making process. They prioritise our product roadmap based on actual insights and testing, rather than gut feeling or an idea that came about after last Friday’s pub drinks. (Although we aren’t denying those ideas are sometimes spot on. 😂)
“If we have data, let’s look at data. If all we have are opinions, let’s go with mine.” — Jim Barksdale
Transparency and honesty are key Cuvva values. They aren’t just limited to our interactions with customers - a transparent company culture is critical. After all, it’s no good creating insights if the people who need them don’t have access.
So analysis isn’t locked away and only wheeled out for specific purposes. We host weekly company updates and show-and-tell events, where trends and results are presented, warts and all.
Spreadsheets and visualisations are saved in shared folders, freely accessible to staff. And the whole company is encouraged to ask questions and openly talk about data in Slack - except when the data’s truly sensitive.
Data is a huge part of how our engineers work. Whether they’re building backend services, testing pricing models, implementing new features in the app, or creating dynamic web pages, you can bet data will be involved at all stages of the process.
When changes are made, they’re made because the data suggested it would be a good idea. But it doesn’t stop there. For any new changes our engineers implement, they consider how the impact of these changes will be tracked.
Specifications from the product managers (PMs) will explain the intent of any change, but the details are left to the engineers. We’ve found that by giving them the freedom to solve problems themselves, they come up with some really innovative solutions to capture data.
To complement this, we’ve defined standards for event tracking, database structures and more general data principles. These guard rails make it easier for new joiners to understand the systems, and feel confident implementing features.
An effective PM understands that data sits at the core of their role. At Cuvva, new app features start their journey with data-led profiling, user research or interview sessions to form and validate hypotheses.
We use an objectives and key results (OKR) framework at both a company and team level to guide quarterly roadmaps and make our priorities clear. Progress against OKRs is visually displayed across office dashboards, and product tests are set up with measurement in mind. We strive to find the right balance between qualitative and quantitative metrics.
When a PM is happy with the results of a test (backed up by the data of course), the feature is rolled out gradually to our customers. And all the while our data analysts are on hand to help with everything from formulating test strategies through to calculating statistical significance.
Having PMs and engineers focused on data is a great start. But the true power of data culture comes from engaging the real end users. At Cuvva, this means empowering our Customer Operations team and others in the Company who don’t necessarily have any data-specific training.
Even with accompanying how-to guides, data visualisation tools can feel a little intimidating. So on-boarding sessions and refresher training courses are provided to make teams comfortable building their own queries.
“Give a woman a data and you feed her for a day. Teach a woman to SQL query and you feed her for a lifetime.” - not exact quote
This combination of tools and training allows most of the team to self-serve with minimal support from the Data team. This works for two reasons. Firstly, it empowers the team to look for a way to answer their own questions before requesting help.
And secondly, a real data-driven culture means data is being used at every level of the business.
We’ve found that staff on the front-line often have timely questions that deliver real insights. And sometimes these insights have the potential to shift priorities at the highest level in the business.
With everyone at Cuvva thinking positively about data, the Data team itself has a huge responsibility to support the self-serve environment we’ve created. We also need to automate the boring stuff, and focus on data architecture as products scale and new data sources come into the mix.
At the moment, the team is made up of a Product & Engineering Manager and Data Analyst (with another analyst joining us shortly). Our backend engineers from other teams helped lay the foundations of our current data architecture.
But we’re now on the lookout for a Senior Data Engineer who can take the lead in this area. And in the future, we plan to build out the team, hiring data scientists and machine learning specialists.
Spending analytical time crafting similar dashboards over and over is a waste of resources. Instead, priorities include building master templates for queries, documenting definitions, and working closely with the wider team to make sure their needs are met.
For the engineering side of the team, building a robust, scalable and speedy centralised data warehouse is the main goal. As Cuvva is growing quickly, flexibility is a must. Decisions about longer-term solutions aren’t being rushed - we expect managing through change is easier than ripping everything apart and rebuilding from scratch at a later date.
Building a positive data culture takes effort from every level of the organisation. There’s no silver bullet or single hire that can do all the work. Develop broad goals linked to clear, interpretable metrics, then work out what needs to be technically implemented to make them measurable.
Finding the data evangelists that naturally exist in each team helps bolster the culture. Give them freedom to embed their methods - they’ll be more effective converting others towards an analytical approach than a lengthy memo or company-wide email.
Lastly, don’t expect change to happen overnight. Celebrate small wins and let things snowball as empowered and engaged teams realise the benefits of data.
If you want to help us build out our data culture as we grow, check out our current openings. We’re on the lookout for skilled data engineers and analysts to join our growing team.
Follow us on...