3 Challenges of Data Adoption
And tips for data leaders on how to overcome them
Countless companies are embarking on a data journey. And increasingly they — correctly — start by designing a data strategy. But even with a great plan and positive intentions, success is not certain.
One of the reasons why data initiatives are failing is the lack of data adoption across the organisation. Adoption needs to spread from the C-suite to individuals at all levels — all the way to the front-line.
If adoption is low among the C-suite, the data strategy often fails at the first roadblock. When the adoption is thin throughout the organisation, it makes the execution of the data strategy impossibly difficult.
Simply, the higher the adoption, the more successful a data strategy will be in creating tangible business outcomes. But how to drive the adoption? And what are the greatest challenges when doing so?
Challenge #1: What business problem are we solving?
The first challenge is related to the most critical question — what business problem are we solving?
Why should data matter? Why should we pay attention? How would we benefit? How would the business benefit?
The simple answer is to tightly link the data strategy with a business strategy. And ensure that every single data project is targeting a defined business issue.
But, that’s easier said than done. Why? Because when a data professional asks the executives what business problem they want to solve, she often gets answers like: “we need to upgrade the ERP system”, “we need to improve data quality”, or “we need to develop a chatbot”.
That’s all good, but these are not business problems (they are IT and data problems). As a data scientist, we should not accept that answer but rather keep probing to get to a real business problem. We want to hear things like: “we need to increase the ARR”, “we need to scale our business to ten new markets”, or “we need to increase the conversion rate from free trials to a paid subscription”.
If we succeed in pinpointing the actual business problem, we also ‘hook’ a strong senior sponsor because our data project solves their problem! And with that, everything becomes easier — access to budgets and resources, support when things don’t go well, influential role model for others and consequently even easier cooperation with other teams.
On top of this, other business colleagues will start to see a link between their business problems and solutions data and analytics have to offer. We will see a bridge being built between the business and data.
Challenge #2: Ability to make data-driven decisions.
The second challenge is the ability to make data-driven decisions. And it’s not only about business becoming more data driven. It’s about the convergence of business and data worlds.
Let me explain.
Everyone in an organisation needs to work on their data fluency. Everyone. This is a broader concept than data literacy. It goes beyond the ability to read charts. It’s about thinking in data. It’s about the language of the business.
When driving data fluency, two things are helpful to keep in mind:
- People should be compelled by the value of data. That way they are intrigued to learn and feel they would benefit.
- They shouldn’t be intimidated by data. It shouldn’t be considered a scary world, full of difficult and artificial concepts.
On top of data fluency, people need to have access to data in order to be able to make data-driven decisions. And — importantly — the tools providing access must be designed around their users.
A useful consideration is the difference between analysts and non-analysts. Benn Stancil uses a nice analogy saying that, when working with data, analysts work like scientists who keep asking more questions and creating more hypotheses as they work through the data.
Whereas, non-analysts work like journalists. When they have a question, they collect the most important metrics, add their experience, overlay it with a story and — bam — there is the answer.
That’s why data professionals must get closer to the business, and why businesspeople should become more data fluent.
Staying with the bridge analogy — this is where two-way traffic needs to be created, in a constant flow, back and forth.
Challenge #3: De-mystifying and democratising data.
Once we ensure data matters (challenge #1) and data-driven decision making is enabled (challenge #2), we need to lower the bar and spread adoption further.
Four concepts can be helpful in achieving that.
Product management approach. Similarly to how we drive adoption of other products, we should use a product management approach to drive adoption of data products. We need to be responsible for a data product’s overall success — determining user needs, owning the product vision, executing it, marketing the product to the users and supporting them when needed.
No technical jargon. Technical language can be intimidating, preventing people from being more data fluent. We shouldn’t make it harder for them, and so should use business language whenever possible.
Explainable & ethical AI/ML. When advanced analytics, AI or ML is part of a data solution, its methods and outputs should be crisply articulated to non-data scientists. The models should be transparent, trustworthy, and explainable. So that people understand their workings.
Digital twin. Finally, it helps to think of data as a reflection of the real world. People tend to find the concept of a digital twin intuitive. It helps them comprehend the constrains, limitations and assumptions related to (available and missing) data — including its quality — beacuse this is what the ‘see’ in their real, business-as-usual world.
That is how we increase the capacity of the bridge. Giving more people (e.g., other teams and departments) access and comprehension of the data and data solutions.
Conclusion
Wide-spread adoption is critical for success of any data initiative. Otherwise, it will never reach its full potential.
However, driving adoption comes with some great challenges:
- What business problem are we solving?
- Ability to make data-driven decisions.
- De-mystifying and democratising data.
These are meaningful challenges, but not impossible to address. A diligent, pragmatic, and empathetic approach can help us build a bridge between business and data, that carries increasing and varied two-way traffic.
The story of this article has been created for a webinar titled “Prepare your data for adoption from the front line to the C-Suite” done together with colleagues from Keboola and ThoughtSpot. The recording can be found here.
As ever, I’m indefinitely grateful to Chelsea Wilkinson for patiently shaping my thoughts into a publishable format.
