Extraction vs Realisation of Value
Making your data solution matter.
Background
I love having a chat with Dr. Rakesh Menon, a close personal friend and ex-colleague at Citi. We have very similar ways of thinking when it comes to the practice of data analytics / data science domain. In a recent conversation, we briefly landed on the topic of value extraction vs value realisation. For many, this would seem like an argument of semantics, but for us, it was far from that.
Both Rakesh and I are acutely aware of the shortcomings of the data analytics / data science community, where much of their work doesn’t translate into implemented solutions or towards achieving their intended outcomes. A key contributing factor is the gap in the cognitive side of the practice; the lack of deep thoughtfulness in both problem articulation and solution design.
And so I dedicate my 62nd article to discussing why value extraction isn’t the same thing as value realisation in the practice of data analytics / data science.
(I write a weekly series of articles where I call out bad thinking and bad practices in data analytics / data science which you can find here.)
The Value Chain
In a conversation years ago, a Chief Analytics Officer (CAO) friend of mine gave me this fantastic perspective of how the work of data analytics / data science fits into the larger scheme of problem-solving. It’s a translation journey, and the diagram below succinctly sums it up. “A business problem needs to be represented as an equivalent data problem so that a data solution can be formulated. The data solution then needs to be translated into a business solution that addresses the original business problem.”
I equate the two middle translation stages to value extraction. It begins with data exploration and ends with data solutioning. Simply put, value extraction is the act of distilling actionable information signals from data. It’s about putting the information signals to work; asking the so-what questions.
Beyond the obvious diagnostic analytics, this could also include the building of predictive models (using heuristics or more advanced algorithms) and data solutions, or even data products. Undoubtedly, value creation is time-consuming and intellectually challenging work. And most data analysts / data scientists will have no problem embracing it as the core of their job scope.
But data solutions don’t necessarily translate to business solutions or business capabilities. They often die on the branch due to problem-solution mismatch and inability to integrate into the organisation’s production environment. And that’s the realm of value realisation.
Value Realisation
Value realisation is the act of turning a data solution into a business solution. But the process starts much earlier in the way the business problem is defined, as illustrated in the diagram above. Now, data analysts / data scientists tend to leave value realisation to their business stakeholders, claiming that since they are embedded in the day-to-day operating process, they are going to be more familiar with the nuances and should ultimately be responsible for that final stage in the execution of the data solution.
From my experience, there’s a good chance that business stakeholders are operating on incorrect axioms and assumptions. Consider the practice of Marketing. For example, contrary to what most marketeers believe, robust evidence indicates that (a) focusing on retention instead of acquisition isn’t going to make your business more profitable, (b) mass marketing strategy trumps personalised marketing, and (c) both physical and mental availability is what drives your brand’s success (see the book How Brands Grow by Byron Sharp). Imagine the amount of false equivalences and faulty assumptions in the other disciplines! This is where data analysts / data scientists can play can important role in value realisation.
Inwards to Outwards
From the diagram above, one can argue that value creation is a set of inward-focused activities — the data analyst / data scientist is solving for the challenges of the business. Success is getting the data solution right for the stakeholders; getting the information signals to align. Value realisation, on the other hand, is an outward-focused exercise; focusing on the broader implications for the customer, the reason for the existence of the organisation. Rather than blindly adhering to a domain’s default thinking and approaches, the data analyst / data scientist should consider if the derived insights from the data solution hint at a more universal phenomenon. The goal is to generalise, simplify and inter-relate. Everything should fit into a finite set of operating principles. Like the principle of physical and mental availability in Marketing — why should the customer care about your brand if you haven’t solve for its physically availability?
Here’s a set of questions that the data analyst / data scientist can ask to turn their value extraction into value realisation:
- How would the customer experience this data solution vs how should the customer experience it? What is the nature of the gap?
- How does this data solution amplify or reinforce the existing intelligence capabilities of the organisation? What information signals are being reinforced? Is there a generalised theory of customer behaviour that this data solution would fit into?
Conclusion
There will always be brilliant people working on the technical difficulties of “data problem → data solution”. And we should celebrate that. But it won’t matter until it matters for the customer; it won’t matter until it matters as a competitive advantage for the organisation. Figuring out how the data solution fits into to the larger scheme of things, into the customer journey and experience, into the customer’s value chain, is critical to turning the value extracted into realised value.
