Data Doesn't Automatically Deliver Value: Data Products Don't Either

Time travel with me a bit to 2015. You’re in a boardroom, listening to the promise of how data would transform the entire company. “The missing piece is data. We’ve never been able to truly understand our customer, how they have or how to optimize our processes and products. That’s why we need to invest heavily into this data space, today. If we don’t, we’re going to fall behind.” If you’re reading this, you might note to yourself “that isn’t just 2015…” You’re right.

For as “fast” as the data space purports to move, the problems it proposes to solve seem to remain fairly static over time. There is this elusive idea of “value creation” from data. The thing that keeps me up at night is that it is too easy to fall into the trap of assuming if we provide the data, the value will come. But here’s the issue: data doesn’t inherently contain value, and capturing it and providing it doesn’t automatically solve a set of problems. I blame part of this on the hype of the data space over the past 10 years. I blame the other part on asking the wrong question.

Typically, I see the question “What data/models do we need in this area of the business that we don’t have now?” But answers to that question often completely miss the value proposition. Instead, I’d like to see this question answered: “What do we want people to do differently as a result of having this data asset and why is that type of action important?” The latter question(s) get at the value proposition of the data. Data isn’t just data - it becomes a part of an action or series of decisions. Data becomes an asset for someone.

Making data an asset for someone is the core of the idea of thinking about data as a product. The primary role of a data product manager is to identify the gap between the data itself and the data become an asset for a customer and/or company and to create a plan for how to close that gap in a way that is reasonable. There are many, many ways that data can become an asset: golden datasets trusted company wide (and externally), metrics that provide a true pulse on the health of the business, experiments that result in critical product decisions, models in production that shape user behavior and provide the right content to people.

Most of what I’ve said above might not seem particularly surprising if you’ve read any previous editions of this newsletter. Why, then, am I talking about it at length? Because it is very easy to say this, but much harder to do it. In fact, I’d bet most of you reading this probably have said something similar in previous years. Yet we don’t see radical behavior changes around data or data products within companies. Here are 5 ideas as to why that is (and potentially how to work through these):

1. Responding to the “problem of the moment” results in momentarily useful products. There’s often a big problem, and as a result, a big push to solve that problem. In many cases, there’s a “can’t miss” data solution. The issue is, this “big problem” may not be a big problem at all. It may be a momentary pain felt by a small number of customers. In this case, there’s no way a data product can transform people’s behavior, because the problem the product was intended to solve was not persistent. Or it affected such a small group that it doesn’t make a blip on the radar.

2. Solving for the last 3 years of the organization, instead of the next 3 years. Product managers and people respond to customer’s pain. It is quite literally part of their job. We end up in conversations that go like this “would it be helpful for us to solve this for you?” or “What’s most frustrating for you when you are working with data?” People are going to naturally focus on the frustrations of past years. In some cases, it makes sense to solve for these issues as they persist. But in others, the frustrations of the past years won’t be relevant for the business in the next years. If we live entirely in solving past problems, rather than anticipating where we are going, we end up in a tough spot.

3. We don’t ask for commitment from the right people. When we build data products, we often get commitment from engineering, from data science and analysts and from leadership. But we often fail to get commitment from the people who ultimately use the product the most - the actual customers! This doesn’t mean they sign a contract for how they use the data, but there should be thorough discussion of expectations. If you don’t set these expectations early and discuss them often, the changes you want to see with a data product are going to have a hard time materializing.

4. Failing to think about incentives around behavior change. Technical people often laugh at the phrase “if you build it, they will come”, not just because of the Field of Dreams reference but because they know actually changing customer behavior is hard. This is where we need to have a real discussion about incentives. Behavior change doesn’t happen without a reason. People need to find real value in the data and data products available to them. We either need to a) make sure the product integrates naturally into the workflow so behavior change isn’t really needed or b) create very strong incentives that allow people to see why they should change their workflow.

5. We care too much about measuring impact. Not everything that matters can be measured. Not every change we make is going to show up in a metric, at least in the short term. Sponsorship for many data products is focused on “this much revenue will move in the next year” or “we will save this much money over the next 6 months”. These are great to have (I think data product managers should be able to articulate this type of impact), but these type of statements miss the big idea: what *value* are you creating for your customers and how do you know you are creating that value? Answering this question is hugely important.

Whether you are in that boardroom in 2015, 2021, or 5 years from now, you will hear incredible stories about how data can transform a company, about how data is the secret sauce that will unlock opportunity and about how data is the biggest asset that companies have. This might be true. But it won’t come to fruition without intentional thought about how that change will occur. Change comes through people, and people need reasons to change.