Leveraging Retrospective ROI to Reposition Data Strategy

Where’s the Common Data Thread?

Imagine you’re part of a data science team inside a global industrial manufacturer. You’re building analytics products to help drive everything from product development to customer uptime. Your team sits within engineering—but your work touches commercial teams, product managers, business development, even customer-facing software groups.

You’re in a decentralized system with three main lines of effort: (1) helping engineering teams improve the core product based on field usage data; (2) building consultative, data-driven offerings like maintenance insights that customers can purchase; and (3) supporting software products like tracking systems for connected equipment. All three use the same core data—telematics, warranty, service records—but there’s no common thread connecting the effort. No one’s thinking across use cases. No one’s responsible for spotting patterns that apply across business lines.

And no one’s job is to say: “Here’s what the data tells us about how our products are used. Here’s how that insight should shape the next generation of offerings.”

So the work happens, but it’s mostly ad hoc. Some product managers request insights. Others don’t know what to ask for. And while the engineering analytics team delivers reports and dashboards as requested, they’re rarely in a position to proactively guide strategy.

For leaders trying to demonstrate ROI, that’s a problem. Because when the organization doesn’t know where to point the team, it starts to question the value of the hard work being executed by the D&A team.

Start with a Retrospective ROI

One of the most effective ways to make the case for centralized data and analytics investment is to look backward.

If your team is regularly fielding questions about the cost of analytics or struggling to get buy-in for infrastructure investments, a retrospective analysis can often shift the conversation. Start by identifying five to ten major initiatives from the last two to three years—projects that required significant effort to assemble, clean, and harmonize data.

Then, ask probing questions like: How many of those datasets were reused? How often did the team rebuild pipelines or manually connect systems for a single use case? What did it cost—in hours, resources, or opportunity—to deliver each one?

From there, begin to sketch a counterfactual. What would the investment profile look like if shared infrastructure had been in place? How much of the effort could have been reused? How much faster could follow-on projects have gone?

It also helps to go one step further. In cases where projects required data from multiple domains, estimate what cross-functional value might have been captured with better integration. And where possible, quantify potential revenue or margin improvements from products built on shared rather than isolated data sets.

The goal isn’t to boil the ocean. Even a small set of well-analyzed projects can create a compelling narrative: without investment, we’re reinventing the wheel. With it, we’re building once and reusing often, which means we’re cutting cost, accelerating delivery, and laying the groundwork for growth.

Centralizing Data Efforts: A Path to Alignment and Investment

In one recent conversation, a client team shared a use case that sharpened the distinction between centralizing data and centralizing analytics.

They’d built a solution to help customers better manage maintenance and downtime. The value was real: over three years, the projected cost savings topped $2 million. But as the client explained, that figure understated the opportunity. Without a central data infrastructure, the initiative relied heavily on manual work, such as data cleaning, one-off integrations, and limited reuse across customers. The real constraint wasn’t in the analytics. It was in the plumbing. With a better-managed data layer, they could have scaled the solution more broadly and more efficiently.

Internally, the team faced similar challenges. On engineering analytics projects, they often spent up to 25% of their time reconciling definitions and debating data meaning. That hidden tax drained capacity and delayed delivery, especially when paired with other staffing constraints. Several high-value initiatives were sidelined simply because there wasn’t enough bandwidth to do the work and manage the friction.

I find this example so powerful because it demonstrates the human levers in D&A progress. The client saw a clear path to value through centralizing data, but they were careful to distinguish that from centralizing analytics. That nuance mattered. Centralizing analytics, in this case, could trigger organizational pushback; other teams may feel like their autonomy or domain expertise is being threatened. But a centrally managed data layer? That’s an enabler. It makes everyone faster, more consistent, and more effective—without stepping on toes.

Further, when you’re selling an investment in foundational data capabilities, lead with shared wins. Don’t just show how it would make your team more efficient. Show how it makes everyone better. And when resistance arises, make sure it’s aimed at the right thing. Most of the time, people aren’t pushing back on the value of clean, connected data. They’re pushing back on perceived control. A thoughtful pitch can separate the two.

When Value Crosses Lines, So Should Your Strategy

It’s easy to underestimate how political the analytics ROI conversation can get, especially when no one clearly owns the thing that’s generating the value.

I’ve seen it happen time and again. A data product or insight lives in the middle of multiple functions. Engineering generates the data. IT pipes it in. Business teams want access to the insights. And analytics ends up bridging the gaps. But when you try to formalize that value—by pitching a center of excellence, for instance—it’s not clear who should sponsor it. No one wants to lose control, it’s human nature. And if you're not careful, a perfectly rational investment can start to feel like a threat.

The trick is to know when you're entering this territory and have a game plan ready. That means slowing down to listen. And I mean really listen. Not to convince or position, but to understand. You want to know what excites each stakeholder, what worries them, and what they’re afraid of losing. That takes more than a slide deck. It takes a couple of coffees, one-on-one. It takes curiosity, patience, and some trust in the long game.

In cases like these, I’ve found that it helps to reposition the value as infrastructure. You’re not building a new power base. You’re building plumbing. Nobody fights over who owns finance or HR—those are just part of how the business runs. The same can be true for a shared data platform or a cross-functional analytics capability, but only if you frame it that way from the start.

Still, structure matters. Even if your pitch resonates, the reporting line can quietly kill your momentum if the function is nested under someone with competing priorities. If one leader owns both the platform and a product area, it's only natural their team will optimize for their own needs. Others will be deprioritized. So be direct. Call out those dynamics early. If the business is ready for a dedicated leader, say so. If it’s not, make the case that wherever this function lives, it must operate like a shared service, as opposed to a subordinate product team. This will help you preserve the credibility of the investment and ensure the organization can actually capture the value it's already agreed to pursue.