Has Anyone Seen My ROI?

ReadBest PracticesFun Yet InformativeIndustry InsightsStats & Trends
Written By Jeff Winter

Turn the sound on. Trust me, it’s worth it

In the video that goes with this article, I’m wandering around an office asking:

“Has anyone seen my ROI?”

I look under the budget.
I root around in the spreadsheet.
I even check that drawer where random cables and orphaned dongles go to die.

Still no ROI.

It’s a joke on camera, but it’s not far from how many leaders actually experience ROI on big transformation or AI initiatives. At kickoff, ROI is everywhere: on slide 3 of the business case, in the board pack, in the town hall talking points. Eighteen months later, it’s mysteriously absent from the P&L… and everyone is quietly wondering if it slipped out the fire exit during user acceptance testing. The problem isn’t that ROI is a bad metric. It’s that we keep asking it to do a job it was never hired for.

When ROI Is a Flashlight (and When It Becomes a Black Hole)

In the right context, ROI is extremely effective. It provides clarity, enforces discipline, and enables comparability across competing initiatives. When processes are stable, inputs are known, and outputs are predictable, ROI does exactly what it is supposed to do. It helps leaders allocate capital with confidence.

The problem is not ROI itself. The problem is where organizations choose to apply it.

Increasingly, companies are using ROI to evaluate initiatives that are fundamentally different in nature. Transformation efforts, in particular, do not behave like cost reduction projects. They introduce new workflows, reshape decision-making, and create capabilities that did not previously exist. Yet they are often evaluated using the same financial logic as a scrap reduction program or a line efficiency improvement.

That mismatch creates predictable outcomes.

Leaders ask for precise ROI estimates on initiatives where the conditions are not yet stable. Teams respond by either overengineering assumptions to meet expectations or delaying action until the numbers appear “credible.” In both cases, the organization sacrifices speed and, often, the opportunity itself.

Analysts like McKinsey have been repeating for years that roughly 70% of large-scale transformations fail to achieve their objectives, which means the promised benefits never fully land. Bain & Company’s 2023 Transformation & Change Survey puts the number even higher, claiming that 88% of business transformations fail to reach their original ambitions.

At the same time, a 2023 PwC Pulse Survey found that 88% of executives say they struggle to capture value from their technology investments. In other words: the ROI we promised at kickoff is proving stubbornly hard to find once the thing goes live.

The AI story is even messier:

  • Boston Consulting Group’s 2025 research shows only about 5% of companies worldwide are “AI future-built” and actually generating meaningful value at scale from AI, while nearly 60% report little or no impact so far.

  • A 2025 MIT study summarized in mainstream coverage found that 95% of enterprise generative AI implementations show no measurable impact on profit and loss, not because the models fail, but because they’re poorly integrated into real workflows.

So we’re investing heavily, and yet in most organizations, the “ROI column” on AI and transformation is suspiciously empty. That’s not just a leadership problem; it’s a metrics problem. When you point traditional ROI at big, messy, multi-year change, it doesn’t act like a flashlight. It acts like a black hole. All the nuance gets sucked in; very little useful signal comes out.

Why Transformation Makes ROI Disappear

Think back to the video: I start with a confident sense that ROI is “right here somewhere.” Then the scavenger hunt begins. That’s exactly how transformation value behaves inside most organizations. There are three big reasons.

1) We’re using a one-dimensional metric on multi-dimensional value

ROI is designed to measure financial return under relatively fixed conditions. It assumes a level of certainty around inputs, processes, and outcomes. Transformation, by definition, breaks those assumptions. The value often shows up as:

  • Faster decision-making and shorter cycle times

  • Reduced operational and compliance risk

  • Greater agility when the market shifts

  • Better customer experiences and loyalty

  • New capabilities: data, AI models, reusable components, skills

When an organization invests in improving decision-making, for example, the value is rarely isolated to a single cost center. Faster decisions may reduce downtime, improve service levels, and accelerate time to market. But these effects are distributed, interdependent, and often realized over time.

Ocean Tomo’s long-running Intangible Asset Market Value study shows that intangible assets (software, data, IP, brand, know-how, etc.) now account for around 90% of the market value of S&P 500 companies, up from 17% in the 1970s. We live in an economy where “how we decide, adapt, and learn” is where the money is.

Customer experience is a similar story. Forrester’s 2024 US Customer Experience Index found that customer-obsessed organizations reported 41% faster revenue growth, 49% faster profit growth, and 51% better customer retention than those that aren’t customer-obsessed. Those gains are real, but they rarely attach neatly to a single project’s ROI cell within a 12–18 month window.

2) The time horizon is longer than the business case

Most organizations are wired to evaluate investments on relatively short cycles. Annual budgets. Quarterly targets. In some cases, monthly scrutiny. ROI models follow the same cadence. They are designed to answer a very specific question: how quickly does this pay back?

That works well for incremental improvements. It breaks down quickly for transformation. Transformation rarely delivers its full value inside a fiscal year. In many cases, the first phase does not deliver financial return at all. It establishes the conditions required for future return. New data structures. New workflows. New behaviors. None of which show up cleanly in a near-term ROI calculation.

This is where the tension shows up. Leaders are asked to fund initiatives whose value curve does not align with the company’s financial evaluation model. The result is predictable. Projects are either reshaped to fit the timeline or delayed until they can be justified within it.

Neither outcome is particularly helpful. Transformation does not follow a linear value curve. It tends to look more like a step function.

  • Early stages are foundational. They create visibility, establish consistency, and enable coordination. Financial impact is often indirect.

  • Middle stages begin to compound. Decisions improve, variability decreases, and performance becomes more predictable.

  • Later stages deliver disproportionate value. New capabilities translate into new revenue streams, cost structures, and competitive advantages.

Trying to compress that curve into a 12-month ROI model is not just difficult. It is misleading.

3) We never set ROI up to be found

Sometimes the issue is not that ROI is hard to calculate. It is that it was never set up to be found in the first place.

Organizations approve initiatives with vague success criteria, unclear ownership, and no defined measurement plan. Then months later, they ask, “Where is the ROI?” as if it should have appeared on its own. ROI does not emerge by accident. It has to be designed.

The gaps are usually predictable:

  • No baseline established before the initiative starts

  • No agreement on what metrics define success

  • No owner responsible for tracking and realizing value

  • No link between adoption and outcomes

In those conditions, even successful initiatives struggle to prove impact. Not because value was not created, but because it was never structured to be measured. Organizations that consistently realize ROI do a few things differently. They define success upfront, establish baselines, assign accountability, and track leading indicators alongside financial outcomes.

ROI is not something you “find” at the end. It is something you build into the initiative from the start.

Deloitte’s 2023 work on “Mapping Digital Transformation Value” highlights that 81% of organizations still use productivity as the prime measure of digital ROI, and that those using a broader metric set are about 20% more likely to report medium-to-high enterprise value from their digital programs. In other words: most organizations aren’t measuring the full value in the first place—so of course they struggle to “find” ROI later.

Treat ROI Like a Discipline and Design It So It Does Not Go Missing

When ROI disappears in transformation and AI initiatives, the problem is usually not a lack of value. The problem is the way organizations approach ROI in the first place. ROI works when it is treated as a discipline. It disappears when it is treated as a decorative line in a business case. The solution is to redesign how we define, measure, and track value from day one.

Start with testable value hypotheses

Many programs fail because the benefits are written in vague terms. Replace broad statements with measurable hypotheses that can be validated. An example may be to Improve quote cycle time from 8 weeks to 3 weeks to enable faster revenue capture. Each hypothesis needs a baseline, a clear metric, a time horizon, and an owner who is accountable for achieving it. This turns ROI into the roll-up of a series of measurable commitments rather than one abstract target.

Design measurement before kickoff

ROI cannot be found after the fact if the measurement system was never built. Organizations must decide in advance where the data will come from, how baselines will be captured, and which indicators matter. This includes both:

  • Leading indicators such as adoption, cycle times, decision speed, error rates, and customer experience

  • Lagging indicators such as revenue, cost, margin, and retention

When measurement is designed early, ROI becomes easier to track and harder to lose.

Assign value owners rather than project owners

A project going live and a benefit materializing are not the same thing. Someone in the business must own each benefit from start to finish. A value owner must have the authority to change processes, shift behaviors, and adjust incentives until the benefit becomes real. Without clear ownership, benefits drift and ROI quietly disappears.

Review value with the same rigor as project status

Most organizations review uptime, stability, and defect counts. Far fewer review the value being created. Monthly governance needs a value track that asks: which hypotheses are on track? wich need intervention? and what new value has emerged that was not in the original plan? This turns value realization into a living practice instead of a one-time promise.

When organizations adopt these habits, ROI stops being something they search for at the end. It becomes something designed intentionally, measured consistently, and visible throughout the journey.

References:

Jeff Winter
Previous

The Magazine Cover You Really Do Not Want To Be On
Next

When CIOs and COOs Collide: Why AI Is Creating Tension Everywhere