# How to measure innovation ROI and prove it upward > Innovation ROI usually stalls at board level for lack of method, not lack of return. How to set the baseline, attribute the gain and prove the value. - Author: Marlon Trettin - Published: 2026-07-07 · Updated: 2026-08-29 - Language: en - Canonical: https://yowpi.com/en/blog/measure-innovation-roi-and-prove-it --- Innovation ROI does not stall for lack of return. It stalls for lack of method. Measuring is not the calculation at the end of the project, it is the decision you take before starting: which business indicator the initiative will move, what the baseline is, and how to separate the gain from everything else. Without that, the result becomes an anecdote in the boardroom. ## TL;DR - Companies measure what is easy, not what matters. Adoption rate, active users and response speed are operational indicators that do not translate into business value. - Among 25 factors tested, redesigning workflows is the one that most influences EBIT impact ([McKinsey, 2025](https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-how-organizations-are-rewiring-to-capture-value)). Only 21% of companies actually redesigned a workflow, and that group captures the most value. - Measuring ROI is, at bottom, an architecture problem. With no designed process, no baseline and no indicator connected to the business, there is nothing to measure. - Four numbers are enough to start a baseline: time, cost, error rate and volume, recorded before anything changes. ## Why does innovation ROI stall at board level? In most cases, the return is not missing. What is missing is a measurement method agreed before the project starts. Many companies invest in people, infrastructure and tools before defining business priorities and success metrics. The project works operationally, but generates no verifiable impact. Add a common vice. Companies measure what is easy, not what matters. Adoption rate, number of active users and response speed are operational indicators that do not translate into business value. A system with 90% adoption that neither cuts cost nor improves customer satisfaction generated no ROI at all. Boards do not refuse innovation. They refuse claims without numbers. When the innovation manager arrives saying the team loved the tool, they lose the round. When they arrive showing the cycle fell from six days to two and freed a given number of hours a month, they win the next budget. ## Measuring ROI is an architecture problem, not a spreadsheet problem There is a name for why so many companies cannot measure return. They bought a tool without redesigning the process the tool was supposed to improve. McKinsey's [State of AI](https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-how-organizations-are-rewiring-to-capture-value) research from 2025 tested 25 factors and found one that outweighs all the others in EBIT impact: workflow redesign. Only 21% of companies actually redesigned a workflow, and that is precisely the group capturing the most value. That explains a problem boards feel when it comes time to approve the bill: attribution. When an automation cuts handling time, part of the improvement may come from the process redesign rather than the algorithm. If nobody designed the flow beforehand, you do not know what to credit the gain to, and you cannot defend the attribution in a board meeting. Here is the point most presentations skip. The ROI gap is rarely a return gap. It is an architecture gap. You do not have a calculation problem. You have a measurement architecture problem. ## How to set the baseline before you start The baseline is a photograph of the process before the innovation. Without it, any improvement is impossible to prove credibly. The work starts by recording the indicators the initiative is supposed to move, before touching anything. Four numbers are usually enough to begin: - Time: how long the process takes today, end to end. - Cost: what each run costs, including the hours of the people involved. - Error: how often the process fails or generates rework. - Volume: how many times it runs per day, week or month. That record is not bureaucracy. It is what separates "we think it improved" from "it fell 40%, here is the measurement". Well-structured projects can show measurable savings within 90 days — but that short timeline is only possible when the baseline exists from day one, not when someone tries to reconstruct it six months later to justify money already spent. ## How to turn the gain into a number a CFO accepts Boards do not decide on technical metrics. They decide on money and time. The ROI report has to speak the CFO's language: cost avoided, revenue added, and hours freed converted into labor cost. The rigor is in the detail. If an automation frees 40 hours a month from an analyst, the value is their hourly cost times the volume, not a vague estimate of efficiency. And the number has to connect to indicators the company already tracks, such as cost per operation, cycle time and NPS. A parallel report only the technical team reads will not sustain a budget. It is worth defining evaluation milestones before go-live — at 30, 60 and 90 days, for instance — each with a clear criterion to continue, correct course or end the initiative. When the measurement is done well, the return shows up. The number that convinces the board is not the vendor's. It is yours, measured in your operation. ## An example of building the case Picture a mid-sized distributor, a hypothetical but common scenario with illustrative numbers, that wants to automate commission calculation for its representatives. Today the process runs in a spreadsheet, consumes two working days a month from one analyst (around 16 hours) and generates recurring disputes over calculation errors. Before buying any solution, the innovation manager records the baseline: 16 hours a month, the analyst's hourly cost, and the number of disputes per closing. After automation, closing drops to a few hours and disputes collapse. Now there is a defensible number: hours freed times hourly cost, plus the avoided cost of disputes that no longer happen. That logic is what underpinned the result at [UniTrust](/en/cases/unitrust), a brokerage that replaced manual commission calculation with an automated module and drastically cut administrative work, and at [Reatop](/en/cases/reatop), which began generating environmental audit reports in seconds with full traceability of the flow. In both cases the gain is measurable because the process was redesigned, not merely digitized. ## Frequently asked questions **I can't justify the investment to leadership right now. Where do I start?** Start with the baseline, not the purchase proposal. Pick a process with clear pain, record current time, cost and volume, and bring leadership the size of the problem as a number before asking for budget. Justifying the investment gets simpler when the cost of the current process is already measured and visible on the table. **I can't see how to measure the result of an innovation project. Is it always subjective?** It does not have to be. The result turns subjective when nobody defines the metric before starting. If you record the current state and pick an indicator the company already tracks — cycle time, cost per operation, error rate — the before-and-after comparison becomes objective. The trick is deciding what to measure at the beginning, not at the end. **My team resists changing the process, and without that there is no gain. How do I handle it?** Resistance usually falls when the team sees its own gain, not just the company's. Culture, not technology, is the main obstacle in these projects. Showing the baseline helps here: when people see how many hours the current process takes from them, the change stops being an imposition and becomes relief. Starting with a small case of visible impact builds the buy-in for the next ones. ## The next step If your innovation projects deliver results but you still struggle to prove it in numbers, the problem is probably in the measurement architecture, not the return. The [Operational Architecture Diagnostic](/en/contact) is a 30-minute conversation to map where your indicators get lost between the process and the boardroom, and what to instrument first.