By: Eva Keller
A CMO walks into a board meeting with an attribution report that nobody in the room fully believes, including the CMO presenting it. The model may be sophisticated, and the dashboard may be polished, but the harder question remains: does the analysis actually demonstrate what marketing contributed, or does it simply produce a number everyone has agreed to accept?
That, according to Shiv Gupta, Managing Partner at The Volute Group and founder of Quantum Sight, is the real marketing ROI problem. “Marketing doesn’t have a math problem as much as it has a credibility problem,” Gupta says. “We have gotten very good at producing numbers. The harder question is whether those numbers represent something we can actually defend.”
Part of the problem is that marketing ROI and marketing attribution are often treated as though they are the same thing. Attribution asks which marketing interaction should receive credit for an outcome, while ROI asks a broader question: what evidence do we have that marketing changed customer behavior, and how did that change ultimately affect financial performance?
That distinction matters most in industries where decisions do not follow a simple path. A mortgage decision can take months, a wealth management relationship may develop over years, and patients may choose a health system based on referrals, insurance coverage, physician reputation, prior experience, search, advertising, and word of mouth. Trying to assign every transaction cleanly to an individual marketing interaction can create the illusion of precision without necessarily improving understanding.
“The objective shouldn’t be to force every marketing dollar into an attribution model,” Gupta says. “The objective is to build the strongest chain of evidence you can from marketing activity to customer behavior to business outcome.” Sometimes that chain is direct, but often marketing works through intermediate effects such as consideration, engagement, acquisition quality, retention, penetration, or customer lifetime value.
Those intermediate effects matter if they can be shown to connect reliably to financial outcomes. Instead of asking only which campaign gets credit for a sale, marketers can ask whether an investment changed a meaningful customer or business metric and whether that metric has a demonstrated relationship with financial performance.
The evidence can come from different places. Controlled experiments may work in some situations, while statistical modeling, historical variation, predictive relationships, or direct observation may be more appropriate in others. The method can vary, but the standard should not.
“The issue isn’t whether every relationship can be measured perfectly,” Gupta says. “It is whether we understand how strong the evidence is, what assumptions are being made, and where our confidence should stop.” That is where marketing measurement often gets into trouble, because organizations frequently demand a single ROI number even when the underlying evidence is much less certain.
A model may conclude, for example, that marketing generated exactly $47.3 million in incremental revenue. But if it cannot sufficiently account for seasonality, competitive activity, pricing, sales efforts, prior customer behavior, or other external influences, the decimal point does not make the answer more valid. “One of the biggest mistakes in analytics is confusing precision with truth,” Gupta says. “I’d rather give an executive a credible range and explain the confidence behind it than give them a precise number the evidence can’t support.”
That philosophy shaped the measurement work originally developed within Quantum Sight and now part of The Volute Group. Volute’s SightIQ methodology was independently certified in 2026 by the Marketing Accountability Standards Board through its Marketing Metric Accountability Protocol, reinforcing an approach centered on empirical relationships between marketing activities, intermediate outcomes, and business performance.
In practice, that means some lines can be drawn confidently, others cautiously, and some not at all. “Not knowing something is not a measurement failure,” Gupta says. “Pretending you know it is.” The point is not to eliminate uncertainty, but to make it visible enough that executives understand what is known, how strongly it is known, and what still needs to be tested.
This also changes the purpose of marketing ROI. Too often, measurement is treated primarily as a defensive exercise intended to prove that marketing deserves its budget, but Gupta argues that this sets the bar too low. “The most important purpose of ROI isn’t proving that marketing worked yesterday. It’s helping us decide where the next dollar should go tomorrow.”
That shifts the conversation from historical credit to investment decisions. Where should the company increase spending, where should it reduce it, which activities appear to create incremental growth, where are returns beginning to diminish, and which customer metrics actually predict business outcomes? Those are more useful questions than simply asking which channel gets credit for the last sale.
This mindset also creates a better conversation with the CFO. Marketing no longer has to pretend that every activity can be traced directly to revenue; instead, it can explain which relationships have been demonstrated, which are supported probabilistically, and which remain hypotheses that need further testing. That makes marketing measurement less like an accounting exercise and more like an investment discipline.
The distinction is becoming more important as direct attribution gets harder. Privacy restrictions, fragmented customer journeys, offline activity, cross-device behavior, and influences that occur outside a company’s observable data make perfect attribution increasingly unrealistic. That does not mean ROI becomes unknowable; it means marketers need a broader and more disciplined definition of evidence.
Marketing’s ROI problem, then, is not simply a math problem. It is a credibility problem, and credibility comes from being rigorous about what the evidence demonstrates, transparent about what it does not, and disciplined enough to use both to make better investment decisions.
