Enterprise marketing has long suffered from a strange contradiction. Companies spend more on marketing than ever before, track more signals than they know what to do with, and yet still struggle to answer a basic question: “Why did this campaign work?”
Executives can see outcomes—leads, conversions, revenue—but the causal story behind those outcomes often remains fuzzy. That gap matters. At enterprise scale, even small attribution errors can distort budget decisions by millions of dollars. Over time, those distortions compound.
Dhivya Nagasubramanian’s work sits squarely in that gap. Rather than adding yet another dashboard or metric, she helped change how attribution itself is framed and calculated inside large organizations, enabling attribution systems designed for enterprise decision-making at scale.
Why Marketing Measurement Is Mission-Critical
Marketing has always been about allocating capital under uncertainty. But as channels multiplied—email, paid search, display, social, direct mail—the decision process didn’t get more rigorous. It often got messier.
In practice, budgets were guided by familiar patterns:
- What worked last year
- What teams already knew how to run
- What showed clean numbers in isolation
These approaches were clearly incomplete. Different customers respond to different sequences of exposure. Some need repeated reinforcement. Others convert after a single, well-timed interaction. The hard problem was never identifying channels—it was understanding how channels interacted over time.
Traditional attribution models weren’t built for that.
The Challenge
For years, enterprises leaned on first-touch and last-touch attribution models because they were simple and explainable.
- First-touch attribution credited the channel that introduced a customer to the brand.
- Last-touch attribution credited the channel immediately preceding conversion.
Both were easy to compute—and both were deeply flawed.
Customer journeys are rarely linear. A buyer might encounter a display ad, read an email, search organically, receive a retargeting ad, and only then convert. First- and last-touch models ignored the sequence, interaction, and interdependence of these events. Channels that played a crucial nurturing role were systematically undervalued. Others were over-credited simply because they happened to appear at the beginning or end.
The result? Skewed dashboards, misleading ROI calculations, and budget decisions built on partial truth.
The Innovation
Facing these limitations, Dhivya Nagasubramanian advanced an attribution approach that reframed marketing impact measurement from isolated event crediting to sequence-aware, counterfactual analysis across the customer journey.
Rather than treating marketing interactions as isolated events, Dhivya approached attribution as a journey. Customers don’t move in straight lines—they encounter brands through a series of touchpoints over time, some reinforcing interest, others quietly shaping decisions long before a purchase occurs.
Her work introduced a new way of answering a question that had long frustrated marketing leaders: Which interactions truly matter, and which only appear to matter?
Instead of focusing on the first or last interaction, her approach evaluates what happens when a marketing channel is removed entirely from the customer journey. If conversions drop meaningfully, that channel is clearly doing important work. If they don’t, its apparent success may be overstated.
This shift reframes attribution from a debate about credit to a measurement of impact. Channels are valued not by when they appear, but by how much they actually influence outcomes. For enterprise marketers, this means decisions grounded in evidence rather than assumption—allowing budgets to be aligned with what genuinely drives growth.
Fixing a Quiet but Costly Mathematical Error
While re-deriving the Markov-chain–based odds ratio formulation from first principles, Dhivya identified a mathematical error in the published Channel attribution methodology—an error that had propagated from academic literature into real-world implementations across the globe.
The issue caused systematic mis-scaling of attribution outputs. In isolation, the discrepancy appeared subtle. In enterprise deployment, it wasn’t. Attribution models guide budget allocation, and even minor distortions can cascade into substantial financial misallocation across industries.
Dhivya independently reviewed the underlying statistical framework and corrected a calculation issue that affected attribution outputs in enterprise-scale marketing analytics systems that used multi-touch channel attribution methodology.
Broader picture
As enterprises confront rising customer acquisition costs and tighter scrutiny over capital efficiency, attribution has emerged as a material business risk within enterprise marketing organizations rather than a purely analytical exercise, particularly as attribution outputs increasingly guide executive-level budget allocation. With annual marketing budgets often reaching into the hundreds of millions or more, even modest inaccuracies in attribution—on the order of 5 to 10 percent—can quietly divert tens of millions of dollars toward underperforming channels. Over time, these errors compound, reinforcing distorted investment patterns and contributing to CAC inflation, with industry research suggesting that organizations lacking robust attribution models may misallocate up to 30 percent of their marketing spend, according to industry research published by Giant Partners. ( GiantPartners)
Leadership in Action
Dhivya’s leadership in advancing large-scale attribution modeling underscores the pivotal role of data scientists in redefining how enterprises understand and act on marketing effectiveness. By bridging rigorous statistical modeling with real-world deployment constraints, she anticipated the growing need for attribution systems that are not only theoretically sound but also operationally scalable and decision-ready. Her work influenced how attribution insights are translated into enterprise action.
“Effective measurement isn’t just about assigning credit—it’s about building systems that decision-makers can trust,” Nagasubramanian notes. “Attribution must work at scale, across channels and markets, and remain robust as business strategies evolve. The goal was to design models that could grow with the organization, not break under complexity.”
Conclusion
Dhivya’s work contributed to a shift in how enterprises evaluate what drives growth. By transforming attribution from a fragmented, assumption-driven exercise into a scalable, decision-ready system, she reshaped how organizations measure effectiveness, allocate investment, and act with confidence in increasingly complex markets.
The impact is not abstract. It shows up in smarter budget decisions, fairer evaluation of channels, and strategies grounded in evidence rather than intuition. Every optimized campaign, every informed executive decision, and every dollar allocated with clarity reflects the systems she helped build.
In an era where data volume outpaces understanding, her work focused on translating complex attribution models into insights usable by enterprise teams. Her work continues to influence how enterprises operate, compete, and grow, leaving a lasting mark on the discipline of marketing measurement itself.
