Consultant at Amazon Advertising, Shruti Dash, is building models that tell fashion brands what’s actually driving their sales: ads or discounts
Proving whether advertising actually drives sales or just captures purchases that would have happened anyway has become retail’s most expensive measurement problem. The Interactive Advertising Bureau and IAB Europe released comprehensive guidelines establishing standardised measurement frameworks for commerce media campaigns. With commerce media spending accelerating toward $100 billion by 2028, the guidelines address a problem that has plagued retailers for years: proving whether their advertising budgets drive genuine business growth or simply capture sales that would have happened anyway. Research cited in the IAB framework reveals that 78% of stakeholders identify media measurement as requiring industry alignment, while 70% of buyers cite lack of standards as investment barriers.
For fashion and sportswear advertisers, this measurement problem becomes even more complex. These brands operate on tight seasonal calendars with overlapping promotional cycles and frequent campaign launches, making it nearly impossible to determine whether sales came from advertising, discounts, or simply the right product at the right time. Shruti Dash, a consultant in analytics at Amazon Advertising who specializes in measurement for the Fashion and Apparel vertical, has spent her career developing frameworks that untangle exactly these questions for some of the world’s largest brands.
Working at the intersection of econometrics and marketing science, Dash designs measurement systems that help advertisers separate signal from noise – quantifying the true incremental impact of multi-million-dollar media investments. Her specialty is building models that account for the complex interactions between advertising and promotional activity, providing fashion brands with clarity about what’s actually driving their sales.
Her ad-promotion optimization framework received the Data-Driven Product of the Year award in the Advertising Tech category at the American Business Expo Award in December 2025. She has also contributed to academic discourse in advertising measurement, publishing “Challenges and Prospects of Using Synthetic Control Groups in Measuring Digital Advertising Effectiveness: A Data Quality Perspective” in the Universal Library of Innovative Research and Studies.
When Discounts and Ads Run Together, Who Gets the Credit?
One of the most persistent challenges in retail advertising is determining how much credit belongs to media campaigns versus promotional offers. When fashion brands run advertising while simultaneously offering discounts, traditional attribution models typically overstate one channel’s contribution while undervaluing the other. A brand launching a new collection might run national media campaigns while offering early-access discounts to loyalty members, but which element actually drove the sales?
Dash, who holds a Master’s degree in Quantitative Management from Duke University and has worked across Amazon and Flywheel (a commerce acceleration platform that provides analytics and optimisation tools for Amazon sellers and brands), building analytics products for global brands, has been developing a methodology specifically designed to address this gap. Her framework integrates econometric modelling with elasticity estimation to quantify the combined and separate effects of advertising and promotions on sales. Rather than treating these as independent levers, the approach examines how they interact.
“What we’re really trying to understand is not just ‘did advertising work’ but ‘how did advertising work in the context of everything else happening,'” Dash says. “For example, we might find that TV advertising drives a 2x lift in baseline sales, but when combined with a 20% discount promotion, the total lift is 3.5x rather than 4x. That tells the brand that the promotion is cannibalising some of the ad effectiveness, and they should consider staggering the timing or adjusting the discount depth.”
The methodology is being designed specifically for fashion and sportswear advertisers operating on tight seasonal calendars with frequent promotional cycles. Understanding the interaction effects between media and promotions allows for more precise budget allocation in future launches. Dash’s work aims to give brands a clearer view of the true incremental value of each investment, moving beyond platform-reported numbers that often conflate correlation with causation.
Building a Culture of Rigorous Testing
Within Amazon Advertising’s Fashion and Apparel organization, Dash serves as the experimentation subject matter expert, driving adoption of rigorous testing standards and supporting science-based decision-making across account teams.
Designing experiments that account for these dynamics requires more than standard A/B testing. “You can’t just run a standard test and assume the results are valid,” Dash notes. “You need frameworks flexible enough to accommodate fashion’s volatility while maintaining statistical rigor.”
Her frameworks help Amazon’s teams evaluate campaign effectiveness with greater confidence. When a major sportswear brand wants to understand whether increasing spend on streaming video will drive incremental purchases, Dash’s models estimate the likely impact while accounting for concurrent promotional activity, seasonal trends, and competitive dynamics. This type of analysis influences how millions of dollars get allocated across media channels.
Major fashion and sportswear advertisers increasingly expect their media partners to provide sophisticated measurement capabilities. They want to understand not just impression counts, but what those impressions actually accomplished. Did awareness increase? Did consideration shift? Most importantly, what was the true incremental sales lift – the additional revenue that would not have occurred without the media investment?
Answering these questions requires combining multiple data sources, applying appropriate statistical techniques, and communicating results in ways that non-technical stakeholders can act on. Dash’s role involves all three components: accessing the right data, building models that capture true causal relationships, and translating findings into actionable recommendations for brand partners.
Shaping Industry Standards
Dash’s measurement frameworks have influenced how Amazon Advertising approaches campaign evaluation for fashion and sportswear clients. Account teams now have access to more sophisticated tools for quantifying media impact, including geo-experimental designs that test campaigns in controlled markets, Bayesian hierarchical models that pool data across similar campaigns, and synthetic control methods that create counterfactual scenarios – strengthening Amazon’s position as a strategic partner capable of answering the questions that matter most to advertisers.
Looking ahead, Dash plans to expand the use of advanced econometric and causal-inference methods in advertising measurement. Specifically, she is working to integrate propensity score matching (creates comparable treatment and control groups from observational data), difference-in-differences analysis (measures changes over time while controlling for confounding factors), and instrumental variable approaches (isolate causal effects when randomization isn’t possible). Her current work on the ad-promotion optimisation framework represents one piece of a larger vision: creating approaches for measuring true media effectiveness that account for all the complex interactions in modern marketing.
“My goal is to help shape industry standards for how we evaluate advertising effectiveness,” Dash says. “Right now, too many measurement approaches rely on vendor-specific attribution models that don’t capture true causality. As marketing continues to evolve with privacy regulations, signal loss, and channel fragmentation, we need frameworks grounded in sound methodology that can be trusted across organizations. That means not just developing better models, but sharing approaches that have been tested at scale so the entire industry can benefit.”
Beyond developing new methods, Dash focuses on mentoring the next generation of data scientists entering marketing analytics. As someone who navigated significant transitions: moving countries, rebuilding after an early-career layoff, and advancing through complex roles – she understands the value of guidance.
As commerce media accelerates toward $100 billion in spending and measurement standardization becomes an industry imperative, practitioners who can build rigorous, unbiased frameworks will shape how billions of advertising dollars get allocated. Dash’s work suggests that advertising effectiveness isn’t primarily a technology problem requiring better AI models—it’s an epistemological problem requiring better questions. As commerce media consolidates into a $100 billion market dominated by a handful of platforms, the retailers that win won’t be those with the largest budgets, but those who can most accurately determine which parts actually matter. In an industry where billions get allocated based on attribution models designed by the same platforms selling the ads, the ability to separate causation from correlation may be the most valuable capability brands can build.
