Return on advertising spend is a useful performance metric, but only with accurate attribution. In the modern digital marketing ecosystem, Return on Advertising Spend (ROAS) has long served as the primary North Star for ecommerce brands. It is a deceptively simple calculation: total sales attributed to ads divided by the total cost of those ads. When an organization reports a ROAS of 5:1, it implies that every dollar poured into the advertising furnace yields five dollars in revenue. However, as the digital landscape grows increasingly complex—defined by fragmented customer journeys, privacy-centric browser updates, and the rise of retail media networks—this metric is increasingly proving to be a blunt instrument in a precision-focused industry.
The reliance on ROAS as an absolute source of truth is, according to many industry analysts, a dangerous oversimplification. Mike Murphy, Vice President of Marketing at the attribution firm Incremental, posits that the metric harbors a significant "Achilles heel." When attribution models are flawed, they risk providing marketers with a false sense of security, potentially leading to the misallocation of millions of dollars in capital.
The Evolution of the Attribution Crisis
The historical trajectory of digital advertising began with simple tracking cookies, which allowed marketers to follow a user’s path from an ad click to a purchase with relative ease. For over a decade, last-touch attribution—giving 100% of the credit for a sale to the final ad clicked—was the industry standard. It was easy to calculate and provided a clear, albeit narrow, view of performance.
However, the ecosystem has shifted dramatically since 2020. The implementation of Apple’s App Tracking Transparency (ATT) framework, the gradual sunsetting of third-party cookies by major browsers, and the rise of "walled garden" retail media networks have created a black box around the consumer journey. Today, a shopper might see a retail media ad on Amazon, research the product on a mobile device via a social media influencer’s post, and eventually purchase the item on a laptop using a direct search. In this scenario, the initial retail media ad may have been the catalyst, but if it was not the "last touch," it receives zero credit under traditional reporting, leading to a massive underestimation of the ad’s true value.
The Fallacy of Correlation vs. Causality
The fundamental issue with standard ROAS is that it conflates correlation with causation. When a retailer spends $10,000 on retail media and records $50,000 in sales, the 5:1 ratio is often treated as a success metric for the campaign. But as Murphy notes, this model "credits the last ad touchpoint before a sale, whether or not it caused anything."
This creates a significant reporting gap. In many instances, the ad is taking credit for organic sales that would have occurred regardless of the marketing spend. This is particularly prevalent in retail media, such as sponsored search results on marketplaces like Amazon or Walmart. Because consumers on these platforms are already in a high-intent buying state, they are likely to purchase the product even if the ad had not been served. If the ad serves as a replacement for an organic link that would have been clicked anyway, the "incremental" value of that ad is effectively zero, even if the reporting dashboard shows a high ROAS.
Understanding Incrementality: The True Value Driver
To rectify this, leading ecommerce brands are moving away from surface-level ROAS and toward a metric known as "incrementality." Incrementality measures the actual lift in sales that would not have occurred in the absence of the advertisement.
The distinction is critical. Consider a scenario where a company spends $10,000 on retail media. If the campaign runs while the brand’s product is already appearing in the top organic search results, the ad is largely redundant. If the marketing team discovers that the ad only appears when the organic result is already visible, the true ROAS might be 3:1 rather than the reported 5:1. The remaining 2:1 represents "cannibalized" organic revenue.
This realization forces a shift in budget allocation strategy. Marketers must ask: Is the spend buying new customers, or is it merely subsidizing existing ones? If the goal is growth, the focus must be on incremental gains; if the goal is brand defense, the current ROAS model may be acceptable, but it should not be confused with efficient growth.
The Inverse Problem: Undercounting Performance
While ROAS often inflates performance by taking credit for organic sales, it can simultaneously undercount performance when attribution fails. This occurs when a user interacts with an ad on one device or platform and completes the transaction elsewhere.
Google and other major platforms have attempted to bridge this gap through "conversion modeling." This statistical approach uses machine learning to estimate conversions that cannot be observed directly due to privacy restrictions or cross-device limitations. Google has publicly stated that without such modeling, reported conversions would represent only a fraction of actual campaign performance. Consequently, while some marketers are worried about inflated ROAS due to non-incremental sales, others are missing growth opportunities because their systems fail to capture the full, cross-platform impact of their advertising efforts.
Strategies for Validation and Testing
How can an advertiser determine if their ROAS is accurate or merely an illusion? The most effective, albeit rigorous, method is the "holdout test." By stopping advertising for a specific subset of products or a specific geographic region for a defined period, marketers can establish a baseline of "natural" sales.
Comparing the sales performance of the non-advertised group against the advertised group provides a directional view of the ad’s actual impact. For smaller budgets, this is a manual, albeit manageable, process. For enterprise-level advertisers, advanced retail media networks are increasingly offering randomized control trials (RCTs) or geo-testing, which provide a more statistically significant, automated way to measure incrementality without manually pausing campaigns.
The Financial "Source of Truth"
Ultimately, digital marketers must recognize that ROAS is a proxy, not a balance sheet. The disconnect between ad-platform reporting and actual business health can lead to dangerous outcomes. If a marketing department scales its budget based on high ROAS figures that do not translate into higher net profits, the business is effectively burning cash under the guise of efficiency.
The integration of marketing data with the company’s Profit and Loss (P&L) statement is the final, necessary step in maturity. As Murphy emphasizes, "Your P&L should be your first source of truth—it doesn’t lie." If a company increases its ad spend, there should be a measurable, corresponding increase in the total contribution margin. If the ad spend rises but the bottom-line profit remains stagnant, the ROAS figures are almost certainly capturing non-incremental traffic.
Future Implications for Ecommerce
As we look toward the future, the reliance on single-channel, platform-provided ROAS will likely decline in favor of holistic marketing mix modeling (MMM). These models analyze historical data, seasonality, macroeconomic trends, and advertising spend across all channels to estimate the true elasticity of marketing investment.
The shift represents a move toward greater transparency and financial accountability. Advertisers are no longer willing to accept the "black box" reporting provided by ad tech giants. Instead, the focus is shifting toward an integrated view of profitability where ROAS is merely one of many inputs, rather than the final word on campaign success. For the modern ecommerce professional, the challenge of the next decade will not be in maximizing clicks, but in identifying the true causal relationship between every dollar spent and every dollar earned. By balancing granular attribution with macro-level P&L verification, companies can navigate the current landscape with far greater precision and long-term financial security.
