Executive Overview
Return on Advertising Spend (ROAS) has long occupied a sacred place in the digital marketer’s pantheon of key performance indicators (KPIs). Touted as the ultimate compass for connecting capital investment directly to top-line revenue, its mathematical foundation is deceptively simple: divide total attributed sales by the cost of ads. For decades, this clean, unambiguous ratio has empowered brand managers, growth marketers, and C-suite executives to compare disparate campaigns, justify budgets, and make rapid-fire resource allocations across Google, Meta, and an exploding ecosystem of retail media networks.
Yet, beneath the glossy dashboards and triumphant presentations lies a foundational vulnerability. ROAS, on its own, is fundamentally blind to the nuances of human behavior, consumer intent, and the messy reality of multi-touch customer journeys. According to industry experts like Mike Murphy, Vice President of Marketing at attribution firm Incremental, traditional ROAS suffers from an “Achilles heel”: its total reliance on flawed attribution models.
When calculated using standard last-touch or platform-reported metrics, ROAS can drastically mislead organizations. It routinely takes unearned credit for organic conversions that would have occurred naturally anyway, while simultaneously failing to capture off-platform, cross-device, or delayed sales generated by top-of-funnel efforts. In an era where retail media networks (RMNs) command billions of advertising dollars and consumer privacy restrictions complicate tracking, relying on surface-level ROAS is a high-stakes gamble.
This comprehensive investigation explores the systemic flaws plaguing modern ROAS calculations, examines the critical divide between mere attribution and true incrementality, and outlines actionable testing methodologies that modern e-commerce brands must adopt to ensure their profit and loss (P&L) statements align with their marketing dashboards.
Detailed Chronology: The Evolution and Current Crisis of ROAS
To understand why ROAS has become simultaneously indispensable and deeply problematic, we must trace how performance measurement evolved alongside the digital economy.
The Era of Simple Attribution (Early 2000s to 2010s)
In the early days of performance marketing, digital advertising operated largely on direct-response models. Banner ads, early search engine marketing (SEM), and email campaigns were measurable through straightforward tracking pixels and cookie-based monitoring. Last-touch attribution—giving 100% of the credit for a sale to the final ad clicked before checkout—reigned supreme. Because the digital landscape was less fragmented and consumer journeys shorter, last-touch ROAS provided a reasonably accurate directional signal. Advertisers could spend a dollar, watch a cookie track a direct sale, and calculate their returns with high confidence.
The Multi-Device and Privacy Revolution (2010s to 2020s)
As the digital marketplace matured, consumer behavior shifted dramatically. Shoppers began moving fluidly across multiple devices (smartphones, tablets, desktop computers) and interacting with dozens of touchpoints before making a single purchase. Concurrently, privacy regulations such as the European Union’s GDPR, the California Consumer Privacy Act (CCPA), and Apple’s App Tracking Transparency (ATT) framework dismantled traditional cookie-based tracking.
Despite these seismic structural shifts, many advertising platforms and internal marketing teams continued relying on legacy attribution frameworks. Platforms naturally leaned into attribution models that flattered their own performance, often claiming credit for conversions they did not strictly cause. ROAS transformed from an objective financial metric into a heavily platform-dependent estimation tool.
The Retail Media Boom and the Incrementality Reckoning (Present Day)
Today, retail media networks like Amazon Ads, Walmart Connect, and Target’s Roundel dominate digital ad budgets. Brands pour billions of dollars into these platforms because consumer purchase intent on retail sites is inherently exceptionally high.
However, this high-intent environment exposes the fatal flaw of standard ROAS: it struggles to differentiate between a customer who bought because of an ad and a customer who was already searching for the product and would have purchased it organically. As e-commerce profit margins tighten amid macroeconomic pressures, brands are waking up to the realization that an impressive 5:1 ROAS on paper may actually be masking stagnant or declining true growth. The industry has officially entered the era of the "incrementality reckoning," where marketers must separate hollow attribution from genuine business growth.
Supporting Context & Metrics: Unpacking the Attribution Blind Spot
To truly grasp why standard ROAS calculations frequently collapse under scrutiny, one must dissect the mechanics of how advertising platforms assign value to consumer interactions.
The Illusion of Last-Touch Success
Consider a classic e-commerce scenario: a mid-sized merchant invests $10,000 into a retail media campaign on a major marketplace. Using a standard last-touch attribution model, the analytics dashboard reports that the campaign generated $50,000 in direct sales.
The resulting ROAS ratio is 5:1—every single dollar invested yielded a fivefold return in revenue. To a busy executive, the logical directive is unambiguous: scale the budget immediately. Double the ad spend to $20,000, and expect $100,000 in return.
However, as Mike Murphy points out, this decision-making process is fundamentally flawed. Standard ROAS credits the final ad touchpoint prior to the sale, completely ignoring whether that touchpoint actually caused the transaction.
"That leaves a very big gap," Murphy explains. "ROAS can take credit for organic sales that would have happened anyway, or for a sale an earlier touchpoint drove. It also only sees what can be tracked directly."
In retail media environments, where shoppers are already actively browsing with credit cards in hand, organic search visibility often does the heavy lifting. If an ad appears above or alongside an organic product listing, the platform’s attribution engine frequently sweeps in and claims full credit for the conversion, even if the shopper would have clicked the organic listing moments later.
Incrementality: The True North of Marketing Efficiency
To combat the illusions of traditional attribution, sophisticated marketers are turning to incrementality testing. Incrementality measures the exact lift an advertising campaign produces above and beyond what would have happened naturally in the complete absence of that advertising.
Let us revisit our $10,000 retail media campaign. Upon closer investigation, the brand’s data science team discovers a critical operational reality: approximately 50% of the time the retail media ad appears, the brand’s products are not organically visible on the primary search results page. In those specific instances, the ad acts as a true driver of discovery, capturing demand that otherwise would have been lost to competitors.
Conversely, during the other 50% of the time, the brand’s organic listing already dominates the top of the search page. When the ad appears alongside the organic result, the ad captures the click, but the organic result would likely have secured the sale anyway.
When accounting for true incrementality—allocating full credit only when the ad created a non-organic outcome—the real ROAS drops from a flattering 5:1 down to a more sober 3:1. The campaign is still profitable, but the margin for error is far narrower than the platform dashboard suggested.
The Reverse Problem: Under-Attribution and Missing Revenue
While over-attribution artificially inflates ROAS, the opposite phenomenon occurs with equal frequency: under-attribution.
An ad campaign may successfully drive consumer demand, but the digital tracking infrastructure fails to record the transaction. This frequently occurs in omnichannel retail environments. For example:
- A shopper encounters a sponsored ad on Amazon or Google while browsing on their mobile device during a morning commute.
- Later that evening, the consumer visits the merchant’s direct-to-consumer (D2C) web store on a desktop computer to complete the purchase.
- Alternatively, the consumer sees a digital ad, visits a physical retail brick-and-mortar storefront a few days later, and buys the item in person.
In both scenarios, the advertising investment directly influenced consumer behavior, but standard attribution models fail to connect the dots. The ROAS reported by the ad platform appears artificially low because a significant portion of the revenue generated remains invisible to the tracking pixel.
To combat this, tech giants have rolled out complex solutions. Google Ads, for instance, utilizes automated conversion modeling to estimate unobservable conversions across multiple devices and privacy-restricted browsing environments. Google explicitly notes that without predictive conversion modeling, its reported metrics would represent only a fraction of total campaign performance. When attribution fails to capture the full scope of an ad’s impact, ROAS registers as deceptively low, risking the premature termination of highly effective marketing initiatives.
Official Statements and Industry Insights
As the digital marketing ecosystem grapples with these measurement crises, industry leaders are increasingly vocal about the need to overhaul how performance is evaluated.
Mike Murphy of Incremental has been a leading voice in urging brands to look past the surface-level metrics served up by self-service ad platforms. In professional analyses and industry communications, Murphy emphasizes that digital advertising cannot be managed in an analytical vacuum.
"Your P&L should be your first source of truth — it doesn’t lie," Murphy asserts.
While advertising platforms are incentivized to demonstrate high returns to encourage sustained spending, corporate financial statements reflect the unvarnished reality of cash flow, inventory movement, and bottom-line profitability. Murphy argues that marketers must bridge the gap between media dashboards and financial accounting. When media spend increases month-over-month, corporate P&Ls should reflect a proportional expansion in gross profit and net revenue. If ad spending scales aggressively while overall business profitability stagnates or declines, leadership must interrogate the validity of their reported ROAS metrics.
Furthermore, platform architects and independent analysts alike agree that self-service attribution models are inherently biased. Because ad networks operate as both the player and the referee—hosting the ads, tracking the user interactions, and reporting the resulting performance—their native ROAS metrics should always be treated as hypotheses rather than absolute facts.
Future Outlook: Testing Methodologies and the Path Forward
Navigating the post-cookie, privacy-first era requires e-commerce organizations to adopt rigorous validation frameworks. Relying solely on platform-reported ROAS is no longer a viable strategy for scaling sustainable growth. Moving forward, modern digital marketing operations must embrace proactive testing and holistic financial alignment.
1. Implementing Holdout and Geo-Tests
To determine whether attributed sales are genuinely incremental, brands must incorporate controlled testing into their marketing strategies:
- Holdout Tests (For Smaller Budgets): For brands with modest advertising budgets, Murphy suggests executing simple holdout experiments. A company can completely pause advertising spend for a specific, randomized subset of products or customer segments for several weeks, while maintaining normal ad spend for a control group. By comparing the sales trajectories of the tested products against the control group, marketers can establish a directional baseline of true ad effectiveness.
- Geographic and Randomized Trials (For Enterprise Budgets): Advertisers managing substantial capital investments—and maintaining close partnerships with major retail media networks—have access to advanced testing methodologies. Geo-testing involves running campaigns in specific metropolitan markets while withholding ads in matched control markets. This provides statistically significant, randomized insights that bypass the self-serving attribution models of native ad platforms.
2. Diversifying Performance Metrics
While ROAS will undoubtedly remain a useful directional metric for day-to-day campaign optimization, it must never operate as a standalone arbiter of success. Forward-thinking marketing teams are broadening their analytical frameworks to include complementary indicators:
- Customer Acquisition Cost (CAC) and Lifetime Value (LTV): Evaluating whether acquired customers generate repeat purchases that justify upfront acquisition costs.
- Blended ROAS: Calculating total company revenue divided by total marketing spend across all channels, eliminating channel-specific attribution bias.
- Contribution Margin: Measuring the exact profitability of a product line after subtracting variable costs, shipping, and ad spend.
3. The P&L as the Ultimate Source of Truth
Ultimately, the future of digital advertising measurement belongs to the marriage of marketing analytics and financial accounting. As capital becomes more expensive and profit margins tighter, the era of unchecked digital spending based on vanity ROAS metrics is drawing to a close.
Brands that thrive in the coming years will be those that treat ad platform dashboards with healthy skepticism, aggressively test for true incrementality, and anchor every strategic marketing decision to the immutable truth of the corporate P&L.
