Not Every Conversion Is Impact. Multipliers Show You What Is.

Platform ROAS answers a different question than you think

Every ad platform follows the same logic:

A user converted. Did they see or click an ad? If yes, the platform takes credit.

That answers a specific question: which conversions had an ad touchpoint?

But that is not the question marketers actually care about. The real question is: which conversions happened because of advertising?

Those are not the same.

A repeat customer searching your brand every week will likely convert whether or not ads are running. Attribution still credits the ad. The result is ROAS that looks strong in-platform but overstated relative to actual business impact.

This is the gap incrementality measurement is designed to close.

What an incrementality multiplier actually does

An incrementality multiplier adjusts platform-reported performance to reflect causal impact. It estimates how much of your reported conversions or revenue were genuinely driven by media versus baseline demand.

BA (Base Attributed) is platform or UTM-reported conversions. AA (Advanced Attributed) is the incrementality-adjusted figure.

If Facebook reports 10,000 new customer conversions and your multiplier is 70%, then 7,000 are incremental and 3,000 would have happened anyway. That difference determines whether you scale spend or question performance.

Where incrementality multipliers come from

Not all multipliers are equally reliable. Their value depends on how they are derived.

Geo-tested multipliers (highest confidence). Geo experiments split markets into exposed and holdout groups, then measure the difference in outcomes. This isolates causal lift directly. The logic is identical to a pharmaceutical clinical trial: a drug is given to one group and withheld from another, and the difference in outcomes is attributed to the drug, not to any underlying health trend that was already underway. Geo tests apply the same principle to media. If MMM is observational, geo testing is experimental, and the resulting multipliers are the strongest available evidence of true media impact.

MMM-derived multipliers. Marketing Mix Modeling analyzes historical spend and outcomes to estimate channel contributions. If a geo test is a clinical trial, MMM is the equivalent of epidemiological research. It cannot assign individuals to treatment groups, but by analyzing patterns across a large population over time, it identifies which factors consistently move outcomes. MMM does not isolate causality as cleanly as geo tests, but it captures cross-channel interactions, lag effects, and diminishing returns. Especially useful when you need a portfolio-level view.

Expert multipliers. In the absence of data, experienced practitioners can provide directional estimates based on vertical experience, channel behavior, and spend levels. Expert multipliers are judgment-based, so their reliability scales directly with the relevance of the expert's experience to your specific business context. Treat them as interim inputs.

Industry benchmarks. Benchmarks provide a starting point when no other data exists. They are not tailored to your business and should not be used as long-term decision inputs.

What multiplier values actually tell you

A multiplier is not a scorecard. It is context.

Low multipliers are common in retargeting and branded search. These tactics capture existing demand rather than create it. That does not make them ineffective, but it does mean their role is different.

M-Squared channel benchmarks for new customer acquisition:

Channels like TikTok and YouTube often appear weaker because their impact shows up later through other channels. A user discovers a brand on TikTok, searches for it three days later on Google, and converts through that search session. Platform attribution credits the search. Incrementality, when measured properly, credits the TikTok exposure that started the journey. The 30 to 34% values do not mean these channels are ineffective. They mean their true contribution shows up differently than their reported numbers suggest.

Multipliers change. Your measurement should too.

Multipliers are not static. They vary based on creative quality, spend levels, audience saturation, and seasonality. A multiplier from Q3 does not describe performance in Q1.

Quarterly refresh. M-Squared's standard practice is to recalibrate multipliers each quarter through MMM refresh. For most businesses, this is the right rhythm.

Mid-cycle updates. Significant creative changes, budget reallocation, or a new channel entering the mix can shift incrementality faster than a quarter. In those cases, update sooner.

Promotional period multipliers. Peak seasons require their own multipliers. During high-demand windows, baseline purchase intent is elevated, which changes the math entirely.

How often multipliers should change depends on four things: business size, how often marketing strategy shifts, the brand's promotional calendar, and how frequently geo tests or MMM runs are producing new data. A cosmetics brand running quarterly geo tests and adjusting creative each cycle will likely need updates with every MMM refresh. A specialty food retailer with two major promotional windows per year, say holiday gifting and a summer grilling season, may need only quarterly recalibration plus two seasonal adjustments.

Every multiplier should carry a timestamp and a confidence tag. When either the time or the conditions become stale, refresh before making budget decisions.

Peak seasons distort ROAS the most

Black Friday is the clearest example.

During peak demand, baseline conversion rates rise. Attribution claims more credit. ROAS increases. But incrementality often drops.

Consumers were already planning to buy. Ads intercept demand rather than create it. Using standard multipliers during these periods leads to overestimation and poor budget decisions in the following year, because the inflated reads anchor the next planning cycle.

Smart measurement teams build this into MMM explicitly, running holiday-specific geo tests to capture the seasonal adjustment. The result is a promotional multiplier that reflects actual media lift during the event, not the elevated baseline that would have happened anyway.

Applying multipliers in practice

A structured approach matters more than any individual number.

The Marketing Accounting Framework (MAF) is the structure M-Squared uses to operationalize multipliers across the full media mix. Rather than applying a single multiplier to all activity, the MAF separates multipliers by tactic, source, and time period, then triangulates across methods where data allows.

Best practices:

• Prioritize sources: geo, then MMM, then expert, then benchmarks.
• Separate by channel and tactic. A Facebook prospecting multiplier is not the same as a Facebook retargeting multiplier.
• Tag with time window and confidence level.
• Refresh on a defined cadence.
• Investigate differences across methods rather than averaging them. Divergence is a signal, not noise.

The Multiplier Engine is how M-Squared automates this process, maintaining a live library of multipliers by channel, tactic, and time period, updated through the quarterly MMM pipeline and supplemented by geo tests as they are completed.

Why this matters now

Attribution is becoming less reliable. Signal loss, privacy changes, and walled gardens all push platform numbers further from reality. Platform metrics will continue to overcount. That is not a flaw, it is a limitation of methodology.

Incrementality multipliers are how you correct for it. They convert reported performance into decision-grade measurement, the kind that holds up in front of finance, leadership, and actual business outcomes.

Learn more about how Multipliers fit into your measurement stack. Talk to an Expert