How to Sharpen Your Snapchat MMM Measurement for Better Insights

In today's fast-evolving digital landscape, media measurement can be very difficult to navigate. With changes like GDPR, CCPA, Apple's ATT, GA4, and the removal of third-party cookies eroding traditional tracking signals, relying on a singular measurement solution is no longer enough to understand the combined impact of your advertising.

In this new era, the most effective strategy involves a combination of execution (daily monitoring and optimization), experimentation (monthly/quarterly testing to find efficiencies), and evaluation (quarterly holistic assessments). That's where Marketing Mix Models (MMMs) come in — a methodology long relied on by CPG brands, experiencing a resurgence across verticals like commerce and tech, and expanding into app-focused businesses. MMMs are crucial for evaluating long-term outcomes and optimizing ad budget allocation. But how can you ensure your MMM for Snapchat is providing the most accurate and actionable insights?

Drawing from the recent “NCS Sales Lift Based Priors in MMM” meta analysis with Snapchat, Nielsen, and NCSolutions, here are two key tactics to elevate your Snapchat MMM measurement:

1. Integrate Experimentally-Derived ROAS as "Priors" in Your MMM

Think of "priors" in an MMM as starting points or informed guesses about a channel's effectiveness before the model crunches all the numbers. Traditionally, these priors might be based on historical data or MMM vendor’s expert recommendations to provide variability across marketing. Others will use the same prior ROI across all marketing channels to provide the same starting point for all. Snapchat, Nielsen, and NCS’s meta analysis highlights a powerful way to make your MMM even smarter: using sales lift experiments to create enhanced priors.

The Challenge: Often, there can be a misalignment between campaign-specific sales lift studies (like those from NCS, which measure causal effects of advertising on incremental sales through experimentation frameworks) and the broader insights provided by MMMs. This discrepancy can lead to advertisers questioning how to use both results when educating their organization about performance.

Research Objective: The NCS Sales Lift Based Priors in MMM analysis aimed to identify improvements to MMMs by using the experimentally-derived priors based on NCS sales lift results across industries like beauty and food and beverage.

Here's what they found:

• Preserved Model Quality: Incorporating NCS experiment priors preserved a high level of model quality, with no significant change to core model performance statistics like MAPE and R-Squared. This means you gain better insights without sacrificing the reliability of your model.

• Closer Alignment to True Impact: Using these enhanced priors achieved a closer alignment to NCS experiment results, on average being within 7% of the NCS reported ROAS. This increased accuracy helps bring MMMs closer to the true incremental impact.

• Significant ROAS and Effectiveness Increases for Many: Half of the models actually saw an increase in Snap ROAS with the inclusion of NCS priors. For these increasing brands, the average ROAS and Effective more than doubled, leading to an additional $17.1 million in revenue for those campaigns.

• Broad Benefits: Experiment-derived priors are especially helpful for measuring new strategies, tactics where historical data is scarce, or smaller campaigns where execution levels might not be sufficient for a clear read in a traditional MMM. They also enable greater granularity in your results.

What You Can Do: Run well-designed A/B tests and sales lift experiments with partners like NCS for your Snapchat campaigns and leverage findings as early indicators for performance and optimization. Then, leverage the ROAS numbers from these experiments and integrate them into your MMM as priors. This helps your MMM deliver more precise and accurate long-term results by grounding it in real-world, causal sales data.

2. Model Snapchat Performance at More Granular Levels, Starting with the Ad Format

Many MMMs, in their quest for a broad overview, often group Snapchat ad performance at a high level, often at the channel view (i.e. Total Snapchat). While this provides a macro view, it misses crucial details about what specific elements are truly driving performance. To obtain optimal learnings, you need to dig deeper.

Why Ad Format Matters: Snapchat offers a variety of ad formats, each with unique creative and engaging content opportunities. Knowing which ones deliver the best results is critical for optimizing your spend. A recent analysis into Snap’s ad format performance via NCS studies found that augmented reality (AR) formats have a 1.3x higher ROAS, and 1.5x higher dollars per thousand impressions (DPM) compared to the overall Snapchat NCS medians. This makes AR a top performer in ad format results for both ROAS and effectiveness. But does this type of performance play out similarly in Marketing Mix Modeling?

The Challenge: The "NCS Sales Lift Based Priors" analysis revealed that the majority of models (80%) grouped Snapchat at the total Snapchat level, not breaking out results by individual ad format. This means advertisers aren't seeing which specific ad types are performing best and contributing most to sales.

The Findings: While re-running the MMM models using experimental priors, Nielsen was also able to include format-level detail and model by ad format. Nielsen found that there were two ad formats in particular that were strong ROI drivers: Story Ads and AR (Lens/Filters).

One of the significant benefits of using experiments like NCS sales lift is that they help to unlock granular performance drivers like ad format level and can serve as a comparison point when breaking out similarly (i.e by ad format) within your MMM. This allows advertisers to understand the specific performance drivers on Snapchat and evaluate tradeoffs between ROI and effectiveness more deeply and allow for more confident optimizations for future initiatives.

What You Can Do: When setting up or refining your MMM, break down your Snapchat spend and performance data by ad format. Prioritize understanding the impact of specific formats like AR, especially given its strong performance indicators in other measurement methods.

The Bigger Picture: Remembering that no single measurement tool holds all the answers, Snap has developed a checklist to help MMM-focused advertisers master the 3 E's at Snap: execution, experimentation, evaluation. By running experiments and integrating their learnings into a granular MMM, you are moving towards a more holistic and accurate understanding of your media investment, ensuring your Snapchat campaigns deliver optimal long-term results.

Source: Snapchat, Nielsen and NCSolutions 2025: Meta analysis with results based on 10 individual Nielsen MMMs, in which original priors were replaced by NCS lift result priors for the overlapping time period.