Blog Series: The ROI-Driven Marketing Playbook
In this article, we’ll cover the final set of tools in a company’s performance marketing toolkit. This is the last article in our series, and we’ll be focusing on the bigger picture – understanding a customer’s start-to-finish journey and using that understanding to make smarter spend decisions across your entire marketing budget.
We’ll dig into customer journey mapping and the concept of a media mix model, a unified view of marketing impact that accounts for the multi-touch, multi-device, and multi-channel nature of customer purchase journeys.
But before we do, let’s recap how we got here. In the first four articles in the Grow Smart series, we covered The Fundamentals of Performance Marketing ROI, The Early Days of Performance Marketing ROI, Master Attribution, and Unlocking Incrementality. By now, you’ve likely realized a few things: performance marketing is messy, attribution is imperfect, and even the best experiments have limitations. That’s where customer journey mapping and media mix modeling come into play. These are advanced tools that allow companies to go beyond campaign-level optimization to build a unified, end-to-end view of how marketing spend influences customer behavior across channels, devices, and touchpoints.
Let’s start with a few new concepts.
Customer Journey Mapping
Customer journey mapping is the process of documenting the path a customer takes from their first interaction with your brand to their final purchase (and ideally, beyond). It’s a visual and data-driven representation of the steps a customer takes, the channels they interact with, the devices they use, and the touchpoints that influence their decision-making.
It’s important to clarify that customer journey mapping is not an attribution model. It does not assign percentages of credit to each touchpoint in a purchase path. Instead, it helps you understand the structure and sequence of customer behavior: what common paths exist, what combinations of touchpoints are typical, and where friction points may occur in the customer journey. Attribution modeling, by contrast, attempts to quantify how much credit each touchpoint gets for a conversion, often expressed as a percentage.
Think of customer journey mapping as qualitative and diagnostic, while attribution modeling is quantitative and formulaic. Customer journey maps help companies identify the moments that matter. Then you can apply attribution or incrementality techniques to those moments to measure their impact.
An example: a journey map might reveal that a large percentage of first-time customers consistently engage with product comparison pages before purchasing. This insight highlights that these pages are a key moment of consideration. The company might then invest in A/B testing those pages to improve conversion rates, or ensure they are more prominently linked from ad landing pages.
Another example: a journey map might show that a significant number of users drop off after clicking a retargeting ad but before checking out. Your team might investigate how customers who click a retargeting ad are different, whether there’s friction in the checkout flow, or test a new promo offer during that stage.
Either way, journey mapping surfaces behavioral patterns that inform where to focus optimization efforts, and attribution or incrementality testing can then help quantify the value of those efforts.
A typical customer journey might include:
- Awareness (e.g., seeing a TikTok ad or streaming TV commercial)
- Research (e.g., visiting your website from Google Search)
- Consideration (e.g., browsing reviews or opening a retargeting email)
- Purchase (e.g., converting through a promo code in a direct channel)
- Repeat behavior or churn (e.g., subscribing, unsubscribing or abandoning)
Why does this matter? Because as we already know, most customers don’t convert after seeing just one ad. They engage with brands over time and across platforms. Without mapping this journey in a sequential way, you’re likely not understanding the paths that customers take prior to converting. If you’re looking at attribution in a vacuum and relying on black box attribution models to assign credit to each touchpoint, you’re missing dependencies and potentially distorting true attribution. At a minimum, if you’re looking only at last-touch attribution, you’re almost certainly undervaluing the early-stage marketing that drove awareness.
To summarize, customer journey mapping helps marketers understand which paths are common, which touchpoints are influential, and where customers drop off. When paired with a robust attribution and incrementality framework, it can transform your marketing approach from being channel optimizers to end-to-end journey architects.
Media Mix Modeling
Media Mix Modeling (MMM) is a statistical approach that helps you understand the relationship between your marketing spend and business outcomes, typically revenue or conversions, across all marketing channels. MMM is especially valuable in multi-channel environments that include DTC, retail, and wholesale channels. Unlike multi-touch attribution, which focuses on individual user behavior, MMM works at an aggregate level and is best suited for analyzing long-term trends and top-down budget allocations.
MMM uses historical data (ad spend, revenue, seasonality, promotions, etc.) to isolate the effect of each marketing channel on your KPIs. Specifically, MMM relies on multivariate regression analysis, a statistical technique that models the relationship between multiple independent variables (like ad spend by channel, holidays, and promotions) and a single dependent variable (such as total revenue or conversions). The model assigns a coefficient to each input variable, which quantifies its impact on the output variable while controlling for the effects of the other variables in the model. In other words, it estimates how much of the observed variation in business outcomes can be explained by each marketing input, net of noise and confounding factors.
Media mix models can also be designed to account for the fact that different marketing channels influence customer behavior on different timelines. This is typically done by including lagged variables in the model. Lagged variables allow the model to measure not just the impact of spend in the current period, but also its effect in subsequent weeks or months. In this way, a well-designed MMM can distinguish between channels that drive immediate response, such as paid search or email, and channels that operate higher in the funnel, such as paid social or TV, where impact occurs over a longer horizon.
The end result of MMM is a set of coefficients that act as a scorecard. The coefficients tell you how much each channel contributes to revenue after controlling for the other variables. It’s a powerful tool, backed by statistical theory, that gives you a clear estimate of each channel’s contribution to overall revenue performance. These scorecard coefficients have an added benefit: they can be used to calculate the marginal ROI of each channel, helping you assess how additional spend is likely to perform at the margin.
An example: Consider a DTC brand that spends across paid search, paid social, TV, and email. Using three years of data, the company designs an MMM with total weekly revenue as the dependent variable.
- The model estimates that for every additional $1,000 spent on paid search in a given week, revenue increases by approximately $3,200 in that same week, implying a coefficient of 3.2.
- Paid social shows a smaller immediate effect, with a same-week coefficient of 1.4, but additional coefficients of 0.9 in the following week and 0.5 two weeks later, indicating that social spend tends to influence customer decisions over time rather than driving instant conversions. Note that a coefficient below 1.0 does not mean the channel reduces revenue; it means the incremental revenue is less than the incremental marketing spend.
- TV shows a modest same-week coefficient of 0.6, with a stronger effect appearing in the week after the ads run, with a coefficient of 1.2.
- Email displays a strong same-week response, with a coefficient of 4.0, reflecting its role as a high-intent, conversion-focused channel.
Based on these results, the company shifts some budget from channels that primarily drive immediate response into channels that support longer consideration cycles, while maintaining investment in email to capture demand once customers are ready to convert. The outcome is a more balanced budget that aligns spend with how customers actually move through the purchase journey.
Running an MMM can be complex both technically and organizationally. Fortunately, there are software tools and specialized consultancies that can help. On the software side, tools like Recast, Rockerbox, and Nielsen’s MMM platform offer structured environments to ingest your historical data, define the right independent variables, and run statistically robust models. These tools often include user-friendly dashboards, built-in data connectors, and features to simulate different budget allocation scenarios based on your model’s outputs.
For companies looking for a more guided or hands-on approach, there are also high-touch consultancies that specialize in media mix modeling. Firms like Analytic Partners, Gain Theory, and marketing science teams at major agencies offer experience in structuring MMMs from scratch. They not only bring statistical rigor to the modeling process, but also help your team decide what data to use, which independent variables to include, and how to interpret and act on the results. This is especially helpful if you’re not sure whether to model digital and offline channels together, how to treat brand versus performance spend, or how to incorporate external factors like seasonality and macroeconomic trends.
Whether you’re using a SaaS platform or working with a consultancy, the biggest lift often comes from data readiness. You’ll need to assemble clean, granular historical data for your MMM model. This includes data on marketing spend, revenue, pricing or promotions, and other contextual variables. Deciding which independent variables to include is both an art and a science. Include too few and your model may miss key drivers of performance; include too many and you risk overfitting or diluting the signal. This is why domain expertise and close collaboration between marketing, analytics, and finance is critical.
Let’s say you spend $1M per month across Meta, Google, email, TV, and influencers. MMM might tell you that Meta and email have the highest marginal ROI, Google is flat, and TV shows long-term brand lift but low immediate conversion. With this insight, you can reallocate spend, test hypotheses, and scale what’s working while cutting waste.
What makes MMM powerful is its ability to handle:
- Multi-channel behavior
- Offline media (e.g., TV, print, out-of-home)
- Seasonality and promotions
- Diminishing returns at high spend levels
As marketing organizations mature, the goal shifts. Early on, success looks like winning individual channels, campaigns, or tactics. But over time, the challenge becomes system-level optimization: understanding how channels interact, how customer behavior unfolds across time, and how spend decisions compound across the entire journey.
Customer journey mapping and media mix modeling are the tools that enable this shift. Journey mapping provides the structural understanding of how customers actually move from awareness to conversion, revealing dependencies, friction points, and moments of influence. Media mix modeling brings statistical rigor to those insights, translating spend into measurable business outcomes and helping teams allocate capital with discipline.
Together, these approaches move marketing decision-making closer to how finance already operates: balancing short-term returns with long-term value creation, measuring marginal impact, and allocating resources based on evidence rather than intuition or platform-reported metrics.
The companies that win with performance marketing over the long run are not the ones that chase the highest ROAS in any given week. They are the ones that understand the full system—how demand is created, captured, and sustained—and invest accordingly.
That is what it means to grow smart.
About the Author:
Karsten Loose is co-founder and Managing Partner at Karlon Group, a fractional finance and accounting firm that helps companies build, scale, and optimize their finance and accounting functions. Karlon Group works with companies across SaaS, e-commerce, manufacturing and technology, offering a full suite of finance and accounting support tailored to each client’s changing needs.