Navigating the Complexities of Conversion Models

What’s up everyone, today I have the pleasure of sitting down with Ron Jacobson, Co-founder and CEO of Rockerbox

Summary: Multi-touch attribution doesn’t tell you what really caused a conversion or revenue, it’s a credit distribution system. It’s still a useful guidepost in understanding where your efforts are making an impact. Incrementality testing, on the other hand, digs deeper—helping you pinpoint what’s really driving results by answering, "What would’ve happened without this campaign?" But to get there, it’s not about finding the perfect model, it’s about asking the right questions. Don’t get stuck in the basics like Google Analytics. True measurement demands first-party data and statistical modeling, especially as third-party cookies fade. For startups, the goal is momentum—nail one channel before diving into complex measurement. Build success first, then refine with tools like MTA or MMM to truly understand what drives growth.

About Ron

• Ron started his career as a software engineer before transitioning to product management at AppNexus where he ran the platform analytics team and later the real time platform product team
• He then took the entrepreneurial plunge Co-founding Rockerbox, first as a programmatic advertising platform then a multi touch attribution platform 
• And today they’ve added a suite of marketing measurement tools that also leverage marketing mix modeling. 

Rethinking the Role of Multi-Touch Attribution

Multi-touch attribution (MTA) often sparks debate around its effectiveness in driving marketing decisions. While many recognize it as a flawed tool, few fully grasp the extent to which it misses a crucial element: causality. When asked whether MTA should be seen as a credit distribution mechanism rather than a way to measure causality, Ron agrees wholeheartedly, explaining that this is exactly how his team has framed the discussion for years.

Ron emphasizes that MTA’s purpose isn’t to assign cause-and-effect between marketing touchpoints and revenue generation. Instead, it's a retrospective tool designed to distribute credit across various touchpoints in a customer’s journey. He argues that marketing teams need to shift their focus from chasing causality to understanding how customers interact with marketing efforts. This approach helps marketers assess what channels or strategies might be working, even if the exact causal impact remains elusive.

A specific example Ron highlights is when clients test new channels like OTT, CTV, or linear TV. Frequently, these clients aren’t sure if the new channel is even making an impact. The issue, he notes, isn’t necessarily that the marketing is ineffective—it’s that the data simply doesn’t reflect customer engagement due to gaps in tools like Google Analytics. While causality is still out of reach, MTA can at least show that the new channel is on the customer’s path to purchase, providing some reassurance that the efforts are not entirely in vain.

Ron points out that this shift in perspective helps marketing teams function more effectively. Rather than getting bogged down by the impossibility of determining exact causality, teams can use MTA to answer more immediate, practical questions: What are the touchpoints that seem to drive the most engagement? Where should we focus next? It’s not about perfectly predicting outcomes, but about gathering insights that improve day-to-day operations.

Key takeaway: MTA isn’t designed to establish causality, but rather to help distribute credit among touchpoints. When marketers focus on how customers engage with their efforts rather than trying to measure cause-and-effect, MTA becomes a valuable tool in refining strategy.

Understanding the Value of Path to Conversion

When diving into the value of the path to conversion, we often struggle with the fact that it doesn’t fully address causality. Just because a customer clicks on a Google link and converts doesn’t necessarily mean that click caused the purchase. It’s possible the customer had already been influenced by a social ad or an email from days prior. Understanding the motivations behind these actions remains elusive.

Ron’s take on this is refreshingly straightforward. He suggests ignoring the model entirely when pitching multi-touch attribution (MTA). Instead, focus on the question: What can you learn from understanding the customer’s path to conversion? By treating MTA as an alternative lens to last-click or first-touch attribution, Ron emphasizes that it provides more context but doesn’t necessarily give a definitive answer to causality. He argues that last-touch attribution, for example, isn’t the best method for understanding the full customer journey.

The real value of analyzing the path to conversion, according to Ron, comes from the variety of questions you can answer. Questions like time to conversion, comparing paths for new versus retained customers, or how adding a new channel influences customer behavior. Retention, in particular, has gained importance as rising interest rates push companies to focus on profitability, and understanding how existing customers engage without paid media is crucial.

Ron points out that the path to conversion isn’t just a credit distribution mechanism but a core dataset that allows marketers to do their jobs more effectively. By looking beyond conversions alone and examining full paths, even those that don’t lead to a sale, marketers can better assess conversion rates and session data. Still, he concedes that none of this answers the critical question of whether marketing spend was truly incremental or whether a customer would have converted without it.

Key takeaway: While path to conversion analysis doesn’t solve for causality, it opens the door to deeper insights. Marketers can use it to answer key questions about customer behavior, retention, and channel effectiveness, but should remain aware of its limitations in proving incremental impact.

Defining Incrementality in Marketing

When we discuss incrementality, the core question is simple: Would the business results still have happened without marketing? It’s a shift in mindset from how we traditionally report on marketing outcomes. Instead of simply attributing revenue to specific touchpoints, incrementality forces us to ask whether that revenue would exist at all if we hadn’t spent that marketing dollar.

Ron emphasizes the importance of having a baseline when assessing incrementality. Without this, everything looks like it’s driven by marketing, which isn’t always true. For him, the key is understanding the marginal return on that last dollar spent. In other words, is each dollar spent still driving profitable results? This approach helps marketers gauge if they’re spending wisely and achieving their business goals.

The real challenge comes in determining the best methodologies to uncover incrementality. Ron explains that while modeling tools like multi-touch attribution (MTA) aren’t designed to measure incrementality, they provide valuable insights when combined with testing methodologies. He highlights that running a holdout test, for example, can reveal incremental results, and applying that test’s findings to MTA reporting allows marketers to optimize daily decisions while still understanding broader trends.

Ultimately, Ron advises marketers to focus less on the methodologies themselves and more on the questions they need answers to. Whether you’re trying to allocate next quarter’s budget or determine the effectiveness of a new creative, the right approach depends on what you’re trying to uncover. By starting with the right questions, marketers can select the best tools or methods to answer them, rather than getting caught up in finding a one-size-fits-all solution.

Key takeaway: In...