More Than Data

When the Pitch Went Sideways: Lessons from Using an Open-source MMM Package

Jan 31, 2025By More Than Data
More Than Data

Let’s face it: most people prefer to read about soaring victories, groundbreaking innovations, or miraculous turnarounds—the kind of feel-good stories that restore faith in the human spirit. This, unfortunately, is not one of those stories. This is the tale of an unsuccessful pitch we made to a client using the famous open-source marketing mix modeling (MMM) package, Meta Robyn (a.k.a., Facebook Robyn). But don’t despair! We promise it’s at least mildly entertaining and, spoiler alert, it has a happy ending (kind of). Think of it as a comedy of errors—with actionable takeaways.

The Setting: Meta Robyn and the Small Media Agency

We decided to use Meta Robyn in our pitch for a simple reason: it’s famous and free. Like a startup borrowing a Tesla for a meeting to seem more impressive, we thought leveraging Meta Robyn’s reputation would help us secure the trust of a small media agency with just 20 employees. The agency had no internal analytics team, no coders, and no capacity to sense-check the outputs of a complex open-source tool. But hey, the middle person who introduced us to the agency was a trusted bridge between us and the client’s CEO, so what could possibly go wrong?

A lot, apparently.

The Pitch: A Comedy of Errors

We prepared a pitch deck that emphasized the value of marketing measurement and minimized the technical jargon. The middle person, with whom we’d built years of trust, gave it a thumbs-up before the big meeting. Armed with confidence, we headed into the pitch meeting ready to impress. The first half went splendidly. The agency’s CEO nodded, asked insightful questions, and seemed intrigued.

Then came the live demo of Meta Robyn.

Here’s where things took a turn for the… awkward:

"We don’t have analytics professionals. What do we do with this?" 

The agency’s CEO seemed to visibly recoil when we opened Meta Robyn’s script. It might as well have been hieroglyphics to them. Without an analytics expert on their team, they felt lost and intimidated.

"This is taking forever…" 

To demonstrate Meta Robyn’s capabilities, we ran a sample model. Unfortunately, Meta Robyn building models with evolutionary computing takes time. Our underpowered laptop chugged along like it was running on hopes and prayers, and the client’s patience dwindled. We tried to fill the silence with small talk, but it was clear—they were bored.

"We believe TV and OOH drive most of our sales. Can we test that?" 

The agency had strong assumptions about which media channels drove their KPI (sales). They wanted to test these theories. Unfortunately, Meta Robyn didn’t support customizable testing of media assumptions. Strike three.

"DIY or outsource? Neither works for us." 

When we proposed becoming their analytics partner to handle the modeling on their behalf, they balked at the cost. On the flip side, they didn’t want to outsource everything—they wanted a tool they could use themselves. Meta Robyn’s complexity made this DIY dream impossible.

The result? 

An increasingly skeptical CEO, a restless room, and a polite yet unmistakable decline of our proposal. We left the meeting empty-handed, questioning both our choice of tool and our approach.

The Lessons: What We Learned

Failure isn’t fun, but it’s instructive. Here’s what we learned from our Meta Robyn misadventure:

  • Complexity kills confidence. Small agencies without analytics professionals are intimidated by complex tools. A solution needs to be simple, user-friendly, and accessible to non-technical users.
  • DIY flexibility matters. Agencies want the ability to customize and test their own assumptions—like determining the impact of specific media channels on KPIs. Meta Robyn’s rigidity made this impossible.
  • Speed is key. Long run times, especially in live demos, are a major buzzkill. No one wants to sit through a lengthy process when they expect quick results.
  • Budget constraints are real. Many smaller agencies can’t afford to outsource analytics, yet they need solutions that fit within their limited budgets.

The Comeback: Introducing Minute MMM

Every failure carries a seed of opportunity. We realized there was a gap in the market for a marketing mix modeling tool designed specifically for small agencies—one that’s easy to use, fast, flexible, and free. So, we went back to the drawing board and created Minute MMM.

Minute MMM is our answer to all the challenges we faced during that ill-fated pitch. Here’s what makes it different:

  • No coding required. Even non-technical users can build models with ease.
  • DIY-friendly. Agencies can test and customize their own assumptions, empowering them to take control of their marketing decisions.
  • Lightning-fast results. Models are built and results are delivered in under 10 minutes (check out this thrilling (or not) video on YouTube—it has no subtitles, no narration, just Minute MMM running with a timer to track the runtime).
  • Affordable for SMEs. In fact, we made Minute MMM completely free for small media agencies.

A Final Word

Our failed Meta Robyn pitch taught us a valuable lesson: great tools aren’t enough if they don’t meet the needs of the user. By putting ourselves in our clients’ shoes, we were able to create something we’re truly proud of. Today, More Than Data is helping small media agencies achieve big things—no coding skills, long wait times, or massive budgets required.

So, to all the small media agencies out there: if you’ve been looking for a marketing mix modeling tool that actually works for you, give More Than Data a try. It’s free, it’s fast, and it’s built with your needs in mind.

Visit our website to get started. Because sometimes, failure is just the beginning of a better story.