When Is Marketing Mix Modeling Most Worthwhile?

When Is Marketing Mix Modeling Most Worthwhile?
Table of Contents
  1. Template item

This is an excerpt from MarketerHire's weekly newsletter, Raisin Bread. To get a tasty marketing snack in your inbox every week, subscribe here.

On the privacy-first web, it’s getting harder to spot impactful channels. 

Try to track clicks and conversions, and you’ll often overvalue bottom-of-funnel channels and undervalue discovery channels, The Information recently reported.

Marketing mix modeling (MMM) offers a bigger-picture approach. 

This type of statistical model accounts for relationships between two variables you know first-hand: channel-level spend and revenue over time. 

A growing number of tools, like Facebook’s free Robyn and SaaS platform Recast, help automate the MMM setup process.

But are insights from  MMM worth the time and tinkering it takes to set up? Recast co-founder Michael Kaminsky told MarketerHire MMM is most likely to produce value when… 

You want to measure channel incrementality better. 

MM can capture the incremental revenue impact of performance channels — like Facebook Ads — and less trackable channels — like linear TV — in a realistic, apples-to-apples way. 

So if you use multiple channels, each with different performance tracking capabilities, MMM can come in handy, Kaminsky explained. Done correctly, it minimizes  bias towards bottom-of-funnel channels. 

You have strong historical data available. 

How much data you need, exactly, depends on what you’re selling and how it’s bought. 

The longer the consideration cycle for your sales, the more channels you use for marketing and the more precision you want from the model — the more historical data you need on…

  • How much revenue you brought in per day/week month, and 
  • How much you spent on each of your marketing channels

You’re selling online and offline.

In 2015, Kaminsky started at Harry’s, a DTC men’s shaving brand, heavily invested in performance marketing.  

But then Harry’s started selling not just online, but through brick-and-mortar chains like Target and Walmart — and invested in harder-to-measure marketing, like podcast and TV ads.

“All of [the] machinery that we had built [for] tracking people across the internet no longer was very useful,” Kaminsky said. “The DTC business… was just one part of the total business.”

Media mix modeling helped Harry’s navigate the transition — and will likely also help… 

  • DTC brands that open owned brick-and-mortar stores, like Warby Parker
  • Legacy brands that open e-commerce shops, like Nike
  • Brands that sell via walled-garden channels, like Amazon 

Our takeaway?

MMM captures the big picture, relative impact of channels in a way channel-specific metrics just can’t — especially when you’re making sales online and off, and you have strong historical data available on spend per channel. 

Mae RiceMae Rice
Mae Rice is editor in chief at MarketerHire. A long-time content marketer, she loves learning about the weird and wonderful feedback loops that connect marketing and culture.
Hire Marketers
Explainers

When Is Marketing Mix Modeling Most Worthwhile?

September 8, 2023
April 19, 2022
Mae Rice

Marketing mix modeling (MMM) takes more setup than native dashboards — but it offers clearer, less biased insights into each channel’s incremental impact, too. We asked Recast co-founder Michael Kaminsky when it’s most impactful.

Table of Contents

This is an excerpt from MarketerHire's weekly newsletter, Raisin Bread. To get a tasty marketing snack in your inbox every week, subscribe here.

On the privacy-first web, it’s getting harder to spot impactful channels. 

Try to track clicks and conversions, and you’ll often overvalue bottom-of-funnel channels and undervalue discovery channels, The Information recently reported.

Marketing mix modeling (MMM) offers a bigger-picture approach. 

This type of statistical model accounts for relationships between two variables you know first-hand: channel-level spend and revenue over time. 

A growing number of tools, like Facebook’s free Robyn and SaaS platform Recast, help automate the MMM setup process.

But are insights from  MMM worth the time and tinkering it takes to set up? Recast co-founder Michael Kaminsky told MarketerHire MMM is most likely to produce value when… 

You want to measure channel incrementality better. 

MM can capture the incremental revenue impact of performance channels — like Facebook Ads — and less trackable channels — like linear TV — in a realistic, apples-to-apples way. 

So if you use multiple channels, each with different performance tracking capabilities, MMM can come in handy, Kaminsky explained. Done correctly, it minimizes  bias towards bottom-of-funnel channels. 

You have strong historical data available. 

How much data you need, exactly, depends on what you’re selling and how it’s bought. 

The longer the consideration cycle for your sales, the more channels you use for marketing and the more precision you want from the model — the more historical data you need on…

  • How much revenue you brought in per day/week month, and 
  • How much you spent on each of your marketing channels

You’re selling online and offline.

In 2015, Kaminsky started at Harry’s, a DTC men’s shaving brand, heavily invested in performance marketing.  

But then Harry’s started selling not just online, but through brick-and-mortar chains like Target and Walmart — and invested in harder-to-measure marketing, like podcast and TV ads.

“All of [the] machinery that we had built [for] tracking people across the internet no longer was very useful,” Kaminsky said. “The DTC business… was just one part of the total business.”

Media mix modeling helped Harry’s navigate the transition — and will likely also help… 

  • DTC brands that open owned brick-and-mortar stores, like Warby Parker
  • Legacy brands that open e-commerce shops, like Nike
  • Brands that sell via walled-garden channels, like Amazon 

Our takeaway?

MMM captures the big picture, relative impact of channels in a way channel-specific metrics just can’t — especially when you’re making sales online and off, and you have strong historical data available on spend per channel. 

Mae Rice
about the author

Mae Rice is editor in chief at MarketerHire. A long-time content marketer, she loves learning about the weird and wonderful feedback loops that connect marketing and culture.

Hire a Marketer