The Recast Philosophy

One of the first things people generally want to know about Recast is how it compares with their other media measurement systems. Generally, the clients we work with have set up an attribution system and have run a few lift tests, and perhaps their data science team has built, or their TV agency has provided them with, a media mix model (MMM). What does Recast replace? Where will the numbers be different or similar?

To break it down, we generally have to take a step back and explain how we frame the world of media effectiveness measurement, which, in our view, encompasses three qualitatively different approaches to measurement:

  • Touch-Based Attribution, which focuses on telemetry, and has the goal of assigning customers, orders, and even order items, to specific channels that get “credit” for the conversion
  • Experimentation, which involves — you guessed it — running an experiment, with control and treatment groups across experimental units, which could be individuals, households, geographies, stores, etc., with the goal of estimating a treatment effect
  • Regression, which uses observational data combined with a statistical model to produce coefficients that parameterize the relationship between channel spend and the relevant business metric (new customers, revenue, etc.)

Each of these approaches have their strengths and weaknesses, and all have their place in a modern measurement system. We cover the strengths and weaknesses of each of these approaches in detail in this blog post, but here we’ll focus on how companies use Recast (a regression-based tool) in this context.

Why should you use a top-down model for marketing measurement?

Top-down regression-based models like Recast augment touch-based attribution systems in a few important ways:

  • Reducing the need for using e-commerce coupon-codes for customer tracking
  • Providing an apples-to-apples comparison of channel effectiveness across the marketing funnel
  • Most importantly, providing an estimate of causal impacts that help marketers predict where their dollars will be most effective.

While many organizations already do some amount of experimentation to attempt to answer these questions, for various reasons, most companies are not able to experiment as frequently and in as many channels as they would need to in order to effectively manage their marketing budget through the method of experimentation. That means that if you want a view on causal impacts in your marketing budget, you need a model like Recast.

How to incorporate Recast into decision-making process

The team we described at the outset, who “set up an attribution system and have run a few lift tests, and perhaps their data science team has built, or their TV agency has provided them with, a media mix model (MMM)” is already employing a mix of all three approaches. How should they be integrated in a principled way?

While there’s no right answer, this is how we think about it:

First, we should remember that all this measurement is in service of setting a budget. While, in a vacuum we may wish to get the best numbers possible, in the nitty gritty of the real world, the numbers need to be actionable and timely — so the goal with Recast is to provide actionable numbers at the cadence your business works on.

We see companies most successfully using Recast when they use it in conjunction with both touch-based attribution modeling and experimentation:

  • Touch-based modeling is used for day-to-day channel management and for optimizing creative within a channel.
  • Recast is used for longer-term budgeting and optimizing spend across channels.
  • Experimentation is used to resolve discrepancies between the other two models. This set up allows marketers to take advantage of all of the most powerful features of the different types of marketing measurement systems.

What makes Recast different?

We’ve designed recast from the ground-up to work with companies in a modern retail environment which means omni-channel marketing and omni-channel distribution. Recast is designed to:

  • Work equally well with programmatic advertising channels like Facebook and Search as well as with traditional media channels like TV and print ads
    • Many of the “old school” media-mix-modeling companies do not handle programmatic channels correctly
  • Be updated weekly so that you can respond to changes in channel performance in real-time.
    • Old-school media mix modeling companies generally only perform updates quarterly (if that).
  • Estimate how channel performance changes over time, from week-to-week
    • Old-school media mix models using outdated regression methods estimate a single “impact” for a channel over the quarter. From our experience with channels like Facebook, we know that that is a poor assumption.

We have built Recast from the ground up to address the needs of modern marketing teams, and we would relish the chance to show you how you can save millions of dollars in marketing spend using the Recast system. Reach out to us at or click here to get started.

About The Author