One of the key strengths of Recast’s model is that it operates within a fully Bayesian framework. Unlike traditional models, a Bayesian approach requires explicit assumptions from the outset. While all models rely on assumptions, Bayesian models stand out by allowing those assumptions to directly align with the knowledge and context that marketers already have.
Objective of Model Discovery
The goal of the Pre-Build Discovery working session is to understand how your business operates so Recast can structure the model accurately. Discovery ensures the final model reflects how you view your marketing program.
During the session, we’ll map your funnel by working backward from your KPI, defining each step of the journey, typical time lags, and channel bucketing.
We’ll walk through live exercises on the call:
Conversion Funnel Discovery
Every organization uses different language to describe the stages a customer moves through, from first touch to conversion and beyond. We want to understand your terminology so we can adopt it in our communications and ensure the model reflects how your business works.
We’ll also discuss how long customers typically spend in each stage of the journey. Examples of funnel terminology:
-
Awareness → Consideration → Conversion
-
TOFU → MOFU → BOFU
-
Reach → Act → Convert → Engage
Channel Taxonomy
Once we understand your customer journey, we’ll explore how your channels fit in, bucketing them into the identified stages.
Examples
-
“Our affiliate program targets customers in the BOFU stage.”
-
“Linear TV is used to drive awareness.”
-
“Facebook ASC is optimized for people in the consideration or conversion phases.”
Channel Details
In the last part of the session, we’ll may confirm remaining configuration details. These include:
-
Whether channels saturate. (Recast’s default assumption is, yes all channels saturate. However, this allows us to toggle this effect on/off for some channels)
-
Whether the impact of channel spend is realized fully on the day of spend.
-
What do marketers actually pay for when buying inventory on this channel? (e.g. impressions, clicks, conversions, spots)
-
Is there a delivery lag between when spend is recognized and the absolute earliest an impression can be seen? (e.g. Direct Mail spend recognize on send date)
-
How spend is reported to Recast.
Who Should Join?
We’ve found smaller groups here tend to result in a more productive conversation, as long as the group is composed of stakeholders who can speak to both strategy and structure across marketing channels. This group is critical in de-risking model setup so that the outputs of your first model align with how the broader team thinks about decision making, and insights can flow easily through the organization.
How should I prepare for the Discovery Session?
Since the focus of this call is strategy, we don’t need attribution reports or performance dashboards. However, it is helpful to come prepared with:
-
A general understanding of how long it typically takes customers to move through your decision funnel
-
How you think about your channels relative to the different funnel stages 3. Which channels you expect to behave similarly
-
Additional Details
a. The cost-basis of your channels (clicks/conversions/etc),
b. How spend for channels is recognized (execution date, conversion date, drop date, invoice date, smoothed)
c. Whether a channel saturates/cost efficiency diminishes (generally yes, w affiliate being a common exception)
d. Whether a channel has delivery lag from the time of spend (physical mail an example of where impressions cannot be physically seen until days after initial send/spend recognition)