Skip to the content
logo mainlogo darklogo light
  • Home
  • Solution
    • Core MMM
    • Core Direct
    • Core Lift
    • Core Forecast
    • Core Advisory
  • Method
  • Business Case
  • Research Hub
  • About Us
  • Contact Us
  • EN
    • IT
logo main
  • Home
  • Solution
    • Core MMM
    • Core Direct
    • Core Lift
    • Core Forecast
    • Core Advisory
  • Method
  • Business Case
  • Research Hub
  • About Us
  • Contact Us
  • EN
    • IT
logo main
  • Home
  • Solution
    • Core MMM
    • Core Direct
    • Core Lift
    • Core Forecast
    • Core Advisory
  • Method
  • Business Case
  • Research Hub
  • About Us
  • Contact Us
  • EN
    • IT
HomeCoreThe right KPI isn’t always the easiest one to measure
A white sphere and a light wood cube on a cream background: the choice between two KPIs in Marketing Mix Modeling

The right KPI isn’t always the easiest one to measure

Tracking is changing: which KPI should you choose to continue evaluating media ROI?
A comparison between Upper and Lower Funnel shows why the metric that is easiest to model isn't always the one closest to business value.

The Challenge

The site redesign introduced a new digital user journey and made the mid-funnel KPI underlying previous Marketing Mix Modeling models unusable.

A new metric was needed to evaluate media ROI, without losing continuity and consistency with the historical data. There were two alternatives, both imperfect:

  • Upper Funnel (onboarding and lead opening): large volumes, but a partial view of the conversion path and lower consistency with the historical KPI.
  • Lower Funnel (activation): lower volumes, but closer proximity to the financial outcome and better consistency with the historical mid-funnel KPI.

Our Approach

A purely automatic selection could have rewarded the KPI that was statistically simplest to model. However, simplicity of modeling and business relevance do not always coincide.

We evaluated the two alternatives on two levels. To verify consistency with the past, we compared the series over the overlap period, indexed the curves, and analyzed the correlations with the old KPI. To evaluate business relevance, we then verified which metric best represented the brand’s dynamics by analyzing individual levers separately.

In short: statistics narrows the field, business experience guides the choice.

Measurable Results, Stronger Decisions

Consistency with historical data: in the period analyzed, the Lower Funnel showed a very high correlation with the historical metric, clearly higher than that of the Upper Funnel. This evidence overturned the initial hypothesis and was evaluated alongside the KPI's ability to represent individual business levers.

Readable promotions: the new KPI makes spikes corresponding to commercial promotions visible, providing the model with a metric to analyze their contribution.

Member-Get-Member isolated: the effect of referral programs is separated from other levers.

Physical distribution in the model: in the case analyzed, the Lower Funnel KPI allows for the inclusion of the physical network's estimated contribution in the MMM. With the Upper Funnel KPI under consideration, this contribution would have remained invisible.

A solid model must measure the right value

A Marketing Mix Model can be statistically solid and, at the same time, poorly measure what generates value. Technology is decisive, but it only works when data, model, and business reality stay aligned.

By combining statistical evidence, method, and business logic, management chose the Lower Funnel as the new reference KPI: a shared foundation for the next planning cycles.

A Common Decision Platform

For the Brand Manager: a reliable read on the effect of promotions and Member-Get-Member programs.

For the CMO: evaluate media ROI using a metric closer to the financial outcome. Comparison with the previous KPI maintains a historical reference to guide planning.

For the CFO: a model that also includes the contribution of physical distribution and connects media to the value generated.

Has your tracking changed, and your KPI no longer holds up?

Let’s talk: we can compare available metrics to identify the new reference KPI for your Marketing Mix Model, evaluating consistency with historical data and proximity to business value.

Click here

Date:

November 18, 2026

Category:

Core

Tags:

Business Case
Share
MMM in Assisted SaaS for Ferragamo in the US Market
Core
Costs reduced by 20% thanks to the continuous optimization of Core Forecast
Core
Core Advisory to Protect €300,000 in Business Margin
Core
Subscribe to our newsletter
loader

I hereby declare that I have read the privacy policy of Core Analytics S.r.l. and consent to the processing of my personal data for the purpose of sending newsletters.

Independent Marketing Measurement since 2010.
Turning data into decisions, with you all the way through.

Contacts

info@coreanalytics.it

Social
Core Analytics S.r.l.

Registered Office: Via Monte Rosa 86, 20149 Milano | Operational Headquarters: Via Abbondio Sangiorgio 4, 20145 Milano
Tax ID/VAT Number 07181000964 | REA MI-1941618 | Sh. Cap. 50.000 € i.v.

Privacy policy | Cookie policy