Sign In

Gabriele Franco

Fondatore at Hybrida Marketing

Highlights

Aug 26 · linkedin

⚠️Brands investing > €30k/month in advertising ?

🚨 I genuinely believe most of your companies are wasting 25-30% of their Advertising budget.

The reason this is happening is because current measurement methods are “compromised”.

Privacy regulations and the sunset of cookies are reducing the quality of digital tracking.

With No accurate tracking, digital analytics tools provide misleading measurements that lead to wrong decisions.

These decisions can cost tens or hundreds of $$$

⏱️ It happened to me in 2021, I was managing as performance marketers, one brand investing 1M$/month in advertising.

Their Marketing Mix had 8 different channels.

Google analytics was our primary measurement tool and it showed us a completely unrealistic measurement.

This led us to make decisions completely blindfolded and probably caused tens or hundred of thousands of dollars.

This is why we started Cassandra.

At Cassandra, we combine digital tracking with best-in-class Marketing Mix Modeling (MMM) and Incrementality Testing to truly understand what drives growth.

We turn wasted budgets into growth opportunities by providing incrementality measurements insights.

Unlike traditional tools that only report on PAST performance, our algorithms simulate thousands of media mixes and forecast FUTURE potential revenue and how to get there.

This allows marketing teams to predict the impact of their decisions before implementing them.

We'll share more Brand stories about Marketing Measurement on this page Cassandra

Jun 18 · linkedin

Adopting MMM Right Now?

Follow These Techniques to Build a Trustworthy MMM

  1. Choose the Right Model:

Ensure your model has: ✅ The fewest "0 contribution" channels ✅ Baseline > 0 ✅ Lowest NRMSE ✅ Highest R^2 ✅ Lowest Decomp.RSSD

If you use Robyn, Bernardo Lares has an open source function to automate that part: https://lnkd.in/emjabFA8

  1. Assess Uncertainty: ✅ Examine confidence intervals to understand why they are wide.

Common reasons include:

✅ Insufficient records ✅ Low spend share (<5%): Be cautious ✅ Multicollinearity: Minimize or calibrate affected channels

Conduct holdout/incrementality geo-tests for uncertain channels and calibrate your MMM.

  1. Trust Diminishing Returns: ✅ Compare diminishing returns share to scatterplot distribution using a scatterplot analysis.

  2. Trust Adstock Effects: ✅ Identify the lag effect that maximizes correlation between spend and output. Align this with Upper/Mid/Lower funnel campaigns.

  3. Match with Platform Attribution Data:

✅ Select the model with a lower incremental ROI than the ad platform's reported ROI for lower funnel campaigns.

Google's Marketing Science paper explains it well: https://lnkd.in/edV6vdHD

  1. Backtesting: ✅ Validate forecast accuracy by predicting with trained data and comparing predictions with actual sales.

"Why trust my MMM", is the main topic to increase adoption of MMM.

Software like Cassandra, Recast, Measured, Sellforte, Adtriba are all working towards improving the user experience to prove why you should trust the results automating this diagnosis

Related Questions

What company did Gabriele found?
Where did Gabriele study Economics and Business?
What roles has Gabriele held at Roix?
What startup event does he organize?
Gabriele Franco
Add to my network

Experience

Fondatore at Hybrida Marketing
Startupweekend Marketing Organizer at Startupweekend Latina

Education

Economics, Business, Luiss Guido Carli University

Location

Rome, Latium, Italy