Why 2019 is critical to MMM

It might be 2022, but we're thinking about 2019.

Why?

It's because of the significance of a 3-year historical period?for Marketing Mix Modeling (MMM).?

The statistical strength of an MMM program relies on the degrees of freedom created by the number of data points in the model. Now that we're entering our second month of 2022, we begin to lose out on 2019 data.

For clients in the consumer packaged goods segment that rely on market composites from NielsenIQ or IRI, we recommend planning your database extractions to take place this quarter.?As POS databases are typically rolling on three-year periods, failure to run extractions now will limit the ability of models to understand pre-covid baselines.?

Take organic search trends, for example, shown in the chart below. ‘How to cook’ has maintained its seasonality over the past 5 years including during covid, while ‘toilet paper’, ‘online shopping’, and ‘zoom’ experienced varying heightened interest.?Without a historical benchmark, it would be more challenging to validate the level of difference attributed to the covid pandemic.

Please send us an email at [email protected] if this applies to your business. We’d be happy to discuss options for data extraction before it’s too late!?