March 07, 2017

Fixing File Momentum

Here's a common outcome across my current projects.

  1. Merchandising missteps have hurt the business.
  2. Marketing cut back on new customer acquisition to "align expenses with demand".
  3. The customer file shrinks due to the interaction between (1) and (2).
  4. Merchandising fixes problems, and merchandise productivity actually increases.
  5. But the business shrinks.
  6. Executives get angry.
This is an example of what I am talking about.


A year ago there were 61,521 twelve-month buyers - today there are 59,765 twelve-month buyers - and as we forecast into the future the quantities continue to decrease. If the business is optimized per current merchandise productivity levels, then the business will continue to shrink. That's what data-driven optimization does to a business.

The best marketers leverage forecasting algorithms that incorporate channel dynamics, merchandising dynamics, pricing dynamics, and discounting dynamics. All are blended together using some sort of "algorithm" - I personally like to use Principal Components Analysis combined with probability of segment movement, your mileage will vary. Given the dynamics of your business, you get an overall forecast (illustrated above) and you get forecasts by marketing channel and you get forecasts my merchandise category.

Best of all, the marketer figures out what kind of marketing investment is required to grow the business again. Look at this example.


We need 7% increases in retention and new buyer productivity to grow the business by 5% this year ... and then to maintain 5% growth, we need increases of 12.5% / 16.8% / 20.6% / 24.0% (vs. this year) to keep growing by 5% per year.

The best marketers use forecasting methodology to communicate to Sr. Management what is needed to grow the business. A 7% increase in retention + newbies is then compared to the organic percentage. If a business generates 50% of volume organically / without marketing, then the marketer knows that there will need to be a 20% investment in marketing spend to generate the 10% increase in demand required among the half of the business that is fueled by marketing. The 10% increase among marketing-centric customers and 0% increase among all other customers yields the 5% increase our situation requires.

Make sense?

Contact me (kevinh@minethatdata.com) for your own Forecasting Worksheet.