### This Business Is Cookin' Right Along, Isn't It?

Remember our business from yesterday? Click the image below to take a peek.

This is a business that generated \$88.4 million in demand last year ... and is projected to generate \$92.3 million next year, then \$95.7 million, then \$98.6 million, and then the business grows by one or two million a year thereafter. In other words, the forecast for this business isn't as optimistic in the long-term as it is in the short-term.

Variable Operating Profit, as a percentage of demand, isn't bad ... between 16% and 17% of demand in the short-term ... which will yield a reasonable EBITDA ... so the business is reasonably healthy.

Or is it?

Tomorrow, we will begin to explore how the future trajectory of this business could change.

When I create my simulations, I try to keep the segmentation system simple. I create a statistical model that grades 0-12 month buyers (A / B / C / D / F) ... "A" customers are in the top 20% of the 0-12 month buyer file, "F" customers are in the bottom 20% of the 0-12 month buyer file. Then, I segment customers by 12-month recency chunks. Finally, I have new customers as an additional segment.

For each segment, I measure the following:
• Probability of Repurchase, Next Twelve Months.
• Spend per Repurchaser, Next Twelve Months.
• Probability of Repurchasers Migrating to A / B / C / D / F Segment Next Year.
This simple table of counts, probabilities, spend amounts, and migration patterns yields half of the magic above.

The other half of the magic comes from marketing spend. Customers are frozen as of a year ago, and then the marketing spend allocated against each of those segments is summed. I build laws of diminishing returns for housefile buyers and first-time buyers, allowing me to estimate how response increases as ad spend increases. I use the "square root rule" for housefile buyers ... meaning that for demand attributed to marketing activities, I estimate that a doubling of the marketing budget yields a 41% increase in response. Conversely, halving the marketing budget reduces response by about 30%. For customer acquisition, the diminishing returns relationship isn't as curved, it is flatter.

Tomorrow, we will begin to play with the relationships. This is where the simulation makes a huge difference. We get to see how the business evolves over a five or ten year period of time, given various investment strategies.