## August 04, 2021

### Variability Associated With Forecasts

I ran a forecast for this business, remember?

So yeah, my prediction of \$25.1 million is likely to frighten the CEO ... she's not gonna like seeing her business revert back to 2019 levels (with some gain due to file power).

Yeah, she's gonna ask me how confident I am in the \$25.1 million estimate? She's gonna be worried about inventory commitments. She's gonna be worried that she can't take \$25.1 million to her Board or to her Owner because they'll crucify her ("are you kidding me? You did more than thirty million for two years in a row"). Being crucified is probably her biggest concern. I know I'd have a hard time selling the likely outcome above to ANYBODY, much less to a Board/Owner.

Each forecast has variability tied to it. I'm making five guesses ... number of buyers, rebuy rate, spend per repurchaser, new/reactivated buyers, spend per new/reactivated buyer.

Therefore, I ran an algorithm that simulated 100,000 outcomes. Each outcome varies the five guesses above (via a normal distribution). Then I produce 100,000 forecasts. Yeah, one hundred thousand. Hard to argue with that many outcomes.

The result of one hundred thousand forecasts?

That's fun stuff!

To simplify the outcome, I broke the outcome down into deciles.
• 10% = < \$23.3 million.
• 20% = \$24.0 million.
• 30% = \$24.4 million.
• 40% = \$24.8 million.
• 50% = \$25.1 million.
• 60% = \$25.5 million.
• 70% = \$25.8 million.
• 80% = \$26.3 million.
• 90% = \$26.9 million.
Realistically, there's an 80% chance that my forecast will yield a result between \$23.3 million and \$26.9 million.

You saw my guesses above. If the CEO doesn't like the guesses, we can plug in different guesses and get different outcomes. But as always, we're dealing with guesses. Your expertise at guessing determines the outcome, and the outcome is likely to vary by +/- \$1.8 million (80% of the time ... varies by +/- \$2.8 million 95% of the time).

Even with the variability, the business isn't forecast to be > \$30,000,000. That's likely to be the sticking point here ... even if we change our guesses, we're still likely to end up with a forecast < \$30,000,000 ... the COVID-bump is over and the business is settling back into a "new normal".

That's the discussion we need to have.

Ok - you are likely struggling to come up with a top-line estimate for 2022. For \$4,900 I'll run my algorithm on your business and we'll simulate 100,000 outcomes and see where we end up. Sound good? Contact me (kevinh@minethatdata.com) and we'll get busy.