top of page
Search

Doctor StrangeSales or: How I Learned to Stop Forecasting and Love The Odds

Updated: Jun 14

(Apologies to Stanley Kubric.)


"Gentlemen, you can't fight in here! This is the war room."


A priceless line from the classic movie Doctor Strangelove or: How I Learned to Stop Worrying and Love the Bomb.


In the movie, the characters were playing a global game of poker—with the survival of humanity at stake. In sales, it is the survival of your company. Or your role.


What's more important; exactly hitting your forecast, or beating your goal? Hold that thought.



Countless CRM analytics vendors claim that their "AI" will get you hyper-accurate sales forecasts. The boldest claim we have seen is 98% accurate on week one of the quarter. Which I think means the forecast is within 2% of where you will land.

 

Here's the problem with that. Treating your sales like the (un-changable) weather—that you can't control—is admitting that the forecast is not helping you grow your sales. Unlike the weather, you control the outcome. We’ve written about this here.

 

But there’s another issue. A forecast is never exact. It’s not particularly helpful to "call a forecast" number without stating how likely it is. There is a range and a distribution of possible outcomes. How risky is that number? Calling a single forecast number misses this point. Is it certain you will beat it? Or is it a 50/50 chance?


This is why we need to stop making forecast calls and learn to use odds. In sales, we usually want to know if we will exceed a goal. So instead of saying “The forecast is $10M,” you could say “The odds are 50% that we will be over $10M.” Meaning half the time I would expect to exceed that figure. Yes, you might be over $10M but if that’s your goal, you can give guidance on how likely that is.


If you don’t like even odds, fine. You might add that “The odds are 90% that we will come in at over $9M.” And even more specifically, “If we get the handshake on the Acme deal this week, the odds are 90% that we will be over $9.5M.

 

People generally don’t do this because they don’t have the ability to realistically assess these odds. Unfortunately, most analytics vendors focus on replicating time-worn forecasting techniques. Like forecast category rollups. The more advanced ones use an expected outcome—derived by weighting each deal by a probability. But very few of them tell you your estimated odds.

 

These odds tell you a lot about what to expect. An even more nuanced view uses a probability distribution showing the relative likelihood of different outcomes. Figure 1 shows a sample probability distribution. The curve indicates possible sales (horizontal-axis) versus relative probabilities (vertical-axis). That shaded area is an 80% prediction interval. You instantly see the range of what to expect, the most likely outcome (the highest point), the average expected outcome (vertical line), and the shape of the distribution.

Figure 1. A sales forecast probability distribution showing the relative likelihood of different sales outcomes, the average of all possibilities (vertical line), and an 80% prediction interval (shaded area).

That “shape” is important because there are times when the average expected outcome is unhelpful. Like in figure 2—a bi-modal distribution. The average expected outcome is very unlikely—due to one mega-deal. If it comes in, you will be looking at a distribution like that right hump. If not, then it will look like that left hump.

Figure 2. A bi-modal probability distribution. The hump on the right reflects one low probability large deal. Note that there is a very low probability of the average forecast (vertical line) occurring. Averages can be deceiving.

So, stop calling your forecast number. Start using odds.

 

With forecast odds and probability distributions, you’ll know how to assess what is likely and can decide on how you want to report it to your stakeholders. Optimistic. Pessimistic. Balanced. Your call.

 

Want to know your odds? The easy way. Try the free Funnelcast Essentials. We will show you how that works. Or if you have deal probabilities, you can estimate the likelihoods of your possible sales and get your odds of beating a goal. See our video on how to do Monte Carlo simulations in Excel.


You may have several different probabilities to analyze. For instance, the ones:

  • Associated with your sales stages

  • Your sales team manually sets to express their opinions

  • Generated by a third party forecasting application

No problem. Try each one out to see how good they are.

Comments


bottom of page