Peter, thanks very much to adding to the discussion.

]]>Here is how I present the same to my students and clients in my workbook:http://bit.ly/1qbUiGo:

A Revenue Forecast asserts that a certain amount of revenue will be earned in a certain period of time with a certain probability that the actual result will be within a certain tolerance of the forecasted result.

For example: management may forecast there is a 90% chance of actual revenue being more than 10% less than a certain amount.

% probability of revenue from a source is assigned by management based on their judgment in the face of their collective past experience with similar situations and similar circumstances.

Some managers set their forecasts to equal the Expected Value (Sum of entries each multiplied by their assigned probability of occurring). There are potential problems with this approach in that:

It allows fractional results. For example, $100,000 with a 50% probability of occurring would contribute $50,000 to the forecast even though the actual result will either be $0 or $100,000. Actual Results are more likely to binary because the sale either happens or it doesn’t so fractional results do not occur.

% probabilities assigned often reflect the probability of the revenue ever occurring but do not capture the period in which the revenue will occur. A good approach to forecasting needs to set the probability of revenue in a specific period.

% probabilities assigned often reflect stage of progression through to a sale but do not represent a real assessment of probability. For example, one might assign a 75% probability to all prospects for which a proposal has been submitted because at this point they are 75% of the way through the sale process but it may be, in reality, that only 50% of all proposals are successful.

The way for management to prepare a forecast is to think carefully and critically about each possible sale in order to reach an informed judgment as to whether the projected revenue will happen in the period or that it won’t and then, in the face of that thinking, compute a projection that has the target % chance (where target is probability around 90%) of being no worse than a degree of tolerance (say 10%) below the forecast.

It takes careful, critical and rigorous thinking than most people are willing to invest but when they do, it pays off and they get better at it over time.

]]>Great to hear from you Andy! Great point, we need to focus our energy around winning the deal—not 75% or whatever.

]]>Gary, you cover a lot of things in this outstanding comment. I’ll address a few:

1. For the 3 competitors, each in the same stage of the selling process, each predicting 85% probability of winning–we know that’s statistically impossible. But each don’t necessarily have a 33% chance of winning. One may have 80% and the other 2 10% each. Or one may have 50% another, 30% and another 20%. It just can’t be more than 100%.

2. More than the statistics problem is the issue of not tying the probability directly to the issues having to do with the propensity to buy and their propensity to buy you. To do this, we have to really assess things from the customer point of view—which we know is difficult because of our tendency to be self centered.

3. I strongly believe sales people and managers should be measured on forecast accuracy. Forecasting when we get a deal and how big it will be is very important to the organization. We tend to look at our own performance in isolation from everything else, without understanding the impact on those who count on us.

Thanks for the great discussion!

]]>I recently caught an interview with a professor who teaches risk and probability and how he applied his skills to a life and death health decision; 10-years to live if he did nothing, a transplant that had a 70% success rate that had a 25 to 30% chance of killing him.

He did not make his decision based on the numbers but on the best outcome. When he looked back he deduced, “that probabilities, while useful, are quite limited in their ability to predict what will happen to any one person.” (or sale)

He made a strong point that “We’re never 95 percent alive. We either live or die. *We experience outcomes.* On a population level, I can have 100 people in a room, and some will have something happen to them and some will not. And that’s the hard part because if you happen to be the unlucky one who has that rare event happen to you, you still have the bad thing happen to you in its full awfulness.”

This is true of everything including sales. As I listened to him, I thought about the same point you made about three competitors in the final phase of a sale all thinking they have an 85% probability because of the stage they’ve reached. In reality, their mathematical probability is 1 in 3. Unless you’re selling an inventory item the buyer divides equally among the three vendors, as this professor said about life, you can’t close 95% of the sale, you either make it or you do not. And determining the real probability requires evaluation skills.

David, you and I discussed at great length measuring the soft stuff and the challenge of collecting that type of data from salespeople. One metric we don’t speak enough about that fascinates me is how accurately or inaccurately does each salesperson forecast? When a salesperson is consistently off their opportunity assessment by more than 10%, is it because they are fudging numbers, not understanding what is happening in the sale, not realistic, or something else? Coaching salespeople to have a solid understanding of what is happening or not happening in an opportunity is infinitely more important than forecasting and provides the added bonus of more accurate forecasting.

Isn’t closing sales the ultimate outcome wanted by everyone from the front line to the CEO?

When your salespeople know where they stand in an opportunity and know what to do next to move the sale forward or disqualify and fish in better waters, they’re win rate improves and they’re forecasting improves too. I think there is too much “weight” on forecasting and not enough emphasis on understanding how and why sales are won and lost. What we can best learn from weighted pipelines, IMO, is where to look for opportunities to improve.

]]>