“Just how good are your forecasts?” We get asked that question a lot. It’s a deceptivly simple, but often misleading, question because the answer depends upon what you want out of your forecasts—only one of which is how close they come to reality. Before answering how good your forecasts are, you have first ask "what do we want out of our forecasts?"
There are several possible objectives for a forecast.
In part one of this series, we will focus on that first item. Most commonly, people think of a sales forecast as telling you how much you will sell in some period.
Let’s deconstruct that statement. “How much you will sell…” There’s the (first) rub; how much you will sell of what? Are you focused on new business acquisition (the hardest thing to forecast and what most of our customers are interested in), or are you forecasting subscription renewals for a product that has a known attrition rate? Your ability to forecast accurately is a function of what you are forecasting. The weatherperson in San Diego has a much easier job than the one in the northeast corridor.
Moving on, “… in some period.” And there is rub number two. When are you making your forecast? At the start of your month, quarter, year? Midway through? Perhaps you want to wait until all your opportunities have been updated—so, after your business review meetings. And for how long are you forecasting? Until the end of the month, quarter, year? When you start your forecast, and for how long you forecast determine how accurate your forecasts are.
For short-term forecasts at the end of your month or quarter, your sales team pretty much knows what is going to close—without the benefit of any predictive analytics.
How close your forecast comes to reality is of course very important. But comparing a forecast only to actual results is an unfair contrast. Reality is always more accurate. Which brings us to the third rub, you should compare a forecast to your next best alternative. For short-term forecasts at the end of your month or quarter, your sales team pretty much knows what is going to close—without the benefit of any predictive analytics. So, when you look at the accuracy of a forecast you have to ask, “compared to what, and how much better is it?”
If you have a reasonable estimate of what you can expect, then you have a good start on understanding your gaps to plan. But a forecast can also give insights to how much activity you need to fill your gaps. Doing this requires a forecast that takes into account both existing- and prospective new-pipeline as a function of time.
Ok, so let’s say you are looking at the forecast made at the beginning of your quarter for new business opportunities that close in the quarter. If you have a good handle on what to expect, that’s pretty helpful. At the very least, it can help you set and manage expectations.
But as we pointed out at the start, there are other uses for a forecast—beyond telling you what to expect. If you are only focused on forecasting results, then you are missing out on the power of mining your CRM data.
A good forecast model not only shows you what to expect, but why things are what they are, and how to improve matters. This is a major shortcoming of other forecasting approaches.
Traditional classification forecasts (where deals are classified as Commit or Best Case...) are great for setting priorities about what you will work on, but they are typically inaccurate until the end of the quarter. AI forecasts may provide better accuracy, but they provide little to no insight about why you may be short of your goal, where or when to focus your efforts, and how to improve your ability to meet a plan.
Analytics vendors routinely claim 95%+ forecast accuracies. These are nonsensical claims. While we can define a business that might be that forecastable, it would not reflect reality. When it comes to B2B sales, no one knows what the future holds with that kind of certainty.
Funnelcast simply shows you what's reasonable to expect based on recent trends in your data. It does a pretty good job of that and we will put our models up against others. But more important, it shows you when to focus on specific opportunities. And which basic factors—which you control—have the biggest effect on desired outcomes.