A Forecasting Tutorial

It was surprising to see the reactions to my post, Pipeline, Tutorial. Mistakenly, I had assumed people understood the basics around pipeline management, how to use the pipeline as a tool for maximizing personal performance, and other things. They don’t, I got a lot of great feedback about the post. But it caused other requests, people started asking for posts on other “basics.”

It’s interesting, we toss these concepts around quite a bit, assuming we have a common understanding of them, the reality is we don’t. As a result, we waste a lot of time, misapplying the concepts.

In the previous post, I talked about the distinction between pipelines and forecasts. Mistakenly, too often, we think they are the same thing. Either we take the “weighted total” of the deals we have in the pipeline, as the forecast, or we take the deals in the closing stage. While those deals might be in the forecast, it’s not just because they are in the closing stage of the forecast.

So this post is a quick tutorial in forecasting.

What is a “forecast?” Per the dictionary it is, “…a prediction or estimation about a future event….” Consequently, in complex B2B selling, it becomes a prediction about a specific deal or opportunity. When we expect to get the PO for a specific deal, and the value of that PO. We may aggregate these individual deals, “We expect these deals…. to close by the end of the quarter…. totaling $X. This is where we run into a lot of problems with forecasting, we think it is about a number, it’s not! I’ll be coming back to this in later sections.

Why is the forecast important? The forecast is our commitment to the rest of the company about what business we expect to produce by what date. Too often, sellers think it’s just something management cares about, only to know what revenue to expect. But it’s much more than that, the rest of the organization is dependent on accurate forecasts to do their jobs. If we build products, others need to know what parts to order, how to schedule manufacturing and logistics capacity to meet the commitments we make in accepting an order. If it’s a services company, we have to make sure we have the right resources and skills available to deliver the services. Not long ago, a major client faced a huge challenge. The projects they had expected and staffed for didn’t come in as forecast. Not only did they have a huge team idled, but the business that was booked required completely different skills–skills that were already booked on other projects. So now, they had a “double whammy” effect on the forecast, even though they had made the number. Even for single product SaaS organizatons, customer experience needs to be prepared to on-board and support new customers.

Who is the forecast important to? Hopefully you know the answer, based on the previous section. We think it’s just something management hassles us with to maximize the revenue number. But the reality is that it’s what drives the majority of the activity for the rest of the company. They need accurate forecasts to be able to to their jobs. Fundamentally, the forecast is a deal that we commit to deliver to the rest of the company. “We will get this order, for this $, on this date….”

The forecast is not about a number! Most of the time, when any of us talk about the forecast, we talk about a number, “We forecast $100M by the end of the quarter….” This isn’t helpful to the organization! What inventory do they need, what manufacturing capability is needed, what delivery capabilities are needed, and so forth. A number doesn’t tell us anything. A forecast is about a deal–what it is, when we expect to get it, what is the size. We aggregate all the deals committed to the forecast for the quarter to come up with a total for the commitments. This, in turn, allows the rest of the organization to deliver on those commitments.

A forecast that focuses only on a number is a certain indicator of under performance! When we focus on the total number, and not delivering on the specific deals we committed to make that number, inevitably we under perform our potential–but making the number masks that from us.

1. Here’s an example, virtually every numbers based forecast looks like this. A very large client committed to a forecast number of $500M. If they made it, they would be on target for their annual number.
   1. At the beginning of the quarter, they forecast a number of deals coming in. They weighted these with various probabilities to assure they would make the number. At the end of the quarter, they made the number. Execs were high fiving each other, proud of their forecast process. I started asking questions:
      1. 27% of the deals they had committed to win, they actually lost! 39% of the deals they committed to the quarter slipped to later quarters. So 66% of the deals they said “Count on us….” didn’t come in. They still made the number, but there were other problems….
      2. 20% of the deals they closed, were not even in the qualified pipeline at the beginning of the quarter. Where did they come from, how come they had no visibility to them. If they thought they could do something, why weren’t they in the original forecast. The remainder of deals they got to make the number were deals in later quarters, they pulled in, usually by offering special terms and pricing.
   2. When we look at this, they were proud of making the number. The reality is because they made the number, but with a completely different mix of opportunities, they significantly underperformed their potential. Why weren’t they forecasting any of those other deals? Why did they lose deals they had committed were ours? Why did so many “trust me, it will hit this quarter” deals slip. All of these point to sales error.
   3. And even worse, because of the churn in opportunities closed, the rest of the organization was completely unprepared to fulfill them. This caused shifts in priorities, resources, and so forth, all increasing their operational costs, reducing their net margins.
2. The forecast is a commitment of a specific deal, closing at a certain value (perhaps range), and by certain date (perhaps a certain range).

Forecasts cannot be based weighted probabilities. This is probably the most controversial thing that I’ll say, but here are the issues with probabilistic forecasts:

1. Usually the probabilities we used are based on probabilities we assign to stages in the sales process. Our CRM systems assign a probability to each opportunity, based on where it is in the pipeline. Prospecting may be 20%, qualifying 40, discovery 60, proposing 80, closing 90. None of these represent a buyer propensity to buy, they represent our progress in through the sales process. To understand how senseless this is, imagine competing with 2 other competitors. All 3 of you are in the closing stage, each of seller is forecasting the deal at 90%. Even elementary statistics informs us this is impossible. Probabilities based on where we are in our selling process are, 90% of the time, bad thinking.
2. Then we win a deal that $1M We don’t win 80% of the deal, or anything else. We win it or lose it.

Forecast must be based on customer confirmed commitments. We increase our confidence in committing an opportunity to the forecast, when customer have made commitments to themselves: They are committed to a change, they know the date they need to have a solution in place to achieve their goals. They understand the consequences of missing that date. Note these commitments have nothing to do with a commitment to us, but a commitment to themselves of achieving certain goals. We then have to assess their commitment to do business with us. They may not tell us this, so we will have to engage them in discussions, where we talk about how they will choose, and their assessment of us. (And sometimes, analytic tools can help us, based on past decisions the customer has made.

Committing a deal to a forecast is a collaborative process between the sales person and manager. Sometimes, managers believe they know opportunities better than their people do. When pressured by their own management, they sometimes forecast an opportunity without the sales person’s agreement. In each opportunity we forecast,

To many forecasts are simply wild assed guessing and wishful thinking: We waste too much time, figuring out how to hit a certain number, retrofitting “data” to support our guess. This doesn’t improve the accuracy of our forecasts, it makes them worse. And it diverts time from our ability to focus on hitting our coals.

No forecast is perfect: As much as we want to make them perfect or as close to it as possible, we will never be perfect. “Shit happens….” The customer may be very committed and something happens on their side, other things may not happen for reasons in or out of our control. So if we need to hit a certain quarterly goal, we must commit to a forecast greater than that goal. While no forecast will be perfect, we can get better at it by paying attention to the issues I’ve outlined.

High integrity, strong pipelines help: While the forecasting process is actually a deal by deal approach. A high integrity pipeline is a starting point and important cross check for committing deals to pipelines.

We seem to spend too much time talking about forecasting–particularly managers: This is purely my opinion, but when I look at the collective management time (and impact on sellers) for forecasting, comparing that to time spent on deal strategies, qualification, how we create value, how we find more of the right deals, how we grow in our accounts, how we grow in our territories, how we leverage our time most impactfully with our customer; we spend way too much time talking about forecasting. If we spend more time on doing these other things well, forecasting and forecast accuracy becomes much less an issue.

Advanced topics: Most of what I’ve focused on his forecasting complex B2B deals. There are some nuances which impact forecasts. I’ll cover a few things:

• Forecasting component parts/embedded products: Imagine you are selling components which are part of an industrial product, medical device, technology product. The sales/forecasting cycle actually has two phases. The first it the design win. This means getting your product specified into the design. Using the thinking outlined above, applies to this phase of the forecasting. The second is the manufacturing implementation. This is a collaborative process, where working with the customer, you determine monthly orders, sometimes for several quarters. Often, we forecast based on contracted run rates. Sometimes we adjust the forecasts based on analytics, trend analysis, seasonality.
• Forecasting highly transactional products: Imagine you are selling highly transactional products. Those were the impact of a single purchase is negligible, when you are looking at forecasting all the potential orders you might receive. For example, whether you get a specific small order from one customer, will not have an appreciable impact on the overall performance. Here, we can’t really forecast on a deal by deal basis, rather we can use analytics, trend analysis, run rates, past customer behavior, and other tools to forecast these volumes. Some SaaS products may fall into this category, though not as many as I suspect some think.

I’ll stop here, I’ve covered some basics. But if you do these well, you can refine and improve what and how you do forecasting. What have I missed?