In theory, forecasting should be quite simple. We have a pipeline of opportunities with close dates in the quarter we are forecasting. We have history to help us understand what proportion will actually close, and what proportion of those will be won. The forecast logically follows, but that generally does not get close to the right answer. The dynamics of our sellers, the competition and the customers are always changing and that means getting an accurate sales forecast is going to require a bit more work. This blog is a condensed version of our white paper that looks at the steps that can be taken to create an accurate sales forecast.
Determine what a forecast is, and what it isn’t
A forecast is a commitment to deliver a certain amount of revenue in a certain period of time. It may be informed by a set of opportunities but the commitment is separate from the opportunities. If it is believed that an opportunity will be won, and it is subsequently lost or delayed, it does not invalidate the forecast, but it might make the forecast harder to deliver. The forecast is not precise about exactly where the revenue will come from, but it’s clear that there needs to be an adequate pipeline of opportunities from which to deliver the forecast, or the possibility to create new pipeline.
Creating the forecast
The forecast will start with an inspection of open opportunities to project how much revenue the opportunities will produce. But there are additional steps that will be required to create a forecast that sales management and the executives can act on with confidence. These additional steps will be managed by various people up through the organization, but essentially they refine and tune the forecast.
Here are the elements that are required to create the best possible forecast:
1. Stay connected to the data
A forecast is made at a point in time, and from that point it already starts to become out of date. That’s why we generally see the same fiscal quarter forecasted a number of times as it progresses and as new information becomes available. During the forecast process the opportunities in the CRM continue to develop, and maybe new opportunities get created—but if the forecast is running off a data extract, in Excel for example, it is disconnected and will rapidly become out of date. So to give the forecast the best chance of success, it has to be connected to the source of data at all times.
2. History is our guide
There are people that are perpetually over-optimistic and there are those that are pessimistic. There are others that would prefer not to disclose everything they know. For these reasons, it’s necessary to interpret a forecast and strip out the bias. Some of this is instinctive and a manager will generally have a good sense of the character of their employees. However, a history of forecasts and how they translated into revenue provides very useful inputs.
An approach that we have found to be useful is to ask the forecaster to indicate how confident they are about winning an opportunity. We keep it simple and ask them to indicate High, Medium and Low. Over time, a pattern emerges. We can use that pattern to interpret every new forecast.
3. The value of human involvement in the forecast
It’s clear that human knowledge and judgement is very valuable in the forecasting process. It simply can’t be the case that everything you need to know is in the CRM. You might see a television news report indicating that your customer is suffering difficult trading conditions, or you might know the procurement officer is on vacation for the next two weeks. Humans just know things that are not encapsulated in CRM data.
One of the most important things we can do in the forecasting process is to include the wisdom of all stakeholders so that the forecast reflects their knowledge and experience.
4. Restrict the forecast to the forecastable
One way we improve the forecast is to identify the “swing” opportunities and exclude them from the forecast. The objective here is to restrict the forecast to opportunities that behave normally, or in a predictable way. A swing opportunity generally has some or all of the following characteristics:
- We can’t predict its outcome with any confidence
- We are not in control of its timing
- It’s significantly larger than regular opportunities
To give an example: Your car dealership sell cars to individuals and they vary in price between $10,000 and $50,000. You sell 25 a month for a total revenue of $750,000. A typical sales cycle lasts two weeks. So far, so predictable, right? You are approached by a new company that has started up in the area. They need 25 cars for their sales people. You think that the deal is worth $600,000 after discounts have been negotiated, but never having been here before, you are less than certain.
This is a swing opportunity. If it happens, it will bring in revenue of $600,000, but that might be in this forecast period or a later period. Its $0 or $600,000. If you forecast it as $0 and it comes in, whatever else is in your forecast is pretty irrelevant since you will have knocked the forecast out of the park. If you forecast it at $600,000 and it doesn’t come in, the forecast is going to be a disaster. This is a swing opportunity.
To be clear, it isn’t that we don’t care about swing opportunities, they are just impossible to forecast. The omission from the forecast should be used sparingly and with good reason. So the designation as swing should be re-examined at the start of every forecast cycle.
We know that forecasting is hard and we also know that it’s important. We want our companies to make the best possible decisions based on the evidence available, and to thrive. The way we will achieve the best forecast is to work with up-to-date data, to collaborate across our organization, and to use historic forecast accuracy to better understand the current forecast and to eliminate bias. As a result, in every forecast cycle we will get better at applying these techniques and our forecasts will improve as a result.
Jess is a communications professional and Vortini’s lead content/web developer. Her current interests lie in the intersection of sales technology and machine learning. In her free time she reads a book-a-week, practices yoga, and is an avid gardener.