How to forecast sales: sales forecasting methods and models

Imagine life without meteorologists or technology to help us predict the weather. We’d find it more difficult to plan our daily outdoor activities or our outfits. We may even end up in dangerous situations when severe weather hits unexpectedly.

Similarly, sales teams struggle without an accurate sales forecast or a sales dashboard. They can’t effectively plan to onboard new customers, adjust workload capacity, set achievable goals, or spot critical issues in advance.

At Sell, we’ve worked with thousands of businesses to develop sales processes, build forecasts, and increase sales rep adoption. We’ve learned quite a bit along the way, so we’ve compiled the most essential sales forecasting techniques with examples in this article.

Predict future revenue and get ahead of potential blockers by using a mix of sales forecasting methods that prioritise accuracy.

What is sales forecasting?

Sales forecasting is the data-backed process of predicting how much your company expects to earn over a given period of time. This is based on a number of factors including historical data, industry and economic trends, and your current sales pipeline.

Specific areas that can be impacted include:

1. Preparing post-sales success

   
   Whether you’re selling software or solar panels, there are post-sales activities that need to take place. This can involve purchasing more materials, preparing customer support or developing an implementation timeline. Whatever the case may be, the more accurate you are and the earlier you know your expected sales number, the more prepared your business can be.

2. Faster course correct

   
   Having an accurate view of which deals will close can sometimes paint a bleak picture of missed goals. The earlier you can identify that you will not hit your goal, the faster you can work with marketing to acquire more leads (pipeline) and course correct.

Accurate sales forecasting relies on two critical elements: having the right data and drawing the correct conclusions from it. Neither is easy. If you overestimate sales, you could spend money that you won’t make. If you underestimate sales, you’re left ill-prepared for the coming quarter.

Overall, sales forecasting answers these main questions:

• 
  How much revenue?
• 
  When will that revenue come in?

But just like meteorologists aren’t always spot on with their weather forecasts, sales forecasts aren’t 100 percent correct. So, take these forecasts as predictions—not hard facts. Factors such as marketing, fluctuations in the economy, and hiring or firing employees can all cause deviations.

How to calculate a sales forecast:

Common sales forecasting methods use qualitative and quantitative methods to help you predict total sales, revenue, and new business. Each one takes practice, as well as an objective mindset, to provide your company with accurate forecasts.

5 essential sales forecasting methods

1. Opportunity stages forecasting

Opportunity stages forecasting allows you to calculate the chance of closing a future deal at each stage in the sales pipeline.

Most businesses can break their pipeline down into a general set of stages:

• Prospecting/Incoming

• Qualified

• Quote/Proposal

• Closing

• Won or lost

For example, if you typically end up winning about half of your deals that reach the “Proposal” stage, then you know you’ve got a 50/50 shot for all the deals in that stage during a given quarter.

To use this sales forecasting technique, multiply a deal’s potential by the win likelihood. These numbers can be determined through most customer relationship management (CRM) tools. Next, repeat this process for each deal in your pipeline and add them together.

Let’s say you have a $1,500 deal opportunity at the incoming stage, a $2,000 deal at the qualified stage, and a $1,000 deal at the negotiation stage. Based on the chart above, forecasting would look something like this:

The overall forecast amount for these three deals is $1,400.

One disadvantage to this approach is that it doesn’t take into account individual characteristics of a given deal. This quantitative method is best combined with your sales reps’ opinion on certain deals, so you cover both subjective and objective elements. Your forecast will be more accurate as a result.

2. Length of sales cycle

Forecasting based on the length of your recent sales cycle helps you predict exactly when a deal is likely to close. Rather than analysing success rates based on stage or your sales rep’s gut feeling, this approach makes assessments based on the age of the deal.

For this method, tally up the total number of days it took to close all recent deals. Then, divide that by the number of deals you closed.

Imagine you recently closed five deals. Calculate the period of time it took to close each one, then add up the numbers:

• Deal 1: 62 days
• Deal 2: 60 days
• Deal 3: 59 days
• Deal 4: 55 days
• Deal 5: 60 days
• Total: 296 days

Divide that total by the number of deals (five), and you get your average sales cycle: 59.2 days, or roughly two months.

Now that you know your average sales cycle, you can apply it to the individual opportunities currently in your pipeline. Perhaps a salesperson has reached the “Proposal” stage with a lead after one month—even if this seems like a sure thing, the forecast suggests otherwise. Based on your average sales cycle length of two months, you might predict that the rep has a 50 percent chance of closing the deal. It may take longer than a month for that proposal to actually turn into a win.

3. Qualification frameworks

Forecasting by deal stage is quick and easy, but it can sometimes be inaccurate due to emotion and bias. One strategy for eliminating the impact of emotion on sales forecasting is to use a qualification framework, which generates a score for each deal.

Building out a scoring formula can be used to evaluate the likelihood of a deal closing. A scoring formula is created based on past sales data and then applies the information to predict the likelihood of winning a deal.

For example, in deals with a marketing source, the Referral might be scored at 7.9, while deals that come from Adwords would be scored at 5.1 because Referral deals have historically closed at a higher rate than Adwords deals.

Another qualification framework that we’re fans of here at Sell is MEDDIC. Created by Dick Dunkel and Jack Napoli in the mid-1990s while they were at the legendary sales organisation PTC, MEDDIC outlines six core areas to consider for deal qualification. The team at Lucid Chart has a nice breakdown on MEDDIC here.

At a high level, MEDDIC accounts for:

• Metrics

• Economic Buyer

• Decision Criteria

• Decision Process

• Identify Pain

• Champion

If you’re using a sales methodology like MEDDIC for forecasting, you can assign a point value for simply identifying each criterion. Take this logic and expand it across multiple data points. After scoring each deal, you’ll have a stronger indicator of which deals are likely to be won or lost for your forecast.

4. Regression analysis

Regression analysis provides an in-depth, quantitative assessment of factors that might be affecting sales. Success with this method requires a good grasp of statistics and the factors impacting your company’s sales performance. It also involves calculating the relationships between variables that influence sales.

Regression analysis the most advanced level of forecasting, so it may be more difficult to run and comprehend. But for advanced companies looking to fine-tune their forecasting strategies, this technique can offer valuable information to help with company growth and goals.

The simple regression model equation is Y = a + bX. But let’s break that down. Here’s how you’d go about completing a regression analysis:

Say you want to forecast sales for the next year to plan for budget allocations and determine if more sales reps should be hired. Sales will be your constant, dependent variable (Y)—the factor you’re trying to understand. Now, imagine you want to evaluate how sales calls are affecting your sales. This is your independent variable.

• Dependent variable (Y): Sales (SALES)

• Independent variable (X): Sales calls (SALES CALLS)

You collect data for both your dependent and independent variables over an eight-year period—your annual sales from 2012 to 2020 and the number of sales calls during that time.

Your equation could be SALES = a + b (SALES CALLS), with a representing the intercept and b representing the slope, respectively. Next, use regression software to run the analysis—Excel has this capability. Note that you will not have to compute a or b yourself; the regression software will generate that, too.

You’re looking for the “line of best fit” to approximate the relationship between the variables. For example, your plot might look something like this:

The slope (b) is 0.907, and the intercept (a) is -313.

Based on this model, sales calls look closely correlated to sales and may be leading to more revenue. But remember: just because a variable is correlated doesn’t mean it is the cause. You have to consider a variety of factors too in-depth for this exercise. This is also a simple linear example. You will normally have a multiple linear regression with several independent variables, such as number of emails sent, number of demos given, number of meetings held, etc.

5. Scenario writing

Scenario writing is a qualitative approach used for long-term planning and to account for possible extremes. It is dependent on a subjective understanding of business and sales.

In this approach, you project the likely outcomes based on a specific set of assumptions. You draft several different scenes that could unfold based on the assumptions, say best- and worst-case scenarios for the deals in progress.

Here is an eight-step process for strategically thinking about the planning process for scenario writing:

Let’s say your focal issue is yearly sales. You then move on to key internal factors influencing your sales, such as sales calls, inquiries received, or demo meetings held. External forces that might have an impact are competitors or government restrictions. For critical uncertainties, consider what difficulties might arise over the next year: Will the customer start leaning toward new technology? Will possible government policies affect the nature of your business?

Based on this information, you can begin to develop scenarios. For scenario writing to be effective, plan your potential outcomes around uncertainties with your business, and then create a clear action plan for each one.

Other sales forecasting techniques

• Sales rep classification

  
  This is a qualitative approach to sales forecasting that only looks at the sales rep’s opinion of deals in the pipeline. Rep classification uses your sales reps’ input to identify if a deal is going to close or not.
  To get started, include a forecasting field in your CRM. The expectation would be that any deal in your sales process would have this field to be completed by the sales rep. The forecasting field would contain multiple options to signal what is going to happen to this opportunity in the given time frame (month/quarter).
  Common rep classification categories include:

  • Best Case

  • Commit

  • Pipeline

  • Closed/Won

  • Forecast Stages

  
  While this approach does require that your sales reps give an honest assessment of their skills and potential clients, it can be an effective way to determine if additional steps need to be taken by your reps to close deals, as well as check rep performance.

• Time series analysis

  
  A time series analysis is used when years of a product’s historical sales data is available to identify trends over time. By looking at past data of a certain product or service’s performance, forecasters can set expectations for future performance rates and any recurring changes in rate during the year.
  A time series is a set of chronologically ordered points of raw data. An example of a time series is the set of data points recorded every month over a six-year period of a product’s sales. This type of historical data sales forecasting helps explain:

  • Cyclical patterns in sales repeating every two to three years

  • Trends in the data

  • Growth rates of trends for different data sets

  • Any systematic or seasonal variations in the series of data

  
  One issue with this sales forecasting technique using historical data is that it assumes buyer demand is constant. If anything unexpected happens, your time-series-based model won’t hold up.
  Generally, it’s a good idea to use historical demand as a benchmark rather than the foundation of your sales forecast.

  Use a variety of sales forecast methods for best results

  
  Remember that you aren’t limited to just one technique. People use their smartphone weather apps, watch weather reports on TV, and rely on almanacs and other resources to help determine what to expect from Mother Nature. In a similar way, you can use multiple forecasting techniques and sales reporting tools to get an accurate picture of incoming sales and revenue and evaluate your current sales approach.
  Decide which methods will be most effective for your company, and begin applying them. Don’t get caught up in “analysis paralysis,” either. Although accurate data is important, the aim is for valuable—not perfect—information.
  While using any forecasting technique appropriately takes practice, it will assist you in optimising your sales forecast process and looking to the future.