Sales Forecasting Models – A Modern Approach to Accuracy
Sales forecast accuracy is the number one challenge of many B2B sales leaders. Not only is it a remarkably difficult challenge to tackle, but it is a crucial driver of business success. Inaccuracy in forecasting immediately leads to over or under-investing in the business, missed earnings estimates, market share losses, and company down-sizing.
If There’s a Sales Pipeline, There’s a Sales Forecast Challenge
The reason that sales forecasting is an eternal challenge is not just that it is attempting to predict buyer behavior with only a small amount of insight into the buyer’s true intentions, but that the forecasting technique is trying to model three related challenges at the same time:
- What is the chance that a given deal will close?
- What is the chance that a given deal will close *by the end of this quarter*, which gets closer each day?
- How do I add up deals, each of which has a different size and different probability to come up with a final forecast number?
This challenge is exacerbated by the fact that each deal is difficult to review and evaluate and sales managers work with partial and imperfect information on each account throughout the process.
Individual Deal Review As Core of Challenge
For an overall sales forecast model to be accurate, each deal in pipeline needs to be understood as accurately as possible. In most sales organizations, this means a significant amount of sales rep self-assessment. There may be some heuristics that are used to guide the rep, such as “have I sent pricing?”, “have they given me a verbal indication that they’d like to move forward?”, etc, but fundamentally, these are self-assessments by the rep involved.
Using rep self-assessment to understand each deal is a flawed foundation to rest a sales forecast upon for three reasons; rep optimism, rep message management, and delays in data.
Sales professionals are fundamentally optimistic people; they have to be given the nature of selling. This often means that they have what many jaded sales leaders would refer to as “happy ears”, a tendency to hear what they are hoping to hear. “This looks great” might be interpreted as an indication that the buyer wants to move forward. “Why don’t you send me some follow-up information” can be interpreted as a deep interest. “Call me in a few weeks” might sound like a commitment to next steps. Sales forecasting techniques that are based on this optimistic foundation can lead to unpleasant surprises.
Secondly, since sales people know that their pipeline is being carefully watched and analyzed by management, they are often savvy managers of the message it conveys. Deals can be tweaked just a bit here and there to make sure that their pipeline looks “just right” for their own purposes, ultimately impacting overall pipeline forecast accuracy.
Third, sales professionals, especially top performers, are busy people. They are out there selling, which is what you want them to be doing. Administrative tasks like manually updating CRM systems, are often delayed, ignored, or only completed with significant management cajoling. Missing data means inaccurate forecasts.
Methods of Forecasting
For all the sophisticated tooling that is deployed in the sales space, techniques of forecasting generally combine elements of three core elements:
Stage-Based: In a CRM system such as Salesforce, deals are assigned (through rep self-assessment, often with a set of heuristic rules) to a “Stage” such as “Discovery”, “Qualified”, “Contract-Sent”, or “Negotiation”. Each stage has a percent associated with it from 0% to 100%. Interestingly, the percentages associated with these stages are often pure guesswork, or were the number picked when the CRM system was first configured. The tell-tale sign of this guesswork is that the numbers are often suspiciously perfect numbers like 10%, 25%, 50%, etc. This approach focuses on the question of “how likely is this deal to close?” and struggles to calculate a forecast for a moment in time like “end of this quarter”.
Time-Based: Slightly more transactional businesses tend to use a more time-based method. They will calculate an average time that deals take to close, or an average number of conversations, and use that as the baseline to map each deal to. Although this approach does slightly better at calculating a forecast number for a moment in time like “the end of this quarter”, individual nuances of each buyer are not taken into account in any way.
Forecast-Stages: A common tweak of the stage-based forecasting technique is to add on forecast stages like “best case”, “commit”, “upside”, etc. This is a sales rep’s moment-by-moment assessment of whether the deal will close in the *current* quarter, regardless of its chance to close at any point in time.
These forecast methodologies compound the challenges of individual deal review and tend to produce a forecast that requires constant massaging, especially towards the end of a quarter, and still remains flawed and highly subject to individual reps’ optimism or subtle tweaking.
Facts as a Foundation for Sales Forecasting
What is needed is a sales forecasting methodology that is based on a foundation of facts. This foundation does not remove the need for sales leadership insights, intuition, and careful probing, rather it grounds that intuition in objective and complete data, and focuses the careful probing on key issues that might impact deals in the forecast.
The facts that can be automatically, objectively, and completely available to a modern sales leader include:
Single-Threaded: Are we only talking to one person, or to more than one person at this account?
Exec Engagement: Have we developed relationships with execs on the buying committee?
Functional Roles: What functions (sales, marketing, IT, legal, finance, etc) have been involved in the conversations about this deal?
Deep Relationships: Who have we developed very strong relationships with?
Team Engagement: How many of our (the seller’s) team members are engaged at this account?
Blocking Roles: When were key blocking roles (such as legal or finance) first engaged?
Growing Engagement: Is our engagement at the account growing broader? How many new relationships have we developed since the last review?
Most Recent Communications: When was our last outreach? When was their last communication?
These facts give insight into both the stage and timing that a deal should be correctly assigned.
From Facts To Forecast
With these facts available, instantly and correctly, across all deals in the pipeline, the technique for sales forecasting evolves to one where stage and timing for deals are grounded in objective evidence.
Given the wealth of tools for calculating, analyzing, visualizing, and reporting on stage-based sales forecasts, it makes sense to keep with that general structure. What we can enhance, however, is the determination of two key decisions: what stage a deal should be in, and what timing is appropriate for the deal to close.
To determine the stage a deal should be in, rep self-assessment can be enhanced by fact-based assessments. For example if the deal is supposed to be “Qualified”, we may want to automatically check that it is not single-threaded. If a deal is in “Negotiation”, we need to see a relationship with an executive at the level of Vice President or higher. If a deal is in “Contract Negotiations”, an interaction with legal and/or finance should be seen on the account.
Looking at the timing of a deal can also become much more accurate than old approaches of the average total deal length. Vagueness about how when a deal “starts” often makes those averages highly inaccurate. However, knowing that legal negotiations usually take a minimum of three weeks is an important metric when a deal is slated to close in three days, and the first communication with legal was yesterday.
With objective data guiding when a deal is placed in a stage, and fact-based assessments determining the close date of deals, you can then look historically at a few quarters of deals to apply an accurate percentage to each stage based on what truly happened. If you started with a gut hunch assessment that 25% of deals in “Negotiation” would close, but only 23% of those did (based on the objective view of what it truly means to be in “Negotiation”), you can update that value in your stage-based pipeline forecasting model.
Intelligent Sales Data Leads to Accurate Sales Forecasting
The foundation of an accurate forecast is accurate and complete data on each deal. To achieve this level of accuracy, you need an intelligent data foundation like Nudge to automatically update your CRM system with the facts that matter on each deal in pipeline. Facts such as:
- Every contact your team is interacting with at the deal
- The strength of each relationship
- The decision-making level of each individual involved in the deal
- The functional role of each individual
- The history of communication between your team and theirs
With this core foundation, you can apply simple techniques to assess each deal against its stage in your pipeline to see if it truly belongs there and build up an assessment of overall pipeline health. You can similarly assess the anticipated close date against the expected back-and-forth interactions with various buying roles to see if it’s likely to close on time.
Today’s sales forecasting techniques are deeply flawed, but very necessary for the performance of the businesses they serve. Rebuilding their foundation on a layer of timely, accurate, and complete data allows sales leaders to enhance their accuracy without losing any time, and that quickly leads to better overall business performance.