If someone tells me their biggest problem is forecasting, I congratulate them.
After all, if forecasting is really the problem, they’re focused on optimizing their revenue growth process. That means the core aspects of their business must be in great shape.
And, of course, my congratulations are typically short-lived, because they quickly realize they are still grappling with their core aspects—and they believe that forecasting will solve those issues.
Forecasting gets a lot of attention because its promise is so appealing, but far more often than not, the reason a company is unable to forecast effectively has little or nothing to do with its forecasting process or the technology underlying it. The problem is far more likely to be related to their go-to-market strategy, sales process, or overall performance.
There are only two reasons to focus on forecasting:
- The most common is to increase predictability for revenue and to communicate that to the appropriate stakeholders
- A less common, and more powerful purpose, is to better create the triggers or insights earlier to make an adjustment to either fix a problem before it becomes too costly or to take advantage of an unexpected opportunity
Why do so few businesses excel at forecasting? (According to a recent report, nearly 80% of companies admit they are bad at it.)
And the dominant approach to sales forecasting, where every deal in a pipeline is assigned a probability percentage based on its current funnel stage, simply isn't the answer.
In fact, I'll go so far as to say traditional staged-based probability (where the stage drives the probability) is bullshit.
The current approach produces flawed insights and reinforces weaknesses. It also removes accountability and learning opportunities for salespeople, leaving them to play an increasingly passive role. (For a deeper dive into why this is so, read The Problem with Forecasting.)
Why Forecasting Fails
1. No one asked about the job to be done. Always ask yourself why and what you're forecasting. Sometimes, when people are trying to forecast sales, they're actually trying to forecast revenue.
2. Companies treat forecasting deterministically. Forecasting is probabilistic. Even in the best circumstances, forecasting is not precise. It provides a range—and if your numbers are small, your variance will be large. (More later on what to do in that case.)
3. You can’t “set it and forget it.” Accurate forecasting takes discipline. Most people rarely, if ever, revisit their forecasts—and if they do, they typically examine only the total number. To improve your accuracy, you must periodically revisit and assess each element for accuracy.
4. Forecasting is often a scapegoat. Some companies think they have forecasting issues when they're just in denial about a shortfall. If a company is 50 percent below goal, that's not actually an issue with forecasting.
Several months ago I was working with a company that had been referred to us by one of our clients. The company had been acquired a couple of years earlier by a PE firm, with the strategy of making them the lead in acquiring other companies in their industry. The Chief Revenue Officer was under increased pressure from the board because, according to the CRO, their forecasts properly reflected the strength of their efforts. She insisted there was a problem with the CRM because the forecasts weren’t right.
After digging in, we quickly discovered the forecasts were actually quite accurate. The problem was that their go-to-market efforts and sales process weren’t effective. Sure, they were generating 20% more MQLs than their plan and model called for, but those MQLs weren’t turning into high-quality opportunities.
Unfortunately, she didn’t want to take on the uncomfortable task of leveling with her board or do the hard work of fixing their processes. She continued to insist that it was a reporting and forecasting problem, and she chose to go with a firm that would work with her on “fixing the forecasts.” Six months after we lost that opportunity, I learned that she was fired.
5. It creates artificial pressure. Many companies see forecasting as the equivalent of reporting earnings to Wall Street, where the number must always increase. That means reps are pressured to move opportunities to "commit" before they're ready. And that creates a vicious cycle of inaccurate forecasts.
So, it's clear that traditional forecasting has issues for many reasons, including that companies have a lot of misconceptions about the best way to use it.
How can you use forecasting to increase predictability? Here's what we do at Lift:
1. Determine the job you want forecasting to do. If you're looking for predictability, ask yourself why. It's usually because companies want to better allocate their limited resources—and in sales, the limited resources are usually people, time, and effort. In that case, you may benefit more from scenario planning, which means trying to anticipate various situations, how they could affect your business, and what you can do about them. That's also your best option if you're one of the companies I mentioned earlier with small numbers and a lot of variance.
2. Close the forecasting loop. Examine what went right and what went wrong. If you won a deal forecast at 15 percent, why? If you lost a deal forecast at 80 percent, what went awry?
Most companies only have discussions if the reps committed to something that didn't close, but it's important to examine all the outcomes. If you want better predictions, you must teach salespeople how to improve.
3. Leverage learnings to get better at interventions. What actions should the rep take if a deal isn't likely to close? Should they opt out or change course? Conversely, if they believe a deal will close, they should also consider what could potentially prevent a good outcome. That way, they can have a plan ready, should they need one.
4. Use a forecast confidence property. We started using this at Lift after I read Thinking in Bets by Annie Duke, which posits that one of the most powerful things anyone can do to improve the quality of their decisions and actions is to assign a confidence level. In other words, if a rep thinks they'll win a deal, they should also share their confidence level. I did a small test, asking reps to rate their confidence in predicted wins. It was lower than I expected, so I created this rubric:
This property has helped us direct our efforts effectively and see improved payoffs.
Why? It's human nature to underestimate our low probabilities. We also overestimate our large probabilities.
One of the problems in sales is there's so much noise because there's so much emotion. If you have a bad call, you tend to downgrade the opportunity based on feelings, not logic. In my experience, these ranges are broad enough to counter that tendency.
If you’d like to see an example of this property in use, I put some examples in this blog.
5. Add a "Forecast to Close By" field to your CRM. Normalize your forecasting by using a set timeframe. At Lift, we watch quarters, using a trailing 90 days, but timing depends on each company's overall strategy.
This field is a simple yes/no field, answering one question: "Do you forecast this opportunity will be completed by (chosen timeframe)?" In Lift's case, if it is likely to close in 90 days, it's a yes. If not, it's a no.
In my experience, the traditional CRM field, "Predicted Close Date," is basically a rep's wild guess, which is liable to change many times. In fact, I'm not sure I've ever seen anyone get this date right. In my experience, a date range increases predictability.
6. Revisit your forecasts. A rolling timeframe also means reps must regularly check in to see if things have changed. It also allows you to begin to see patterns.
One reason that companies don't do this is they're trying to create artificial performance by pressuring reps to commit and driving various behaviors that work in the short term. In some cases, this has a high cost associated with it, including high unpredictability.
A reminder: many companies often see a major increase in sales variance right before they see growth. Familiarity with patterns helps rein in panic.
Forecasting can be very effective, but it's not a set-it-and-forget-it process. It requires attention and continuous improvement. It also requires realistic expectations. You can't pressure reps to commit prematurely to opportunities and then blame them for a shortfall. If you want more predictability, forecasting is a fantastic tool for prioritizing opportunities and planning.