SaaS sales forecasting strategies
Accurate sales forecasting enables CFOs and finance leaders to predict future revenue, make strategic decisions, readjust strategies, and manage risks. However, for a lot of SaaS businesses, creating sales forecasts is tricky, time-consuming, and often inaccurate. Achieving near-to-accurate sales forecasts hinges on selecting the right forecasting methods best suited to your business's specific needs.
This article was sponsored by Drivetrain and delves into four sales forecasting approaches, offering insights, triangulation techniques, and best practices, from a recent webinar I was a part of with Drivetrain. You can also explore these strategies in more depth by watching the webinar — now available on demand.
Factors to consider when choosing a sales forecasting technique for your business
Here are two key considerations that determine the sales forecasting approach you choose for your SaaS business:
Go-to-market (GTM)/sales motion: Your growth strategies – sales-led growth, product-led growth, or a mix of multiple growth engines – will inform how you approach sales forecasting.
Forecast period/horizon: Your approach should also consider the timeline over which you’re forecasting. Are you forecasting for the next month, quarter, or half-year? Or, are you doing more long-term planning/forecasting, say, beyond the next 12 months?
With this in mind, let’s discuss four of the most commonly used sales forecasting approaches — judgment-based, probability-weighted, driver-based, and statistical. While they have their own advantages and disadvantages, applying and triangulating these different forecasting methods can help you arrive at a more accurate forecast for your SaaS business.
Judgment-based sales forecasting
Judgment-based forecasting is done at the sales rep and manager level (can also include more senior executives). In this bottom-up forecasting approach, sales reps usually estimate how much revenue they can close every single month, every single quarter, or perhaps every single year, split out across different regions or by month. This information is then used to build the sales forecast.
The judgment-based forecasting method is well-suited for companies that have products or offerings with large-ticket sizes and opportunities. That’s because sales reps typically spend more time on each individual opportunity and have more qualitative information to inform their projections than if they are responsible for hundreds of leads. Note that if you don’t have much historical data, if your data is incomplete, or if you cannot access your data in a timely manner, this approach may be your only option.
Pros and cons
The obvious advantage of this approach is that the sales rep is closer to the “lead”, especially during the demo meetings, and as a result, has a first-hand understanding of qualitative factors, such as the context, the lead’s eagerness for a solution vs. their budget, and similar nuances/complexities.
“Sometimes the biggest advantage of a forecasting methodology can also be the biggest disadvantage.”
In this case, it is the reliance on the judgment of the sales reps, which could be influenced by their inherent biases, inferences/ideas driven by emotions, external pressures, etc., that can lead to inaccuracies in your forecast.
Probability-weighted sales forecasting
The probability-weighted forecasting method takes a more quantitative, top-down approach to forecasting. It involves weighing the deals based on their position in the pipeline and marking them less heavily or more heavily as revenue.
For example, let’s say a lead was generated yesterday, and they are interested in (but not fully committed to) the product. The projected revenue of this lead will be multiplied with a lower probability factor compared to a deal that the sales team is confident about closing tomorrow.
Pros and cons
The advantage is that since the probability-weighted method is data-driven, it avoids almost entirely human interventions and biases (completely opposite to the judgment-based approach). It is also quick to update and adjust since it doesn’t involve one-on-one discussions with sales reps and among teams to validate the information. Whenever there are any changes, those are applied to the rates almost immediately.
Sales managers might be underappreciated in this method, however. The probabilities need to be revisited when sales processes undergo changes, or there are more experienced sales managers on the team with additional information that could improve the accuracy of your forecast. However companies can be a bit slow in reviewing the probabilities, leading to either over forecasting or under forecasting since the changes were not reflected or considered in time.
“There's a very thin line between making your forecast more accurate and spending too much time, and not making it accurate enough by not spending enough time.”
Driver-based sales forecasting
Driver-based forecasting focuses on identifying and then building a model around the key drivers that influence your sales.
With the driver-based forecasting method, your first step is to select the factors ("drivers") that you believe most strongly impact your ARR. Technically, you could choose to forecast with any number of drivers. The key is to select five to 12 of the highest-value drivers, looking closely at your business to make sure you’re not missing any.
Here are a few of the most common drivers:
Number of new customers: New customers mean more ARR. And to the extent that you’re able to retain them (i.e., reduce your churn rate to improve logo retention), they can be the most significant growth driver for SaaS businesses.
Contract Values: ACV can be an important driver impacting your ARR, especially if you’re able to increase the price of your plans or upsell existing customers.
Churn rate: Understanding and reducing churn in your business can have a huge impact on your ARR due to its negative compounding effects. Even a seemingly small churn rate can exponentially grow your ARR losses over time.
Quota attainment: The number or percentage of deals you expect your reps to close, including the turnaround time (sales velocity) as they move through the pipeline is also regarded as a key revenue driver.
Pros and cons
This method offers flexibility and helps you with scenario planning, but it requires a deep understanding of your business.
Note that there is a risk of compounding error with this method. You need to be very careful when you add several drivers on top of one other, as each driver’s margin of error will get multiplied. Your aim should be to ensure that the “overall” error added to the sales forecasting model is minimized.
Statistical sales forecasting
As the name suggests, this approach uses statistical models, such as linear regression, to predict trends or sales based on historical data.
Pros and cons
It is a particularly useful method when you want to calculate ARR or MRR based on consumption or usage. Additionally, this method works quite well when you triangulate it with the judgment-based method.
“Judgment-based methods have human biases, but they take nuances and intricacies into account. But statistical methods — they don't have any or [very few] human biases, only those related to the input into the model.”
The main downside of using the statistical method is the availability of data — for example, companies might have slower sales cycles or generally less data for implementing this method — along with the quality of data (which may not be very good).
How to decide which forecasting method is right for your business
The sales forecasting method(s) you choose depends on a few key factors and the kind of organization you have. In this section, we’ll explore some key considerations along those lines.
1. Sales-led enterprises with high ACV and low deal volume
If you work in an enterprise space with high ACV deals, using a judgment-based forecasting method is the best choice. This is because your sales reps are probably spending a lot of time closing each deal.
“If you have very big deals, then you’d want to spend more time on each individual deal.”
Judgment-based forecasting helps you get more accurate predictions for the next quarter. However, the method tends to become significantly less accurate as you make predictions beyond the next quarter - more so than other forecasting methods.
2. Mid-market / SMB sales-led companies with a well-established customer base
The best method for sales-led companies with a well-established customer base in the mid-market or SMB space can vary depending on what you’re forecasting – existing pipeline or new business.
Forecasting existing pipeline: You can use probability-weighted forecasting techniques where you can leverage historical conversion rates to estimate the likelihood of closing deals. This data-driven approach provides a standardized and objective view.
Forecasting new pipeline: Statistical methods like moving averages or regression analysis are better suited for predicting new pipeline generation (if you have large amounts of historical data). These techniques leverage trends in historical data to estimate future pipeline growth.
3. Mid-market/SMB sales-led companies with clean, reliable CRM data
If your company has clean and reliable CRM data, a sales/quota capacity model can prove helpful for forecasting within a 12-month horizon.
Using a sales capacity model is a different way of doing a driver-based forecast that brings more data into the process. This method works well because the business is in a steady and stable state.
“You can be fairly confident that [your drivers] will work out with a certain accuracy based on certain historical values.”
4. Prosumer companies implementing a product-led growth motion
For prosumer companies implementing a product-led growth (PLG) strategy, it makes sense to look at the demand generation funnel. This is because these types of companies typically attract a large number of leads, which can provide robust data to inform the sales forecasting effort.
A driver-based forecasting method is often the most appropriate approach for these companies because it allows you to identify and quantify the key drivers of customer acquisition and account for additional growth factors like product changes/improvement.
“[With PLG companies] there is a large volume of data because there are a large number of customers that you're typically signing with smaller ticket sizes, a large number of people that flow through the funnel. So, a lot of statistical techniques also become very interesting here.”
5. Companies with hybrid or multiple growth engines
When you have a hybrid or multiple growth engine, you may need to consider a combination of forecasting techniques to get the best results. It’s a good idea to experiment with different methods to see what works best for your unique business. The key is to tailor your approach to the specific dynamics of each growth engine.
Take the example of Squarespace. White it is primarily a hosting service for consumers, it also has a B2B arm with a sales team that targets companies who want to purchase multiple websites. Given these two varied target markets — where each is a growth engine with its own unique drivers — it makes more sense to consider different forecasting methods for better outcomes.
How to successfully apply these forecasting methods at your company
Effectively applying sales forecasting methods is as important as understanding how they work. Here are some tips to make your revenue forecasts more accurate:
1. Use triangulation
It may prove risky to rely on a single forecasting method. That’s where triangulation comes in. Triangulation is the practice of using multiple forecasting techniques in parallel.
“Triangulation is a way to stress-test your forecast and improve the accuracy.”
You combine the advantages of different forecasting methods for a more robust and accurate prediction. For example, you can combine judgment-based forecasts with statistical models to balance human insights with data-driven objectivity. You can even combine the top-down approach (typically aggressive) and bottom-up approach (typically conservative) to arrive at a more nuanced middle ground.
2. Eliminate forecast bias
Every forecasting method is susceptible to bias, which can lead to inaccuracies in your predictions. To mitigate different biases, it is important to recognize their source. These include:
Political pressure: Internal pressures can influence forecasts, leading to overly optimistic or pessimistic predictions.
Statistical bias: Assuming that past trends will continue can be misleading, especially in dynamic markets.
Confirmation bias: Cherry-picking data that supports your pre-existing beliefs and ignoring contradictory data or information will lead to inaccurate results.
Narrative fallacy: Correlation does not equal causation. When you see a pattern in your data, it can be tempting to construct a narrative to explain it, even if there’s no real connection. Being aware of this potential bias can help you avoid it.
Once you identify the biases affecting your forecasts, you can measure them and then take the necessary steps to mitigate those. You can learn more about how to identify and eliminate these biases in your forecasts by tuning into the webinar.
Ready to take a deep dive into SaaS sales forecasting techniques?
SaaS sales forecasting requires a strategic approach, and there are a lot of financial software tools on the market today to help you create reliable, data-driven forecasts. Most financial forecasting software and revenue planning tools will have features that can help with different aspects of your sales forecasting.
The software best suited for your organization should ideally be a comprehensive and robust financial planning and analysis (FP&A) platform, one that offers integrations with all the business systems and apps you need to pull data from to inform your forecast.
Combining the right technology with the best forecasting method or mix of methods will give you more accurate and actionable predictions. To understand what the right combination might be for your unique business, check out this webinar where I discuss the different SaaS sales forecasting approaches, triangulation techniques, and use cases in detail with Drivetrain, a leading FP&A software solution.