Sales forecasting is the process of estimating how much revenue a B2B company will make in the future. Many established companies try to predict sales based on historical data, while startups and early-stage businesses without historical data rely on market research and other insights.
The problem with these approaches is they fail to fully account for the unpredictable changes of the future: In 2020, the shift to distributed work and online commerce lead to significant changes in B2B sales. Because of the COVID-19 pandemic, B2B companies now say that digital interactions are two to three times more important than traditional sales interactions to their customers, according to a 2020 McKinsey survey. Another emerging trend is the importance of self-service in closing deals. The same McKinsey survey found that B2B buyers’ preference for self-service has grown in all stages of the buying process since 2016, especially in the research and evaluation phases.
Because current trends and changes aren’t reflected in historical data, a sales prediction algorithm using historical data is unlikely to yield accurate B2B sales forecasts.
How can companies adapt, react, and create B2B sales forecasts that help them make better business and sales decisions? This starts with choosing the best sales forecasting method for their business — and the right tools for sales forecasting. Let’s dive deeper.
How are sales forecasts determined?
There are three main types of forecasting methods: opinion, historical, and prescriptive.
Opinion forecasting
Opinion forecasting is the oldest approach to sales forecasting. The opinions and perspectives of sellers and business leaders are compiled to project how sales will grow in the coming weeks, months, and quarters. The strength of this approach is that it draws on qualitative insights about market characteristics and prospects in the pipeline to structure predictions. The weakness of this approach is the potential for bias — such as wishful or cautious thinking or lack of awareness of some emerging trend — to make the forecast unreliable.
Historical forecasting
Historical forecasting uses the past sales information of a business to identify trends and make predictions about future sales. More sophisticated approaches to this method may also analyze market data to show what’s trending and why. The strength of this approach is that it employs quantitative analysis to detect patterns that are then projected into the future. The weakness of this approach is that it relies on the faulty assumption that the future will be exactly like the past.
Prescriptive forecasting
Prescriptive forecasting, enabled by artificial intelligence (AI), draws in both qualitative and quantitative data, not only including a company’s historical data but also real-time market and third-party data, such as publicly available data about the prospects in a company’s pipeline. Leveraging this networked data, the sales prediction algorithm, powered by a deep learning neural network, makes actionable, accurate predictions to equip sales teams to move the needle. The strength of this approach is that it draws on more data than just one company’s own data, enabling a fuller picture of the future while minimizing the amount of time spent creating and updating the forecast.
Many B2B companies still use a mix of opinion forecasting and historical forecasting to estimate sales and revenue. But relying solely on the events of the past and personal perspective to make such essential predictions can do more harm to a business than good. When the market shifts in unprecedented ways, or buyers’ behavior doesn’t align with expectations, businesses find themselves unprepared, and sales teams feel increased pressure to match the predictions they made under different circumstances.
Leveraging prescriptive forecasting with Collective[i] keeps the sales forecast updated in real time and gives sellers recommendations for steps they can take toward success.
How do you forecast service sales?
Forecasting the future demand for and return on B2B services doesn’t have to require a lot of manual research and calculations. But for companies using older methods, that is the reality. Here are the typical steps an organization may take to forecast B2B service sales:
Capture the sales pipeline
The first step in forecasting B2B service sales is accurately understanding the current state of the pipeline. B2B buyers prefer many kinds of outreach and follow-up, so it’s likely that each seller at a business is using different approaches to close deals. Collecting information about the prospects and the approach each salesperson or sales team is using will help clarify the state of the pipeline and determine the likelihood that deals at different stages will close. This mix of qualitative and quantitative findings will be the basis of the sales forecast.
Differentiate revenue streams
Especially at a B2B service provider, sellers may be working to generate revenue along multiple streams, including gaining new business and upselling existing clients. Once the pipeline has been captured, or perhaps on a routine basis, all prospects should be labeled according to which revenue stream they will be part of. This provides insight into which revenue stream is currently the highest priority for the sales team — the one that includes the most opportunities for growth over the span of the forecast.
Calculate the historical conversion rate
Determine how effective the sales funnel is at converting the leads that enter it. Companies can do this by examining historical data to assess how many leads entered each stage and how many moved down the pipeline. This will also reveal where the most leakage of prospects is occurring and may even reveal ways to improve the funnel.
Apply weighted projections to the current sales pipeline
Once a sales team has a clear understanding of the pipeline to date, it can apply this understanding to the current pipeline to create a sales forecast. The sales team may assume that leads will churn at a similar rate to the past — which means the pipeline can be analyzed even if the only known data is the current number of prospects.
However, it cannot be taken for granted that leads will churn at the same rate as in the past. So a B2B service sales forecast generated according to these steps is unlikely to be highly reliable. Changes in customer expectations or new competitors in the market could impact your prospects’ decisions, whether making them more at-risk or more likely to buy than in the past.
An AI-enabled prescriptive forecasting tool doesn’t just help a sales team identify and react to the present-day circumstances that influence the sales environment — it also automates data collection and forecasting, removing much of the work required to create a service sales forecast from the to-do list of the sales team.
How do you forecast sales for a new product?
Sales forecasting for a new B2B product starts with estimating demand. Quantitative and qualitative data are both necessary to create a product sales forecast. Here are the steps that sales teams historically have followed:
Conduct market research
First, the business must research the market that exists for the product. For instance, in the case of a B2B SaaS product, it’s important that a business research the needs and pain points of its end users through surveys, focus groups, one-on-one interviews, and more. These insights enable the company to start estimating demand for the product and learn about new features or functionality that could strengthen its position in the market.
Predict based on past product performance
Unless a business is a startup or new to product development, sales data should already exist within the organization for similar products. If the company doesn’t have this information, publicly available data about the sales performance of competitors’ products should provide basic insight into the demand for the new offering.
Solicit opinions from experts
Last, new product sales forecasting can be supported by the opinions of company leadership, sales staff, and even outside subject-matter experts in the relevant market vertical(s).
The fact that market and buyer behavior is always changing demonstrates why forecasting should be timely, especially in the early stages of the product launch. If the forecast isn’t regularly updated to reflect user adoption and market changes, the leadership and product development teams will likely make decisions (and investments) based on a forecast that isn’t going to manifest. Collective[i] Intelligent Forecasts™ provides daily automated adaptive forecasts that give everyone the insights they need.
The real-time functionality of a prescriptive forecasting tool can be deeply beneficial to B2B product sales teams. Collective[i]’s forecasts don’t just project the current state of the sales pipeline but also — through Virtual DealRooms™ — recommend actions that may close deals faster and more effectively.
Who should be involved in the forecasting process?
Companies may benefit from involving several departments besides sales in the forecasting process. These could include:
- Marketing: The marketing team can share insights about campaigns to generate leads, such as which audience demographics are engaging with marketing materials and which are not. It may also be able to collaborate with sales on ideas to better qualify leads through content or target messaging to specific demographics.
- Product development: The product development team can provide transparency into customer feedback and share the timeline of new releases and updates. This information can affect the forecast and timelines to close for different clients: Some deals may be dependent upon the release of new features, while others could be jump-started by the news of what’s coming next. This alignment enables the sales team to make more successful outreach and pitches.
- Customer service: Customer service insights are relevant to the existing customer revenue stream. If an existing customer is in the pipeline for an upsell, understanding their history of interactions with the company can help convert their business. Customer service representatives can also share their perspective on why customers churn after the close, helping salespeople capture lasting revenue.
The involvement of the above departments may be helpful in sales forecasting, but involving them is optional. However, the involvement of the sales department is not optional.
- Sales professionals: Sales team members support the forecasting process by updating the CRM to give the most accurate information about the status of all their deals. Seasoned professionals on the team also share their opinions and past experience to help the sales director understand how ambitious or conservative to make the forecast.
- Sales director: The sales director or VP of sales is often responsible for creating the forecast and monitoring whether performance is on track to achieve the projected numbers.
What is sales forecasting software?
Sales forecasting software automates some or all of the manual processes traditionally completed by the sales team. Through machine learning algorithms that mimic the activities of sellers, processes are automated, augmented, and improved. Collective[i]’s Intelligent WriteBack™ automates CRM data capture from any tool sellers use, eliminating manual data entry. Sales forecasting software also runs deep learning algorithms to analyze large, complex data sets in order to drive predictions.
Which algorithm is used for sales predictions?
There are many machine learning techniques for sales forecasting on the market, though not all of them represent an opportunity for turnkey sales transformation.
Exponential smoothing (SES)
One of the most basic sales forecasting algorithms, SES is used by Microsoft Excel in its forecasting function. This algorithm takes the simple moving average of a given time frame and assigns less value to older data. This produces a smooth view of sales trends over time while minimizing the effects of random spikes and drops that might otherwise throw off the forecast.
Autoregressive integrated moving average (ARIMA)
ARIMA is a method of time-series sales forecasting that relies on the historical data at a company to project sales growth. ARIMA allows for inputs such as market data, seasonal sales cycles, and more. This algorithm can produce more reliable short-term forecasts than SES can, but the scrubbing of data that is required to make it work can be tedious for sales teams.
Neural network
Neural networks are the most advanced sales forecasting algorithms available. They are networks of digital neurons that recognize patterns and make connections similar to how the human brain does. Sales forecasting software that leverages neural networks is able to predict future sales with a higher degree of accuracy than sales forecasting software that use a more basic form of AI. The Collective[i] neural network uses deep learning to constantly adjust its predictions according to buyer behavior, information from sales organizations, and publicly available market data, identifying connections and making recommendations that a less-advanced or siloed algorithm might miss.
Painless B2B sales forecasting is possible with Collective[i]
B2B sales must change, adapt, and innovate in response to the real-time changes of the business environment. Real-time insight into the pain points, needs, and preferences of target customers is no longer just a nice-to-have but a necessity to claim the competitive edge, build and nurture relationships, and make quick decisions that are based on facts, not opinions and guesswork.
Collective[i] makes all this and more achievable through end-to-end digital sales transformation. From automated CRM capture to recommendations about how each seller can leverage personal connections, our solution removes uncertainty from B2B sales forecasting and replaces it with human empowerment.
Explore Collective[i] to see what it could help your business achieve.
