formerly FinancialForce
formerly FinancialForce
Forecasting, already a challenge for traditional services businesses, has grown even more complex with the rise of new services.
How can we as services leaders adapt our forecasting to account for these new revenue streams? In order to best answer that question we need to break forecasting into three categories:
Download our guide: G2 Professional Services Automation Enterprise Report
Service knowns call for the forecasting exercise known as a bottom-up forecast. It’s pretty much just as it sounds: forecasting what we know, meaning anything in our system with a date and a value. Specifically, these knowns would consist of the following:
To be even more specific, service knowns equal the sum of our scheduled backlog plus actuals in a given period. Though this is nothing new, it’s worth noting that any sizable services organization would struggle to calculate its bottom-up forecast without a sophisticated professional services automation (PSA) system.
The term “service unknowns” refers to unscheduled backlog, which can be defined as follows:
Unscheduled backlog can be tricky to calculate given that it is mostly driven by new types of services being sold, which can have minimal or no real plan to deliver. A great example of this would be managed services. A typical managed services deal may consist of four reports, an upgrade, and X hours of consulting time. You may know what all this is valued at, but you won’t have a delivery plan as it’s up to the client to request the work. How do you forecast something you can’t control?
Assuming this isn’t the first time you’ve delivered this type of service, chances are you have a number of these engagements memorialized in your PSA solution. Run a few reports, and you should be able to see how you deliver that type of contract as well as what percentage of revenue was recognized each month for the duration of the engagement. Now that you know what your typical recognition schedule looks like, you can build out a delivery curve for that service. And once you build out the delivery curve, you can apply that to active contracts in your unscheduled backlog.
The delivery curve allows you to forecast how the engagement will be delivered without assigning resources. It’s able to do this by taking the value of the project or contract and anticipating revenue based on the pre-built schedule; delivery curve. A bonus to using delivery curves is it doesn’t require “ghost resources” building out unnecessary schedules for resources and skewing demand, capacity, and utilization reports.
Let’s use the following as an example: Product A was sold for $100,000 and a 12-month contract. With the delivery curve, our forecast could look like this:
Now that the deal is forecasted, using a project start date we can run a forecast at any point in time and capture the project’s anticipated revenue versus just leaving it in our unscheduled backlog. While delivery curves will vary based on type of product or project, the most important thing is that you can turn static, unscheduled backlog into a dynamic forecast of upcoming revenue, taking you one step toward better overall forecasting.
With bookings forecasted, you’ll next want to focus on forecasting sales pipeline. First, that will require isolating the opportunities or products identified as services because we don’t want to forecast the entire opportunity (including licenses and/ or support) as services revenue. Then you need to identify the type of service being sold so we can apply the correct delivery curve—since not all services are delivered at the same rate—and forecast it.
Once you know the types of services in your pipeline, you can apply the same type of delivery curve that we did to our unscheduled backlog. Combined with standard opportunity information, such as close date, weighted probability, and services amount, we know what will be recognized when. This is much more accurate than the traditional “peanut butter spread” forecast so many organizations use. The below example highlights the differences.
In the services world, things can and do change quickly. To account for potential changes in a forecast, it’s common practice to create three different scenarios—typically “Expected,” “Best case,” and “Worst case.”
The following table shows forecast drivers and examples on how to weight them. The different scenarios help us better understand the business and what our results can be, while also giving us a view into what the forecast relies upon. For example, maybe there is a large milestone in the period that is at risk of being delayed. Knowing the impact on the business if it does push helps us better manage expectations and pinpoint what’s needed to make sure it doesn’t push.
Additionally, by leveraging delivery curves—only available with Certinia PSA—you can create an accurate forecast instead of adding plugs or just spreading anticipated revenue. No other commercial PSA tool is capable of modeling how pipeline and unscheduled backlog will be delivered.
The result is what every services organization is chasing: predictability. Though artificial intelligence (AI) systems will very soon provide augmentation, automation, and intelligence to improve this predictability even further, this is the best forecasting achievable today with a best-of-breed solution like Certinia PSA.
Forecasting is just one of the results from integrating your CRM and PSA solutions. The data below from Service Performance Insight (SPI) outlines some key ways that organizations with integrated systems see boosted metrics over those without.
Everything discussed in this ebook relies on data residing in CRM and PSA systems. Assuming a healthy backlog and sales pipeline, this will generate a very reliable forecast for 6-12 months.
Forecasting beyond that will require modeling pipeline and delivery data unavailable through CRM or PSA systems. This is where AI will shine. AI systems like Salesforce Einstein will be able to predict pipeline and backlog for 9-18 months out by evaluating a number of variables ranging from market trends to anticipated performance and growth.
Not only will Certinia incorporate AI for forecasting further out, but we’ll also use it to build better delivery curves since AI can evaluate a volume of elements that would normally require a team of data scientists to consider. Processing data quickly, modeling delivery, and continuing to refine those models will transform how businesses deliver and manage services in the near future.