When Mike Cochran was consulting with one of the largest cellular companies in the United States, the company’s financial planning and analysis (FP&A) team was convinced a reduction in the number of customers for one of its lines of business was behind a big drop in revenue. But after conducting analytics on the last five years of data for that line, the culprit was something else: a change in the mix of devices and services customers were using.
“Their feeling was that the number of customers was the largest driver to their particular revenue line,” Cochran, principal and managing director of The Hackett Group, said in a CFO.com webinar on streamlining the FP&A process. “It was just a standing, expected thing. But when we looked at the data over time, as you might imagine, there’s so much cannibalism in the cellular space that really the rate at which the number of customers was going down and going up wasn’t material enough to support the revenue swings that we were seeing from the data.”
Technology has made it easier to crunch data to learn what’s really going on with a set of numbers, Cochran said, and what he found gave the company a foundation for addressing the revenue drop.
“Many [customers] have kids with cellphones, Apple TVs,” he said. “Every single one of those devices and every one of those services … is a relevant revenue-generating line for this company. We were able to show that if you can increase the average number of products and services for each household even by a basis point, it was a stronger correlative driver to their revenue.”
Many FP&A professionals believe, like the cellular company FP&A team, they’re doing driver-based budgeting and forecasting, but the term carries different meanings. The true measure of the technique, Cochran believes, is to identify lines in your budget that have an influence on your production, however you define that, and then look at what makes a material difference in your margins and has a high degree of volatility.
In the case of the cellular company, the number of customers impacted production and margins but the change was relatively small, suggesting low volatility. But the change in the services used by those customers was high.
Workforce-related costs are a good example of budget items that often get treated like drivers but in many cases aren't. In many instances, Cochran said, it makes more sense to either take the items out or roll them into a single, higher-level number, and focus your attention instead on what he calls the conversion rate of a line item.
In the case of labor hours, rather than look at the quantity of hours multiplied by the average rate per hour per worker, look instead at how the difference between types of labor ties into your production costs.
“What type of production do we get out of a labor union worker and what type of production do we get out of a non-union labor worker?” he said. “We refer to that as the conversion rate. So, when you take the actual dollar per hour as it relates to the production per hour, you come up with a rate that should be based on volume of production.”
Armed with this information, he said, you have options you otherwise wouldn’t have if your leadership asks you for, say, a 10% cut in costs.
“If you’ve implemented an effective driver-based solution, then you can actually have a discussion that the only way this [10% cut] happens is if we increase, per-hour, the volume output of our union labor force,” he said. “What does that mean for your leadership? Well, it means we need a training program. So, I need a $10,000 investment to train my resources up so they can be more efficient and thereby I can get you the 10% of production that you want.”
What the drivers are, and how many you should include in your forecast, will be unique to your organization based on the business you’re in, the size of your operations, and other factors. To figure them out for yourself, start with your profit and loss (P&L) statement.
“Look at every line of your P&L and interrogate it against materiality and variability,” he said. “If there’s a high degree of materiality and a proportionately higher level of variability, it’s a great candidate to be part of a driver-based forecast.”
As a general matter, the 80-20 rule applies: 80% of your materiality and volatility will come from 20% of your lines.
The best way to deal with the lines that aren't part of that 20%, he said, is to run them through an artificial intelligence (AI) or machine learning (ML) application, which can provide a number for creating a forecast.
You'll want to run the 20% of items that are your drivers through a different kind of technology: predictive analytics. AI and ML can help you look back at your non-driver numbers; predictive analytics can help you look ahead at your driver-based numbers.
What you don’t want to do, he said, is simply take last year’s budget and drop it into your planning tool, update some numbers based on what you think is going to happen, and call that a forecast.
“What happens in a lot of cases is, people will take their entire chart of accounts in their enterprise resource planning (ERP) system, drop it in their planning system, and call it a day,” he said. “That is certainly by many accounts not what we would recommend from a best practices perspective.”
“How Finance Leaders Streamline FP&A” was sponsored by Vena and Argyle.