How PayPal Budgets and Forecasts Revenue

There have been countless books and articles written about PayPal—the “mafia,” the stars who came out of it, and the company’s relentless growth. But what’s never covered? How PayPal actually built its forecasting machine. Because at a payments giant handling over $2 trillion in annual transactions, forecasting isn’t just a finance function—it’s a survival skill.

At PayPal, forecasting wasn't just about predicting numbers—it was about understanding how money moves at scale. Erica Gessert, who ran FP&A at PayPal for eight years, shared a behind-the-scenes look at how the company approached budgeting and forecasting revenue as it scaled.

What makes PayPal unique is that it operates at the intersection of payments, technology, and macroeconomics. Unlike a traditional SaaS business with predictable subscription revenue, PayPal's financials are driven by transaction volume, consumer spending behavior, and economic cycles. This complexity forced the FP&A team to develop sophisticated models that could adapt in real time to changing market conditions.

In this piece, we’ll break down PayPal’s budgeting and forecasting playbook, focusing on three key areas:

PayPal’s experience offers valuable lessons for any finance leader managing growth, uncertainty, and the balance between top-down targets and bottom-up execution. It’s a masterclass in harnessing the chaos of a usage based model.

At the core of PayPal’s financial forecasting was Total Payment Volume (TPV)—the total dollar value of transactions flowing through the platform. Unlike a SaaS business where revenue is largely locked in through subscriptions, PayPal’s revenue was a direct function of consumer spending behavior, merchant adoption, and macroeconomic conditions.

When Erica Gessert first joined PayPal in 2015, the company relied on regression-based models that used historical TPV trends to predict future performance. Over time, however, PayPal built a more sophisticated forecasting engine, incorporating dozens of variables across:

At first, regression-based models worked fine in a steady-state environment. But as PayPal expanded across more geographies, currencies, and business lines, it became clear that a single-variable model couldn’t capture the complexity. Revenue swings in one region might be offset by growth in another, and macroeconomic factors like interest rates or currency fluctuations had a bigger impact than historical trends alone could predict.

To improve accuracy, PayPal developed a predictive model called “Galileo”, which used advanced algorithms to refine forecasts. But despite its sophistication, Galileo had a major flaw—it lacked “explainability.”

Ultimately, PayPal settled on a hybrid approach, using Galileo as a baseline but layering in human judgment and business context to adjust forecasts.

Even at a massive scale, PayPal’s experience proves that no model is perfect - even the super smart ones created in a lab. The best forecasting strategies combine data-driven predictions with human intuition, and they evolve based on macro trends, product shifts, and customer behavior. But at the end of the day, even the most sophisticated models need a simple, common-sense check—take a step back and ask,

Forecasting revenue at PayPal wasn’t just about total payment volume (TPV); it was about understanding how different user segments transacted, how transaction types evolved, and how merchants influenced payment flows. Instead of a single top-line model, PayPal broke forecasting into three key behavioral drivers:

For any payments or transaction-based business, volume alone doesn’t tell the full story. Breaking forecasts down by user engagement, transaction mix, and merchant adoption provides a more accurate picture of revenue trends. And if a model is consistently off, it’s often a sign that user behavior has shifted—requiring new assumptions, not just better math.

Not all days are created equal.

At PayPal, forecasting wasn’t just about how much people transacted, but also when they transacted. Some months had more high-volume days than others, which had a material impact on revenue projections.

One particularly tricky year, PayPal’s Q4 forecast was based on historical holiday spending patterns. But Black Friday fell a full week later than usual, which meant a large portion of spending that typically landed in late November was now spilling into December. That single shift had a measurable impact on month-over-month growth rates—proving just how much the calendar could throw off even the most sophisticated models.

Forecasting models should always normalize for seasonality and calendar effects—otherwise, you’re not comparing apples to apples. But it’s not just about the holiday itself—the days before and after matter too. If Christmas falls on a Wednesday, an auto mechanic shop or B2B payments business might not just see lower volume that day, but also the entire week, as people take extended time off.

At PayPal, forecasting wasn’t a one-and-done exercise—it was a continuous process of adjustment, refinement, and recalibration. Every time consumer behavior shifted, a new product launched, or the economy took an unexpected turn, the FP&A team had to adapt.

One of the biggest challenges was balancing the sophistication of predictive models with the need for transparency. PayPal’s “Galileo” model was powerful, but it operated as a black box—great at forecasting, not so good at explaining why a number moved. The finance team quickly learned that even the best algorithm was only as good as the human judgment layered on top of it. In practice, models provided a starting point, but business intuition provided the final answer.

This became especially clear during major economic shifts. When macro conditions were stable, historical trends and geographic overlays provided the best predictors of transaction volume. But when external shocks hit—such as changing consumer habits—PayPal’s team had to scrap traditional models and rebuild them from the ground up. At one point, home and garden, exercise equipment, and pet stores became the best indicators of transaction volume. The team had to pivot their entire forecasting approach in a matter of weeks.

The calendar was another hidden force that dictated financial results. While SaaS companies collect revenue on a predictable basis, transaction-based businesses live and die by the number of business days in a month. At PayPal, forecasting required not just looking at the actual calendar but also the retail calendar, since shifts in major shopping events could distort year-over-year comparisons. In some months, the difference between hitting or missing a forecast came down to whether there were 19 or 23 business days.

At the end of the day, PayPal’s forecasting success wasn’t about relying on any single model or approach—it was about knowing when to trust the numbers and when to challenge them. Because in a business processing over $2 trillion in transactions, getting it wrong wasn’t an option.

To hear Erica walk through PayPal’s budgeting, check out the podcast.