Judgmental forecasting: methods, uses, examples

Judgmental forecasting is the practice of predicting future outcomes using expert opinion, structured intuition, and qualitative reasoning instead of (or alongside) statistical models. Analysts lean on it when historical data is thin — when the market is new, or when conditions have shifted so sharply that the old numbers no longer apply.

The textbook Forecasting: Principles and Practice by Rob J. Hyndman and George Athanasopoulos puts it bluntly: statistical models are usually better when good data exists, but judgmental methods become essential when launching new products, navigating policy shocks, or pricing something the market has never seen before.

You’ll see it inside business intelligence teams, BPO operations leads sizing next quarter’s headcount, and C-suites debating a rebrand. The trick is structure. Done well, judgmental forecasting beats gut feel by a wide margin — done badly, it’s just opinions in a meeting.

How it works

Judgmental forecasting works by converting expert knowledge into a structured prediction. The forecaster frames a question, gathers domain experts, applies a method that controls for bias, then combines or weighs the inputs into a single estimate.

The structure matters because raw human judgment is noisy. People anchor on the first number they hear, give too much weight to recent events, and quietly adjust their answers to match the room. A good method blocks those leaks.

Six methods dominate practice. Each one suits a different shape of problem.

Two names sit behind the most-cited method. The Delphi method was developed at RAND Corporation in the 1950s by Olaf Helmer, Norman Dalkey, and Nicholas Rescher to forecast Cold War threats. It still works because it anonymises panellists, runs in rounds, and feeds back the group’s range before each new vote.

Scenario building has a similar pedigree. Pierre Wack pioneered the technique at Royal Dutch Shell in 1971, and the company’s preparedness for the 1973 oil shock turned scenario work from academic curiosity into mainstream corporate practice.

Examples

Real teams use judgmental forecasting whenever the spreadsheet runs out of road.

Apple’s iPhone variant launches. When Apple ships a new iPhone size or category, product marketing leans on analogy forecasting. The 2014 iPhone 6 Plus was sized against existing phablet sales from Samsung. The 2020 iPhone 12 mini was modelled on iPhone SE adoption curves. In both cases the prior product’s data carried the prediction.

Pfizer and BioNTech, 2020. When the COVID-19 vaccine candidates entered late-stage trials, Pfizer’s commercial team had no historical demand curve to lean on. Internal demand forecasting reportedly used a Delphi-style panel — epidemiologists, supply-chain leads, and country managers to size first-year doses, an effort widely covered in industry press through 2021.

Royal Dutch Shell, 1973 onward. Shell’s scenario-planning unit famously sketched a “high oil price” world before the 1973 oil shock. When OPEC’s embargo hit, Shell adjusted faster than rivals and climbed from the seventh-largest oil company to the second by the late 1970s.

Philippine BPO capacity planning. Manila-based contact-centre operators routinely use composite forecasts when a Western client launches a new product. A client-side analogy forecast pairs with a Delphi round across team leads to set ramp targets for the first 90 days, then a statistical model takes over once live volume data arrives.

Related terms

• Business intelligence: the broader discipline of turning operational data into decisions, of which judgmental forecasting is one tool.
• Forecasting: the umbrella practice covering both judgmental and statistical methods.
• Delphi method: a specific structured technique using anonymous expert rounds.
• Scenario planning: building plausible future “worlds” to stress-test strategy.
• Predictive analytics: the data-driven cousin that runs on statistical models rather than expert opinion.
• Workforce planning: a common downstream use case, sizing teams against forecasted demand.
• Capacity planning: another downstream application, matching infrastructure to predicted load.

FAQ

When should you use judgmental forecasting instead of statistical forecasting?

Use it when historical data is missing, unreliable, or no longer representative. New products, new markets, regulatory shocks, and long-horizon strategy calls all qualify. Where good data exists, statistical models usually win.

Is judgmental forecasting accurate?

It can be, when structured. The Delphi method, Cooke’s method, and scenario planning all consistently beat raw expert intuition because they suppress anchoring, groupthink, and recency bias. Unstructured “ask the boss” forecasting tends to perform worse than a simple statistical baseline.

What’s the difference between the Delphi method and a regular meeting?

A meeting lets the loudest voice anchor the room. Delphi runs in anonymous rounds with controlled feedback, so panellists revise toward the group’s range without knowing who said what. That structure is what produces convergence rather than capitulation.

Can you combine judgmental and statistical forecasting?

Yes, and you usually should. Composite forecasting blends both, and judgmental adjustments to statistical baselines are standard practice in supply chain and finance teams. The combined forecast almost always outperforms either method alone.

What are the biggest risks in judgmental forecasting?

Cognitive bias is the headline risk — anchoring, overconfidence, availability bias, and confirmation bias all distort expert opinion. Poorly chosen panellists, leading questions, and political pressure inside the company compound the problem. Structured methods exist precisely to limit these.

Who actually does judgmental forecasting inside a company?

Usually a mix of business analysts, business intelligence specialists, product managers, and senior operators. In larger companies it’s a named role inside FP&A or strategy. In outsourced operations, the BPO partner often runs the forecast collaboratively with the client.

Need a forecasting-capable analyst on your team without the in-house overhead? Browse vetted outsourcing partners on Outsource Accelerator to find business intelligence and analytics talent ready to support your next planning cycle.