What Evidence Does This PDF Provide for the Forecast?

A forecast in a client PDF is a claim to inspect until the document shows how it was calculated and what supports it. A confident number can guide a project in the wrong direction when it rests on an uncited estimate.

When a spreadsheet changed the argument

In 2013, Thomas Herndon, then a graduate student at the University of Massachusetts Amherst, was trying to reproduce results from a widely cited 2010 paper by economists Carmen Reinhart and Kenneth Rogoff. Their work had helped support arguments about the economic effects of high public debt.

Herndon could not reproduce the published result. He requested the underlying spreadsheet, examined it with Michael Ash and Robert Pollin, and found coding errors, selective exclusions of countries, and an unusual weighting method. Their 2014 paper in the Cambridge Journal of Economics showed that the original conclusion did not hold in the clean, simple form many people had repeated.

The important moment came before the correction, when the original finding travelled as a settled fact. A spreadsheet cell range had become part of a major policy argument.

That is the same risk hiding in a client forecast. The PDF may say revenue will rise, adoption will reach a target, or costs will fall after launch. If the document provides no source, model, comparable project, or stated assumption, the number is a starting point for discussion. It has not earned the status of evidence.

Separate the forecast from the support

A forecast can be useful without being certain. The problem begins when a team turns an estimate into a planning input without asking what must be true for it to happen.

As you listen to a proposal, business case, or research report, pause when the narration reaches a number that changes a decision. Ask:

• What evidence in this document supports this estimate?
• Is this a historical result, a target, or an assumption?
• Which inputs would change the outcome most?
• Does the document name a source or show the calculation?
• What happens to the plan if the estimate is wrong?

These questions create a clean distinction between what the document demonstrates and what it hopes will happen. That distinction matters when a forecast appears on page 27, after the executive summary has already made it sound inevitable.

A source-grounded question is especially useful on a commute or between meetings. Instead of reopening the PDF, scanning tables, and losing your place, you can ask about the passage while staying with the document. Adesa lets you upload a PDF or EPUB, listen through it as an audiobook, download the audio, and ask questions grounded in the document as you go.

The question is deliberately narrow: “What evidence does this PDF provide for the forecast?” A useful answer should point you toward the cited material, reveal that no support appears in the document, or make clear that the claim depends on an assumption.

Watch for language that disguises uncertainty

Forecasts often arrive wrapped in decisive language: “expected,” “anticipated,” “projected,” or “will deliver.” None of those words tells you where the number came from.

Listen for the bridge between the claim and its support. A well-supported forecast usually gives you something concrete to check, such as prior performance, a named customer commitment, a market dataset, a pilot result, or a stated model. An unsupported one may move directly from a broad trend to a precise outcome.

Precision can make this harder to spot. “Revenue is projected to increase by 18.4%” sounds more rigorous than “revenue may increase.” Yet a decimal point does not establish a method. Ask what data produced it.

This is also where a document companion can keep a small uncertainty from becoming a large meeting problem. If the answer shows the PDF contains only an estimate, you can flag it before the forecast becomes a budget, hiring plan, or client promise. For more ways to question unclear material as you read, see What Should You Ask When the First PDF Paragraph Stops Making Sense?.

Turn the gap into a better next step

Finding weak support does not require a dramatic objection. It gives the project lead a better next move: label the figure as an assumption, identify the missing evidence, and assign someone to validate it.

A practical note might read: “The forecast appears uncited in the current PDF. Before we use it for staffing or budget decisions, confirm the source data, assumptions, and downside case.”

That language keeps the conversation factual. It also gives the client room to provide evidence that may exist outside the document.

Herndon’s replication did not prove that debt never affects growth. It showed that a heavily repeated conclusion needed closer inspection of the work beneath it. Treat client forecasts the same way. Let the document show its evidence before the number shapes the plan.