Start with the exact claim you want to test
Company guidance is management's stated expectation for a future period. It is useful evidence about that expectation, but it is not a reported result. To read it carefully, preserve the range, its date, the measure and the conditions attached to it. Then compare revisions on the same basis before deciding what changed.
This guide develops a guidance worksheet rather than a general earnings checklist. You will calculate range midpoints, distinguish a higher forecast from a narrower range, and test what an annual outlook implies for the remaining period. All company names and financial figures in the worked exercise are fictional, and all amounts are in millions of dollars.
Begin with a company factor or timeline entry in ThesisMemo and follow an available source link. Keep your calculations in a separate document or spreadsheet. The product's methodology distinguishes reported facts, possible implications and future checkpoints; use those categories to avoid turning management's expectation into your own statement of certainty.
Save the source and the complete definition
Find the original earnings release or outlook update on the company's investor-relations website, and check any associated SEC report. Investor.gov explains that many companies include an earnings release as an exhibit to an 8-K. The release can precede the fuller financial statements in a quarterly or annual report, so retain both when they answer different questions.
Copy the forecast's measure, fiscal period, units, scope and publication date. Record the actual lower and upper bounds rather than copying only a headline midpoint. If the measure is adjusted, preserve that label and the company's definition. A forecast for adjusted earnings is not directly interchangeable with a reported GAAP earnings figure.
Add the assumptions stated beside the outlook, such as currency, acquisitions or product availability, when applicable. Record what the company actually says; do not supply assumptions it omitted. A forecast described as approximately a value or at least a value is not a two-sided range. Leave the missing bound empty instead of inventing one.
Calculate the midpoint without inventing a probability
Suppose fictional Harbor Software initially forecasts annual revenue between 480 and 520. The midpoint is 500, calculated as the sum of the bounds divided by two. The range width is 40, calculated as the upper bound minus the lower bound. Keep all three original values alongside these calculations.
The midpoint is an arithmetic summary. Without an explicit statement from management, it does not establish the most likely outcome, an expected value, or a probability-weighted forecast. The endpoints also do not automatically define a statistical confidence interval. You cannot infer an 80 percent chance of success merely because a company supplied a range.
For a relative measure of width, divide 40 by the midpoint of 500 to get 8 percent. Label this as full range width relative to midpoint, not forecast error. You do not yet have the eventual result needed to calculate an error. Keep your own scenario probabilities in a separate field, if you choose to use them at all.
Compare revisions using both endpoints
Later, Harbor changes its annual revenue outlook to 500–540 on the same scope and accounting basis. The new midpoint is 520 and the width remains 40. Each endpoint rises by 20. The midpoint increases by 4 percent, calculated as 20 divided by the previous midpoint of 500.
A precise research note would say that the stated revenue range moved higher while its absolute width stayed unchanged. It would not say that revenue grew 4 percent: these are two forecasts for the same period, not two reported periods of sales. Actual growth needs a comparable historical result as its denominator.
Now consider an alternative revision, not a later step in the same story: Harbor changes 480–520 to 490–510. The midpoint remains 500 while the width falls from 40 to 20. The low end rises and the high end falls. Calling this a simple guidance increase loses important information; describe the narrower stated range without claiming that underlying business risk was cut in half.
Separate a stronger completed period from the remaining outlook
Return to the first revision, where the annual range rises to 500–540. Assume Harbor has now reported first-half revenue of 230, with no change in revenue definition or business scope. Subtracting 230 from each endpoint implies second-half revenue of 270–310. The implied second-half midpoint is 290.
For comparison, suppose your earlier worksheet paired the original 480–520 annual guidance with your own first-half assumption of 210. That earlier scenario implied second-half revenue of 270–310 as well. The 20 increase in the annual midpoint is fully absorbed by first-half revenue being 20 above your assumption; the implied remaining-period midpoint has not changed.
That is a conclusion about your scenario, not evidence that management previously guided the first half to 210. Label the 210 explicitly as your assumption. Use this subtraction only for comparable additive amounts such as revenue. Do not subtract year-to-date EPS from annual EPS or subtract percentage margins to construct a remaining-period figure; their denominators require separate treatment.
Keep guidance, consensus and your scenario separate
Use three clearly labeled columns: management guidance, any independently sourced analyst estimate, and your own scenario. Record the date and definition of each. If you do not have a reliable consensus source, leave that column unavailable. Management raising its own outlook does not by itself establish that it exceeded analysts' expectations.
Suppose, solely for this fictional exercise, a comparable analyst estimate was 530 before Harbor issued the revised 500–540 range. The new midpoint of 520 is 10 below that estimate, about 1.9 percent using 530 as the denominator, even though it is 4 percent above the previous guidance midpoint. Both comparisons can be true because they use different baselines.
Neither calculation explains a stock-price move on its own. Timing, other disclosures and the quality of the expectations evidence matter. Keep a descriptive statement about the forecast separate from an unverified claim about why investors bought or sold. Do not retrofit a convenient baseline after observing the price reaction.
Translate assumptions into evidence you can revisit
Read management's explanation of the change and the relevant financial-report discussion. Investor.gov describes MD&A as management's perspective on results, trends and uncertainties. Use that context to identify an actual condition behind the outlook, rather than treating the numerical revision as a complete explanation of business demand.
For example, if a hypothetical outlook depends on a planned shipment, write down the stated shipment condition and the disclosure that could confirm completion. Do not treat a launch announcement as proof that the customer accepted the product or that revenue was recognized. If no quantitative sensitivity is disclosed, avoid assigning a made-up revenue amount to the condition.
Your checkpoint should specify the evidence to look for and what it would resolve. One useful question is whether the next release changes the annual range, the completed-period actuals, or both. Another is whether a revised business scope makes the old range incomparable. Preserve uncertainty when the source does not let you separate those effects.
Finish with a versioned guidance record
Save a row for each genuine update instead of overwriting the previous forecast. Include its original source, date, measure, fiscal period, lower bound, upper bound, midpoint, assumptions and any subsequent reported result. Mark management figures, your calculations and your assumptions separately so another reader can reproduce the comparison.
Before connecting the finding back to ThesisMemo, check the company's displayed research date. The methodology describes intended review intervals, not a guarantee that every new release has already changed the assessment. A newly published outlook can be newer than the research you are reading. Keep that timing gap visible in your external notes.
Close with one supported observation and one unresolved question. For Harbor, the revised annual midpoint is higher, while the remaining-period scenario depends on which completed-period figure or assumption you subtract. That distinction gives you a concrete task for the next report. It does not produce a buy or sell instruction or establish a guaranteed outcome.
Product references
Questions & answers
Is the midpoint of guidance the company’s most likely outcome?
Not necessarily. It is the arithmetic average of the endpoints. Unless management explicitly explains the probabilities, do not call it a probability-weighted expectation or attach a confidence level to the range. Preserve the original bounds and their qualifications beside any midpoint you calculate.
Does a higher annual outlook mean the remaining quarters improved?
Not always. A stronger completed period can account for the increase. For comparable revenue amounts, subtract reported year-to-date revenue from each annual endpoint and inspect what remains. Clearly distinguish any earlier personal assumption from a figure management actually provided.
What does a narrower guidance range tell me?
It tells you that the stated endpoints moved closer together, provided the measure and period are comparable. Check whether the midpoint also changed. A narrower published range alone does not quantify the reduction in business risk or prove that the eventual result will fall inside it.
Can I compare guidance with an analyst estimate?
Yes, when the measure, period, scope and accounting basis match and the estimate has a reliable source and date. Keep that comparison separate from the change against prior management guidance. If a comparable estimate is unavailable, do not invent a consensus or describe a beat.
Does ThesisMemo maintain this guidance worksheet automatically?
This guide describes an external worksheet built from original disclosures and available research source links. It does not assume an in-product guidance calculator, complete consensus feed or automatic version history. Use the product’s displayed dates and retain your own record of the sources and calculations.