Guidance scorecards: compare results on the original basis
A company can report results above an old guidance headline after changing what the metric excludes. The useful test is whether the actual result meets the original range on the original basis. Build a numerical comparability bridge before describing management as having exceeded its earlier commitment.
Freeze the measurement promise
SEC staff interpretations discuss how inconsistent non-GAAP adjustments between periods can mislead unless changes and reasons are explained. This article uses US non-GAAP disclosure as a source context. Its scorecard is an original research method, not a regulatory definition of beating guidance or a determination of compliance.
Record the first range, issue date, target period, metric definition, currency convention, business perimeter and stated assumptions. Keep revisions as separate observations. A narrower later range provides information, but it should not erase what the company originally said or silently change the basis used to evaluate the outcome.
A hypothetical adjustment bridge
Assume a fictional company originally guides to annual adjusted EPS of 2.00–2.20. Its original definition permits adjustments totaling 0.35 in the eventual reported year. Actual GAAP EPS is 1.70. The company later adds another 0.25 exclusion, reporting adjusted EPS of 2.30 under its new definition.
| Hypothetical measure | EPS |
|---|---|
| Actual GAAP EPS | 1.70 |
| Adjustments permitted on original basis | +0.35 |
| Actual result on original basis | 2.05 |
| Additional exclusion under new definition | +0.25 |
| Reported result on revised basis | 2.30 |
The reported 2.30 exceeds the old upper bound of 2.20, but the comparable 2.05 is inside the original range. Against its original midpoint of 2.10, the comparable difference is −0.05, approximately −2.38%. Label the result as within the original range rather than claiming an original-basis beat.
This calculation assumes that the adjustments are additive and the share denominator is compatible. In a real EPS reconciliation, tax effects, diluted shares and continuing-operations scope can prevent a simple subtraction. Use the issuer’s reconciled components, and mark any unavailable original-basis amount as unresolved.
Separate revisions from operating delivery
Suppose the company revised guidance to 2.20–2.40 using the new definition. Reported EPS of 2.30 equals that midpoint. The scorecard can therefore say both that the latest range was met and that original-basis actual EPS was 2.05. These are different comparisons, not contradictory conclusions.
A change in acquisition scope or constant-currency basis needs its own bridge. If management originally excluded an acquired business and later includes it, newly consolidated earnings are not automatically evidence of better performance in the original business. Likewise, a favorable exchange-rate movement should not be described as operating delivery against a fixed-rate assumption without compatible evidence.
Keep absolute targets and growth-rate targets distinct. If guidance expresses growth against a base year that is later restated, the displayed growth rate can change even when the target amount does not. Preserve both the original base value and the target amount so a denominator revision is not mistaken for a change in operating delivery.
Test the full commitment set
Review revenue, margin, cash and investment guidance separately rather than selecting only the most favorable metric. EPS can meet a range while cash conversion or project completion misses its stated objective. Avoid combining these different outcomes into an uncalibrated management credibility score.
The worksheet should include original guidance, subsequent revisions, actual reported result, comparable result, adjustment evidence and remaining unknowns. Observable triggers include a definition change, a missing reconciliation, a business disposal or a switch from annual to quarterly comparison. Each calls for a precise comparability check before a success or failure label.
A forecast is conditional, so also preserve the original assumptions. The scorecard can identify a miss without proving that management acted dishonestly. Conversely, a formally met range does not establish forecasting skill if the metric changed or the range was extremely wide.
FAQ
Should the latest guidance replace the original range?
Keep both. The latest range tests current communication, while the original range tests the earlier measurement promise under its stated conditions.
Can I rank management by how often it beats the midpoint?
Not responsibly without comparable definitions, range widths, assumptions and a suitable sample. The first deliverable is a defensible comparison, not a precision credibility rating.