How Spreadsheet Checkup turns a workbook into a review plan.
The analyzer performs a static, read-only inspection of .xlsx and .xlsm workbooks. It combines deterministic checks with clearly labeled heuristics, groups related evidence, and prioritizes the result so you can start with the strongest signals instead of searching a workbook cell by cell.
From upload to report
The file is validated before analysis
The public tool accepts .xlsx and .xlsm files up to 10 MB. Package-level safety checks reject unsupported or malformed input before the workbook is handed to the analyzer.
The workbook is opened for static inspection
Spreadsheet Checkup reads workbook structure, formulas, values, styles, names, links, hidden state, and related package metadata. Macro presence can be inventoried, but VBA is never executed.
Rules inspect formulas, structure, and data quality
The analyzer looks for broken references, copied-formula drift, formula interruptions, suspicious totals, risky lookups, circular dependencies, external links, hidden content, inconsistent data types, and other review signals.
Repeated evidence is grouped
Instead of flooding the page with nearly identical warnings, related findings can be grouped so the report stays readable while preserving detailed evidence for inspection and download.
Findings are prioritized
Stronger defect signals appear under Needs Attention. Risk patterns that require context appear under Review Next. Informational and technical findings provide additional workbook context without pretending every unusual pattern is wrong.
You decide what is actually a problem
The report explains why something was flagged and what to inspect next. The final judgment stays with the person who understands the workbook's purpose and business rules.
Deterministic checks vs. review signals
Some findings have direct structural evidence. A formula containing #REF!, a reference to a missing worksheet, or a circular dependency can be identified from the workbook itself. Those findings can be presented with relatively high confidence.
Other findings are contextual. A hard-coded number inside a formula may be an undocumented assumption or a perfectly intentional constant. A hidden sheet may be obsolete or essential. Formula drift may be a mistake or an intentional exception. Spreadsheet Checkup labels these as review signals and includes the evidence needed to make that distinction.
How formula-pattern comparison works
Copied formulas often form a repeated relative-reference pattern. For example, a monthly row might use =B2*B3, =C2*C3, =D2*D3, and so on. Spreadsheet Checkup normalizes relative and absolute references so formulas can be compared by structure rather than raw text.
If one cell suddenly uses a materially different pattern, the analyzer can flag the break for review. This is useful because a formula can return a believable number and still be inconsistent with the surrounding logic.
What happens to uploaded workbooks
The application processes a workbook for the current request and does not provide workbook storage, accounts, or uploaded-file history. Analysis result responses use no-store cache instructions. Workbook report pages do not load Google Analytics.
For a fuller explanation of workbook handling and analytics, see the privacy page.
Important limits of static analysis
Spreadsheet Checkup does not run Excel's calculation engine, execute VBA, refresh Power Query, fetch external workbook contents, refresh pivots, or validate charts and Data Model semantics. It also cannot infer every business rule behind a spreadsheet.
That means a clean report is not a certificate of correctness. The value of the scan is coverage and focus: it can consistently inspect many structural risks and point a human reviewer toward the areas most worth checking.
See the method on a workbook built to fail.
The public demo workbook contains deliberate spreadsheet problems, including formula drift, a hard-coded override, a broken reference, an external dependency, a circular reference, hidden content, and inconsistent data.
Read the demo case study