Variance & Flux Analysis
Break a variance into the drivers that caused it — price vs volume, rate vs mix — with the narrative written.
★ 25,690 on GitHub0 downloads on LenderSkills
Install it in one line — paste a single command into Claude, ChatGPT or Cursor and it writes the skill into your project.
Free, and about thirty seconds: a work email and a six-digit code.
Downloading shares your name, firm, role, and email with Anthropic. That introduction is what keeps every skill here free.
What this skill does
Takes a variance and breaks it into the drivers that actually caused it — price against volume, rate against mix — then writes the explanation in plain language instead of restating the number.
What it does
- Decomposes a variance into price/volume or rate/mix components
- Applies materiality thresholds so only real movements get investigated
- Writes the narrative explanation behind each driver
- Lays the bridge out as a waterfall from prior period to current
When to use it
- Budget versus actual needs commentary before it goes up
- A line item moved and nobody can say why yet
- You are preparing flux analysis for the close or for auditors
- Leadership wants the driver behind a number, not the number again
What to have ready
- The line item and the two periods being compared
- Actuals for both, plus budget or forecast if that is the comparison
- Volume, unit, or headcount data if you want a true price/volume split
- Your materiality threshold
How to use it
- Download it with the button above. You'll sign in with a work email — that's how the creator knows who is using their skill.
- Copy the install prompt from the "Run it in Claude" box, and paste it into Claude along with the file. Claude installs the skill, fills in your firm's details where the file has
[BRACKETED]placeholders, and runs a quick test with you. - Then just ask in plain English. Claude picks the skill up on its own when a request matches it. Anything it still needs, it asks you for — it won't invent a number.
"Use the Variance Analysis skill on Q4 revenue versus Q3. Both periods' detail is attached."
What you get back
A decomposed variance: each driver quantified, a waterfall from the prior figure to the current one, and written commentary you can paste into the close package.
Good to know
- Output quality tracks input quality. A missing figure is flagged as missing — it is never estimated to fill a gap.
- Every draft is a starting point for your own review, not a finished decision.
- Runs anywhere Claude does: the Claude app, Claude Desktop, or Claude Code.
- Part of the Anthropic Finance plugin. Built for corporate accounting, and it applies directly to a lending shop's own books and to fund-level accounting.
Source & credit
Imported from Anthropic's open-source knowledge-work-plugins library (Apache-2.0). Author: Anthropic — https://github.com/anthropics/knowledge-work-plugins. The author's content is unmodified apart from a legal disclaimer appended at the end of the package. Full credit to the author. Anthropic did not submit this listing and does not endorse it — the original is free at the link above.
Legal disclaimer
This skill drafts; it does not decide. Its output is not legal, tax, accounting, appraisal, or investment advice, and using it creates no attorney-client, fiduciary, or advisory relationship. Nothing it produces is a commitment to lend or an offer to sell a security. Verify every number, date, citation, and conclusion against the source documents, and have a qualified professional licensed in the relevant jurisdiction review the work before you rely on it or send it to a borrower, investor, counterparty, or regulator. Laws, rates, and market data change. The final work product is yours.
Inside this skill
The actual file you get. Members can read it in full before saving it.
name: variance-analysis description: Decompose financial variances into drivers with narrative explanations and waterfall analysis. Use when analyzing budget vs. actual, period-over-period changes, revenue or expense variances, or preparing variance commentary for leadership. argument-hint: "<line item> <period> vs <comparison>"
Variance Analysis
Important: This skill assists with variance analysis workflows but does not provide financial advice. All analyses should be reviewed by qualified financial professionals before use in reporting.
Techniques for decomposing variances, materiality thresholds, narrative generation, waterfall chart methodology, and budget vs actual vs forecast comparisons.
Variance Decomposition Techniques
Price / Volume Decomposition
The most fundamental variance decomposition. Used for revenue, cost of goods, and any metric that can be expressed as Price x Volume.
Formula:
Total Variance = Actual - Budget (or Prior)
Volume Effect = (Actual Volume - Budget Volume) x Budget Price
Price Effect = (Actual Price - Budget Price) x Actual Volume
Mix Effect = Residual (interaction term), or allocated proportionally
Verification: Volume Effect + Price Effect = Total Variance
(when mix is embedded in the price/volume terms)Three-way decomposition (separating mix):
Volume Effect = (Actual Volume - Budget Volume) x Budget Price x Budget Mix Price Effect = (Actual Price - Budget Price) x Budget Volume x Actual Mix Mix Effect = Budget Price x Budget Volume x (Actual Mix - Budget Mix)
Example — Revenue variance:
- Budget: 10,000 units at $50 = $500,000
- Actual: 11,000 units at $48 = $528,000
- Total variance: +$28,000 favorable
- Volume effect: +1,000 units x $50 = +$50,000 (favorable — sold more units)
Reading the full skill counts as a download. The contributor receives your name, company, and email.
Ratings & reviews
No written reviews yet. If you run this on a live file, say how it went. The next lender reading this page is deciding whether to trust it.
Examination record
Run it in Claude
- Copy the install command above
- Paste it into Claude, ChatGPT or Cursor
- It writes the skill in, fills in your firm’s details, and test-runs it with you
Prefer the file? The prompt above is for a package you already downloaded.
Details
More from Anthropic
Imported from their public repository with attribution. Their other skills, and where to find the original, are on their contributor page.