Budget vs Actual Template with AI Variance Explanations

Tested prompts for budget vs actual variance analysis template compared across 5 leading AI models.

BEST BY JUDGE SCORE Claude Haiku 4.5 9/10

You need a budget vs actual variance analysis template because month-end closed, the numbers are in, and now someone wants to know why marketing overspent by $47K and why revenue missed by 8%. A spreadsheet that just subtracts actuals from budget isn't enough. You need the structure plus the narrative: which variances are material, which are timing, and which signal a real problem.

This page gives you a working template structure (budget, actual, variance dollars, variance percent, favorable/unfavorable flag) paired with AI prompts that generate the variance explanations for you. Instead of staring at a 200-line P&L writing commentary by hand, you feed the data to an AI and get a draft narrative flagging the drivers behind each material line item.

Below you'll find the prompt, sample outputs from four models, example inputs across different industries, and guardrails for when this approach works and when it doesn't. Copy what fits your close process.

When to use this

Use this when you have structured budget and actual data at the GL or department level and need to produce variance commentary on a recurring cadence. It works best when a human reviews the AI draft before it goes to leadership, and when you can feed the model enough context about the business to avoid generic explanations.

Monthly close commentary for a department P&L or cost center report
Quarterly board package variance narratives across revenue and opex
FP&A drafts before the controller or CFO review pass
SaaS ARR vs plan analysis with bookings, churn, and expansion drivers
Project or program budgets tracked against committed spend

When this format breaks down

Audit-ready variance explanations that require sourced journal entry detail and cannot tolerate AI hallucination
Cash flow variance analysis where timing and working capital dynamics need human judgment the model can't infer from totals
Situations where you haven't given the AI any business context, so it invents plausible-sounding but wrong drivers
Highly regulated reporting like SEC filings where every variance explanation must trace to documented source data

The prompt we tested

You are a senior FP&A analyst producing a Budget vs Actual Variance Analysis for a finance team. Using the company context provided below, generate a complete, ready-to-use variance analysis template populated with the supplied data (or reasonable placeholders if data is missing).

Follow these output rules:
Format the output in clean markdown with clear section headers. Flag unfavorable variances with 🔴 and favorable with 🟢, use parentheses for negative numbers, and round all figures to the nearest dollar and percentages to one decimal. Keep explanations concise (under 40 words each) and business-focused rather than generic.

Company context and financial data:
SaaS company, Q3 2024 results. Budget: Subscription Revenue $2.5M, Services Revenue $400K, COGS $870K, Sales & Marketing $950K, R&D $600K, G&A $300K. Actuals: Subscription Revenue $2.38M, Services Revenue $510K, COGS $910K, Sales & Marketing $1.12M, R&D $585K, G&A $315K. Context: lost one enterprise renewal in August, ran an unplanned paid-ads campaign, and signed two larger-than-expected professional services deals.

Deliver the following sections in order:
1. Variance Analysis Table — a markdown table with columns: Line Item | Category (Revenue/COGS/OpEx) | Budget | Actual | $ Variance | % Variance | F/U (Favorable/Unfavorable).
2. Summary Totals — Total Revenue, Gross Profit, Total OpEx, Operating Income with variances.
3. Top 3 Favorable Variances and Top 3 Unfavorable Variances (ranked by $ impact).
4. AI Variance Explanations — for each material variance (>5% or >$10K), a 1-2 sentence plain-English driver hypothesis (volume, price, mix, timing, or one-time event).
5. Recommended Actions — 3-5 concrete next steps for the finance team and budget owners.
6. Reusable Template Notes — a short instruction block explaining how to re-run this template next period (inputs needed, formulas used for $ Variance = Actual − Budget and % Variance = (Actual − Budget) / Budget).

Model comparison

Model	Quality	Cost	Latency	Tokens
Claude Opus 4.7	9	$0.1600	25.3s	759 / 1982
Claude Haiku 4.5	9	$0.0108	16.4s	539 / 2048
GPT-5.4	Not run yet
Gemini 2.5 Pro	Not run yet
Grok 4.1 Fast Reasoning	9	$0.0006	21.8s	608 / 986

Outputs

Claude Opus 4.7 9/10

# Q3 2024 Budget vs Actual Variance Analysis
*SaaS Company | FP&A Review*

---

## 1. Variance Analysis Table

| Line Item | Category | Budget | Actual | $ Variance | % Variance | F/U |
|---|---|---|---|---|---|---|
| Subscription Revenue | Revenue | $2,500,000 | $2,380,000 | ($120,000) | (4.8%) | 🔴 U |
| Services Revenue | Revenue | $400,000 | $510,000 | $110,000 | 27.5% | 🟢 F |
| COGS | COGS | $870,000 | $910,000 | ($40,000) | (4.6%) | 🔴 U |
| Sales & Marketing | OpEx | $950,000 | $1,120,000 | ($170,000) | (17.9%) | 🔴 U |
| R&D | OpEx | $600,000 | $585,000 | $15,000 | 2.5% | 🟢 F |
| G&A | OpEx | $300,000 | $315,000 | ($15,000) | (5.0%) | 🔴 U |

---

## 2. Summary Totals

| Metric | Budget | Actual | $ Variance | % Variance | F/U |
|---|---|---|---|---|---|
| Total Revenue | $2,900,000 | $2,890,000 | ($10,000) | (0.3%) | 🔴 U |
| Gross Profit | $2,030,000 | $1,980,000 | ($50,000) | (2.5%) | 🔴 U |
| Total OpEx | $1,850,000 | $2,020,000 | ($170,000) | (9.2%) | 🔴 U |
| Operating Income | $180,000 | ($40,000) | ($220,000) | (122.2%) | 🔴 U |

---

## 3. Top Variances by $ Impact

### 🟢 Top 3 Favorable
1. **Services Revenue** — +$110,000 (27.5%)
2. **R&D** — +$15,000 (2.5%)
3. *(No other favorable variances this quarter)*

### 🔴 Top 3 Unfavorable
1. **Sales & Marketing** — ($170,000) / (17.9%)
2. **Subscription Revenue** — ($120,000) / (4.8%)
3. **COGS** — ($40,000) / (4.6%)

---

## 4. AI Variance Explanations

- **Subscription Revenue 🔴 ($120K / -4.8%)** — **One-time event + volume:** Loss of an enterprise renewal in August drove ARR contraction; shortfall aligns with approximate size of a mid-tier enterprise contract.
- **Services Revenue 🟢 (+$110K / +27.5%)** — **Volume/mix:** Two larger-than-expected professional services deals signed in-quarter drove material overperformance vs. plan.
- **COGS 🔴 ($40K / -4.6%)** — **Mix:** Higher services revenue carries lower margin than subscription; delivery labor and implementation costs scaled with the two oversized PS deals.
- **Sales & Marketing 🔴 ($170K / -17.9%)** — **One-time event:** Unplanned paid-ads campaign not included in the original budget; spend was discretionary and executed mid-quarter.
- **G&A 🔴 ($15K / -5.0%)** — **Timing:** Minor overrun, likely timing of professional fees or administrative accruals; not structural.
- **R&D 🟢 (+$15K / +2.5%)** — **Timing:** Small underspend, likely hiring lag or delayed contractor engagement; expect normalization next quarter.

---

## 5. Recommended Actions

1. **Root-cause the lost enterprise renewal** — CS and Sales leadership to conduct a churn post-mortem and build a top-20 at-risk renewal watchlist for Q4.
2. **Measure paid-ads campaign ROI** — Marketing to report CAC, MQLs, and pipeline sourced from the unplanned campaign within 2 weeks; codify approval thresholds for unbudgeted spend >$25K.
3. **Reforecast services margin and capacity** — With PS revenue trending hot, validate delivery margin, utilization, and whether additional PS headcount is needed to avoid COGS leakage.
4. **Refresh Q4 operating income outlook** — Model subscription impact of lost logo, sustained PS momentum, and marketing run-rate; flag if FY operating income guidance is at risk.
5. **Tighten variance thresholds in budget policy** — Require budget-owner sign-off and commentary for any line-item variance >5% or >$25K going forward.

---

## 6. Reusable Template Notes

**How to re-run next period:**

**Required inputs:**
- Budget figures by line item (Revenue, COGS, OpEx categories)
- Actual figures by line item from GL close
- Qualitative context: known one-time events, timing shifts, deal-level notes from Sales/CS/Marketing

**Formulas:**
- `$ Variance = Actual − Budget`
- `% Variance = (Actual − Budget) / Budget`
- **Revenue/favorability rule:** Actual > Budget = 🟢 Favorable
- **Cost/favorability rule:** Actual < Budget = 🟢 Favorable
- `Gross Profit = Total Revenue − COGS`
- `Operating Income = Gross Profit − Total OpEx`

**Materiality threshold:** Flag any line with |%| > 5.0% OR |$| > $10,000 for narrative explanation.

**Refresh cadence:** Run within 5 business days of month/quarter close. Populate the table, auto-rank variances, then collect driver commentary from each budget owner before finalizing.

Claude Haiku 4.5 9/10

# Budget vs Actual Variance Analysis
## Q3 2024 | SaaS Company

---

## 1. Variance Analysis Table

| Line Item | Category | Budget | Actual | $ Variance | % Variance | F/U |
|-----------|----------|--------|--------|------------|------------|-----|
| Subscription Revenue | Revenue | $2,500,000 | $2,380,000 | ($120,000) | (4.8%) | 🔴 |
| Services Revenue | Revenue | $400,000 | $510,000 | $110,000 | 27.5% | 🟢 |
| **Total Revenue** | **Revenue** | **$2,900,000** | **$2,890,000** | **($10,000)** | **(0.3%)** | 🔴 |
| COGS | COGS | $870,000 | $910,000 | ($40,000) | (4.6%) | 🔴 |
| **Gross Profit** | **—** | **$2,030,000** | **$1,980,000** | **($50,000)** | **(2.5%)** | 🔴 |
| Gross Margin % | —— | 70.0% | 68.5% | — | (1.5 pp) | 🔴 |
| Sales & Marketing | OpEx | $950,000 | $1,120,000 | ($170,000) | (17.9%) | 🔴 |
| R&D | OpEx | $600,000 | $585,000 | $15,000 | 2.5% | 🟢 |
| G&A | OpEx | $300,000 | $315,000 | ($15,000) | (5.0%) | 🔴 |
| **Total OpEx** | **OpEx** | **$1,850,000** | **$2,020,000** | **($170,000)** | **(9.2%)** | 🔴 |
| **Operating Income** | **—** | **$180,000** | **($40,000)** | **($220,000)** | (122.2%) | 🔴 |
| Operating Margin % | — | 6.2% | (1.4%) | — | (7.6 pp) | 🔴 |

---

## 2. Summary Totals

| Metric | Budget | Actual | $ Variance | % Variance | Status |
|--------|--------|--------|------------|------------|--------|
| **Total Revenue** | $2,900,000 | $2,890,000 | ($10,000) | (0.3%) | 🔴 Slightly Unfavorable |
| **Total COGS** | $870,000 | $910,000 | ($40,000) | (4.6%) | 🔴 Unfavorable |
| **Gross Profit** | $2,030,000 | $1,980,000 | ($50,000) | (2.5%) | 🔴 Unfavorable |
| **Gross Margin %** | 70.0% | 68.5% | — | (150 bps) | 🔴 Pressure on margin |
| **Total OpEx** | $1,850,000 | $2,020,000 | ($170,000) | (9.2%) | 🔴 Unfavorable |
| **Operating Income** | $180,000 | ($40,000) | ($220,000) | (122.2%) | 🔴 **Critical** |
| **Operating Margin %** | 6.2% | (1.4%) | — | (760 bps) | 🔴 **Critical** |

**Key Insight:** Q3 swung from budgeted $180K operating profit to $40K loss, driven by lost subscription renewal, unplanned marketing spend, and elevated COGS on higher services volume.

---

## 3. Top Variances

### 🔴 Top 3 Unfavorable Variances (by $ impact)

1. **Sales & Marketing: ($170,000) | (17.9%)**  
   Unplanned paid-ads campaign to offset subscription churn drove 18% overspend.

2. **Subscription Revenue: ($120,000) | (4.8%)**  
   Enterprise renewal loss in August created $120K revenue shortfall vs. plan.

3. **COGS: ($40,000) | (4.6%)**  
   Higher professional services delivery costs on two larger-than-expected deals exceeded budgeted delivery mix.

### 🟢 Top 3 Favorable Variances (by $ impact)

1. **Services Revenue: $110,000 | 27.5%**  
   Two larger-than-expected professional services deals closed, driving 27.5% revenue upside.

2. **R&D: $15,000 | 2.5%**  
   Engineering hiring slower than planned; favorable underspend on headcount.

3. **Operating Leverage (Mixed):**  
   R&D and G&A combined $15K favorable offset by $170K S&M overrun and $40K COGS overrun.

---

## 4. AI Variance Explanations

### Material Revenue Variances

**Subscription Revenue – ($120,000) | (4.8%) 🔴**  
**Driver: Volume / One-time event**  
Loss of one enterprise renewal in August reduced recurring ARR bookings below plan. No pricing or mix impact; purely volume-driven churn.

**Services Revenue – $110,000 | 27.5% 🟢**  
**Driver: Volume / Contract mix**  
Two larger-than-expected professional services deals signed, expanding deal sizes and project scope beyond Q3 budget assumptions.

---

### Material COGS Variance

**COGS – ($40,000) | (4.6%) 🔴**  
**Driver: Volume / Service mix**  
Higher professional services delivery (driven by larger deals) required more billable hours and external resource costs than planned, shifting COGS up despite overall GM% still healthy at 68.5%.

---

### Material OpEx Variances

**Sales & Marketing – ($170,000) | (17.9%) 🔴**  
**Driver: One-time event / Tactical response**  
Unplanned paid-ads campaign launched to mitigate August renewal loss and drive Q3 pipeline recovery. This was a deliberate but unbudgeted spend ($17.9% overrun).

**R&D – $15,000 | 2.5% 🟢**  
**Driver: Timing / Headcount delays**  
Engineering hiring lagged plan; new roles not filled until late September, resulting in favorable underspend.

**G&A – ($15,000) | (5.0%) 🔴**  
**Driver: Timing**  
Professional services and insurance costs trended higher; minor and expected in mature finance/legal operations.

---

## 5. Recommended Actions

1. **Enterprise Renewal Deep Dive (Finance + Sales)**  
   Conduct post-mortem on August renewal loss. Quantify customer churn drivers (price, product, support) and adjust Q4 renewal forecast. Assess whether $170K unplanned ad spend will recover $120K revenue loss by Q4 or adjust baseline expectations.

2. **Services Profitability & Margin Pressure**  
   Analyze the two larger deals: confirm 60%+ gross margin on services and profitability after delivery. If delivery costs are structural, revisit services pricing model and resource allocation for H2.

3. **S&M Spend Governance**  
   Establish a $50K+ unplanned spend approval process with CFO to prevent future ad-hoc campaigns. Require forecast impact analysis (pipeline-to-close ratio) before deployment.

4. **Reforecasting for Q4 & FY**  
   Update subscription revenue baseline down $120K (or assess if ad spend recovers it by Q4). Adjust OpEx budget for R&D hiring acceleration into Q4. Re-model operating income target given 760 bps margin compression.

5. **Monthly Variance Tracking**  
   Shift to monthly variance reporting for S&M and subscription cohorts to catch renewal churn and campaign ROI in real-time, enabling mid-month course corrections.

---

## 6. Re

Grok 4.1 Fast Reasoning 9/10

# Q3 2024 Budget vs Actual Variance Analysis

## 1. Variance Analysis Table

| Line Item            | Category | Budget    | Actual    | $ Variance | % Variance | F/U    |
|----------------------|----------|-----------|-----------|------------|------------|--------|
| Subscription Revenue | Revenue  | $2,500K  | $2,380K  | ($120K) 🔴| (4.8)%    | U      |
| Services Revenue     | Revenue  | $400K    | $510K    | $110K 🟢  | 27.5%     | F      |
| COGS                 | COGS     | $870K    | $910K    | $40K 🔴   | 4.6%      | U      |
| Sales & Marketing    | OpEx     | $950K    | $1,120K  | $170K 🔴  | 17.9%     | U      |
| R&D                  | OpEx     | $600K    | $585K    | ($15K) 🟢 | (2.5)%    | F      |
| G&A                  | OpEx     | $300K    | $315K    | $15K 🔴   | 5.0%      | U      |

## 2. Summary Totals

| Metric              | Budget   | Actual   | $ Variance | % Variance | F/U    |
|---------------------|----------|----------|------------|------------|--------|
| Total Revenue       | $2,900K | $2,890K | ($10K) 🔴 | (0.3)%    | U      |
| Gross Profit        | $2,030K | $1,980K | ($50K) 🔴 | (2.5)%    | U      |
| Total OpEx          | $1,850K | $2,020K | $170K 🔴  | 9.2%      | U      |
| Operating Income    | $180K   | ($40K)  | ($220K) 🔴| (122.2)%  | U      |

## 3. Top 3 Favorable Variances and Top 3 Unfavorable Variances (ranked by $ impact)

**Top 3 Unfavorable (by $ impact):**  
1. Sales & Marketing: $170K U  
2. Subscription Revenue: $120K U  
3. COGS: $40K U  

**Top 3 Favorable (by $ impact):**  
1. Services Revenue: $110K F  
2. R&D: $15K F  
*(Only two variances exceed $10K threshold)*

## 4. AI Variance Explanations

- **Subscription Revenue ($120K U, 4.8%)**: Lost one enterprise renewal in August reduced volume; no offsetting price or mix gains.  
- **Services Revenue ($110K F, 27.5%)**: Signed two larger-than-expected professional services deals, driving volume uplift.  
- **COGS ($40K U, 4.6%)**: Higher services delivery volume from new deals increased variable costs (one-time event tie-in).  
- **Sales & Marketing ($170K U, 17.9%)**: Unplanned paid-ads campaign spend spike; timing of execution ahead of budget.  
- **R&D ($15K F, 2.5%)**: Improved project efficiency lowered spend; favorable timing on headcount costs.  
- **G&A ($15K U, 5.0%)**: Minor overrun from professional fees; investigate timing or one-off legal expenses.

## 5. Recommended Actions

1. Partner with Sales to diagnose enterprise renewal loss and adjust Q4 pipeline forecasts.  
2. Review paid-ads campaign ROI with Marketing; cap future unplanned spends at 10% of budget.  
3. Collaborate with Services team to model COGS scalability from new deals for Q4 budget update.  
4. Audit G&A overruns with department leads; implement monthly spend reviews.  
5. Update rolling forecast incorporating these drivers; distribute to leadership by EOM.

## 6. Reusable Template Notes

**To re-run next period:**  
- **Inputs needed**: Budget and Actual $ values per line item (Subscription Rev, Services Rev, COGS, S&M, R&D, G&A).  
- **Formulas**: $ Variance = Actual - Budget; % Variance = ($ Variance / Budget) × 100 (round $ to nearest dollar, % to 1 decimal).  
- Copy-paste data into table; recalculate totals/summaries; refresh explanations/actions with new context. Flag >5% or >$10K as material.

What makes these work

01

Feed the AI business context
A variance table alone produces generic commentary. Include notes about known drivers like hiring events, project delays, price changes, or one-time items. The quality of output tracks directly to the context you provide.
02

Set a materiality threshold
Tell the model to only explain variances above a dollar and percent threshold, such as greater than $25K and greater than 5%. This keeps commentary focused and prevents noise about rounding-level differences.
03

Distinguish timing vs permanent
Prompt the AI to flag whether each variance is a timing difference that will reverse or a structural change. This framing is what FP&A leadership actually wants and what separates useful commentary from narration.
04

Ask for action items
Add 'recommend next step' to the prompt. The model will suggest reforecasts, reallocations, or follow-up questions, which gives the analyst a starting point rather than just description.

More example scenarios

#01 · SaaS company monthly opex variance

Input

Department: Engineering. Budget $850K, Actual $923K, Variance +$73K unfavorable (8.6%). Context: We hired 3 senior engineers in the month vs 2 budgeted. Contractor spend also ran $22K over due to a data migration project extended by 3 weeks.

Expected output

Engineering came in $73K unfavorable (8.6%) driven by two factors. Headcount: one additional senior hire landed in-month, adding approximately $38K in salary and benefits versus plan. Contractor overage of $22K reflects the data migration extension; this should normalize next month as the project closes. Remaining $13K relates to one-time equipment and onboarding costs for new hires.

#02 · Retail revenue variance by channel

Input

Q3 revenue: Budget $12.4M, Actual $11.1M, unfavorable $1.3M (10.5%). Ecommerce budget $4.8M actual $5.2M. Retail stores budget $6.2M actual $4.6M. Wholesale budget $1.4M actual $1.3M. Note: two flagship stores were under renovation for 6 weeks in August.

Expected output

Q3 revenue missed by $1.3M (10.5%), concentrated entirely in retail stores which came in $1.6M unfavorable. The flagship renovation program removed an estimated 6 weeks of trading from two high-volume locations, explaining the bulk of the shortfall. Ecommerce offset partially at +$400K favorable (8.3%), suggesting some channel shift. Wholesale tracked roughly to plan.

#03 · Manufacturing COGS variance

Input

Product line: Industrial pumps. Budget COGS $2.1M, Actual $2.38M, unfavorable $280K (13.3%). Volume was 4% below plan. Steel index up 9% vs budget assumption. Labor hours per unit up 7% due to new assembly line ramp.

Expected output

Industrial pumps COGS unfavorable $280K (13.3%) despite 4% lower volume, indicating unit cost deterioration. Primary drivers: steel input costs running 9% above budget assumption contributed approximately $140K. New assembly line inefficiency added roughly $110K through elevated labor hours per unit, expected to normalize as operators complete training in Q4. Residual $30K in indirect overhead absorption from lower volume.

#04 · Agency marketing spend variance

Input

Client: B2B SaaS. Paid media budget $180K, actual $215K, unfavorable $35K. LinkedIn spend $95K vs $70K budget. Google Ads $85K vs $80K budget. Content syndication $35K vs $30K. MQLs delivered: 340 vs target 280.

Expected output

Paid media overspent $35K (19%), concentrated in LinkedIn at +$25K over plan. The overage is performance-justified: MQL delivery came in at 340 against a target of 280 (+21%), with blended cost per MQL improved roughly 2% versus budget. Recommend approving the reallocation going forward and formally shifting $20K per month from underperforming channels to LinkedIn.

#05 · Nonprofit program budget variance

Input

Program: Youth literacy. YTD budget $420K, actual $385K, favorable $35K (8.3%). Salaries on budget. Program supplies $45K vs $65K budget. Travel $12K vs $28K budget. Two planned site expansions delayed to Q4.

Expected output

Youth literacy program YTD favorable by $35K (8.3%), driven by timing rather than structural savings. Program supplies underran $20K and travel underran $16K because two planned site expansions slipped from Q3 to Q4. Expect both line items to catch up in Q4 as expansions launch; full-year spend likely lands within 2% of plan.

Common mistakes to avoid

Letting AI invent drivers
If you don't provide context, the model will generate plausible but fabricated explanations like 'higher than expected travel costs.' Reviewers who know the business will catch it and lose trust in the whole output.
Skipping the human review
AI variance commentary is a draft, not a deliverable. A controller or FP&A lead needs to validate drivers against actual GL detail before anything goes to a CFO or board.
Mixing units and periods
Feeding the model a jumble of YTD, MTD, and forecast numbers without labels produces confused narratives. Structure the input with clear period labels and consistent units.
Ignoring favorable variances
Teams often only explain unfavorable lines. Favorable variances frequently hide timing delays, missed hiring, or deferred projects that will hit next period. The template should force explanation on both sides.
Over-explaining immaterial lines
Without a threshold, the AI will write commentary on every line including $800 variances on office supplies. Set materiality rules in the prompt to keep output executive-ready.

Related queries

Frequently asked questions

What should a budget vs actual variance analysis template include?

At minimum: line item or GL account, budget amount, actual amount, variance in dollars, variance in percent, a favorable/unfavorable flag, and a commentary column. Strong templates also include prior year actual, forecast, and a materiality threshold indicator so reviewers can filter to significant items.

What is a reasonable materiality threshold for variance commentary?

A common rule is greater than 5% and greater than $10K-$50K depending on company size. For large enterprises the dollar threshold scales up. The point is to avoid explaining noise and focus narrative on items that would change a decision.

Can AI actually write accurate variance explanations?

Yes, but only if you feed it the underlying drivers as context. The AI is good at synthesizing numbers and notes into clean narrative. It cannot infer why a number moved without being told. Treat it as a drafting tool, not an investigator.

How is variance analysis different from a flux analysis?

Flux analysis typically compares current period actual to prior period actual or prior year actual, looking at movement over time. Variance analysis compares actual to budget or forecast. Both use the same template structure and the same AI prompting approach works for either.

Should variances be shown as positive or negative for unfavorable?

Conventions vary. Many FP&A teams show unfavorable variances as negative regardless of whether the line is revenue or expense, because it communicates business impact directly. Pick one convention, label it clearly at the top of the template, and apply it consistently.

How often should I run variance analysis?

Monthly is standard for management reporting. Quarterly is typical for board packages. Some high-velocity businesses run weekly variance checks on revenue and key cost drivers. Match the cadence to how quickly you can act on what the analysis reveals.