Margin Levers

Analyze Free

START HERE if you have no Margin Levers account. Analyze up to 250 B2B SaaS customers for free — no API key needed. INPUT: raw CSV text with id and revenue columns (cost optional — pass business_model to estimate from SaaS benchmarks). Minimum 10 customers, maximum 250. SUPPORTED HEADERS: revenue/mrr/arr/monthly_revenue/sales, cost/cogs/monthly_cost/expense, id/customer_id/account_id, name/customer_name/company_name. EXAMPLE WITH COST: "id,name,revenue,cost\n1,Acme,120000,45000". EXAMPLE WITHOUT

Analyze

START HERE if you have customer revenue and cost data as JSON and want to understand profitability. Segments customers into A (most profitable) through F (most unprofitable) using whale curve methodology. INPUT: JSON array of customer objects, each with {id: string, revenue: number, cost: number}; name is optional. Minimum 2, maximum 50,000 customers. Revenue and cost in USD (non-negative). OUTPUT: segments A–F with customer assignments, peak profit point, tail drag percentage, and summary stati

Analyze Csv

Run a customer profitability analysis from raw CSV text. INPUT: a string containing CSV with a header row and data rows. Must include revenue and cost columns; ID and name columns are auto-detected if present. Supported header names: revenue/mrr/arr/sales, cost/cogs/expense, id/customer_id/account_id, name/customer_name/company_name. Supports US ($1,234.56) and European (€1.234,56) number formats. Example: "id,name,revenue,cost\n1,Acme Corp,120000,45000\n2,Beta Inc,85000,38000". Maximum 5 MB, 50

Get Analyses

List the authenticated user's saved profitability analyses, paginated. INPUT: optional limit (1–100, default 20) and offset (default 0). OUTPUT: array of saved analyses, each with id, name, customerCount, totalRevenue, and createdAt. WORKFLOW: use the returned id as analysisId when calling getInsights, runScenario, or generateBoardDeck. Analyses are created by calling analyze or analyzeCSV with persist=true.

Get Insights

Answer 'what should I do about my profitability?' — generates AI-powered insights for a saved analysis. PREREQUISITE: an analysisId from analyze(persist=true), analyzeCSV(persist=true), or getAnalyses. INPUT: analysisId (integer) and type — one of: executive_summary (C-suite narrative overview), segment_deep_dive (detailed per-segment breakdown), action_plan (USE THIS to get prioritized next steps — the most common agent intent), custom_question (ask any specific question about the data — provid

Run Scenario

Answer 'what would happen if I repriced, restructured, or dropped these customers?' — simulates specific customer changes against a saved analysis. PREREQUISITE: an analysisId from analyze(persist=true), analyzeCSV(persist=true), or getAnalyses. INPUT: analysisId and a changes array. Each change has: customerId (string), action ('reprice' = adjust price up/down, 'restructure' = change delivery model, 'drop' = remove customer entirely from book of business), and optional revenueChange/costChange

Get Benchmarks

Answer 'how do I compare to similar B2B SaaS companies?' — fetches anonymized community benchmarks. NO PREREQUISITE — call this before you have an analysis to give the user immediate context. No individual company data is exposed (k-anonymity ≥ 5). INPUT: optional industry filter (default 'saas') and optional customer_count_tier (e.g. '100-500', '500-1000'). OUTPUT: median profit margin, average tail drag %, segment distribution, and percentile rankings. TIER: all tiers. For a structured compari

Generate Board Deck

Generate an AI-powered board-ready presentation deck for a saved analysis. PREREQUISITE: an analysisId from analyze(persist=true), analyzeCSV(persist=true), or getAnalyses. INPUT: analysisId (required), optional companyName (used in deck title), optional audience — one of: "board" (governance focus), "ops_team" (tactical actions), "ceo" (strategic overview, default), "investors" (growth/margin story). OUTPUT: structured deck with slides covering profitability summary, segment breakdown, risk are

List Actions

List tracked customer actions and realized ROI outcomes for the authenticated account. INPUT: optional limit (1-200, default 50). OUTPUT: action items with customerId, actionType, status, previous/current revenue and cost, profitDelta, and completion timestamps. WORKFLOW: use after analysis or remediation to answer what actions are pending, completed, and recovering profit.

Record Action

Record a customer profitability action against a saved analysis so Margin Levers can track outcome ROI. INPUT: analysisId, customerId, customerName, actionType ('removed'|'repriced'|'retained'|'cost_reduced'|'upsold'), and optional previous/target revenue, cost, profit, segment, and notes. OUTPUT: the created action. Use this immediately after a user reprices, restructures, retains, upsells, or exits a customer.

Manage Webhooks

List, create, delete, test, or inspect webhook subscriptions for analysis/action/outcome events. INPUT: operation ('list'|'create'|'delete'|'test'|'deliveries'), plus url/events for create, webhookId for delete, and optional limit/offset/subscriptionId for deliveries. OUTPUT: webhook subscription, delivery, or test results. Signing secrets are returned once on create and must be stored immediately.

Profit Concentration Risk

Answer: 'How exposed am I if my top customers churn?' Computes revenue concentration, stress-tests top-N customer loss, and returns a plain-English risk verdict. INPUT: an analysisId from a previously saved analysis, or pass csv text to analyze fresh. topN controls how many top customers to stress-test (default 3). OUTPUT: concentrationScore 0-100, verdict HIGH/MODERATE/LOW, revenueAtRisk, marginsIfTopNChurn, and a narrative sentence. TIER: Growth.

Ranked Call List

Answer: 'Which customers should I call this week — and what do I say?' Returns a ranked, ready-to-use call list with a one-sentence call objective per customer. INPUT: analysisId from a saved analysis, or csv for a fresh run. limit controls list length (5-50, default 10). focus narrows to: 'margin_recovery' (fix unprofitable), 'churn_risk' (high cost-to-serve), 'upsell' (undermonetized profitable accounts), or 'all' (default). OUTPUT: ranked array with callObjective, recommendedAction, and urgen

Peer Benchmark Comparison

Answer: 'How does my business compare to B2B SaaS peers my size?' Fetches anonymized community benchmarks and compares them against your analysis metrics. INPUT: analysisId (optional — uses most recent saved analysis if omitted). businessModel narrows the peer cohort (product_led, smb, enterprise, hybrid, services). OUTPUT: your metrics vs. peer percentiles, gap list, and a priority action recommendation. TIER: All tiers (no AI cost — pure benchmark comparison).

Price Change Scenario

Answer: 'What happens to my margin if I raise prices by X%?' Models a price change across all customers (or a named segment), applies an assumed churn rate, and returns net margin impact plus the break-even churn rate. INPUT: analysisId (required), priceChangePct (-50 to +100), applyTo ('all', 'segment_A', 'segment_B', 'tail'), assumedChurnPct (0-100, default 5). OUTPUT: currentMetrics, projectedMetrics, netMarginLift, breakEvenChurnRate, and a recommendation string. TIER: Growth.