A growth story lives or dies in the behavior of customers. This chapter turns abstract markets into human decision patterns you can model, price against, and win. We focus on how to segment demand in a way that changes decisions, not just slide labels; how to measure demand quality through retention, expansion, and concentration; and how to convert what customers say and do into defensible model inputs. You will see a recurring theme: outside‑in facts, not averages. Start from what different buyers actually need, how they buy, what they pay, and how they stay, then construct a revenue engine that reflects those differences—by segment, not in aggregate.
Done well, this work produces clarity that moves valuation and terms: which segments are large enough and growing fast enough to matter, where willingness to pay is real, where churn lurks, which channels create or destroy margin, and where competitor pressure will appear first. Done poorly, teams average across segments, confuse roles in the buying center, and design pricing or go‑to‑market moves that look smart in a spreadsheet and fail in the field. The methods here are designed for diligence timelines: fast, falsifiable, and auditable, with every claim tied to evidence and translated into the model within 24 hours.
6.1 Segmentation Framework
Segmentation is the art of refusing to average away truths. The objective is not a pretty typology; it is a set of mutually exclusive, collectively exhaustive groups that differ in economics or behavior enough to change what you should pay, how you should structure, and what you must do post‑close. In diligence, “good” segments are those that let you answer five questions crisply: how big is each segment, how fast is it growing, what will it pay, how sticky is it, and what will it cost to win and serve?
Start by choosing segmentation axes that reflect how value is created in the category. In B2B, buyers with the same SIC code can behave very differently if their job‑to‑be‑done, switching costs, or procurement rules diverge. In consumer, demographics are often weak predictors; need state, occasion, and channel matter more. Your framework should therefore combine four lenses:
- Needs and use case: the job the buyer is hiring the product to do, including required outcomes, risk tolerance, and speed of value.
- Economic value: willingness to pay, realized price corridors, potential LTV, and cost‑to‑serve.
- Behavior and access: purchase cadence, channel preference, decision unit composition, and integration or workflow constraints.
- Structure and context: firmographics/demographics, regulatory regime, geography, and interoperability or ecosystem dependencies.
A practical diligence‑grade framework follows a simple path: hypothesize segments that plausibly differ on willingness to pay and stickiness; design primary research to test those differences; assign crisp rules so any new account can be classified in seconds; then size, price, and model them separately.
What makes a segment “investment‑grade”
- Actionable: you can target it with specific offers, price corridors, messages, and channel routes.
- Economic: it exhibits distinct unit economics (CAC, payback, LTV/CAC, contribution margin, cost‑to‑serve).
- Measurable: you can size it and place accounts into it from observable traits or behaviors.
- Stable enough to plan: it won’t collapse if one feature ships late or one competitor changes a SKU name.
- Material: it is big enough, or profitable enough, to change valuation or post‑close priorities.
A step‑by‑step build for a diligence timeline
- Frame the segmentation job to be done. Write, in one line, why you are segmenting: “to quantify willingness‑to‑pay and stickiness differences that move price and retention assumptions,” or “to identify high‑growth micro‑segments where the target’s right‑to‑win is strongest.” If the statement doesn’t affect price, structure, or go/no‑go, simplify.
- Draft hypothesis segments. Create 5–10 named segments that differ on at least two of the four lenses above. Keep labels plain: “Compliance‑driven hospitals ≥300 beds,” “Mid‑market manufacturers with regulated QA,” “Price‑sensitive SMBs buying through resellers,” “Enterprise early adopters with in‑house integration.” Avoid poetic personas.
- Define crisp assignment rules. For each segment, write 2–3 binary criteria using observable fields (industry codes, size bands, channel used, presence of integration X, regulatory exposure Y, decision cycle length Z). If a junior analyst can’t assign an account quickly, the rule is not crisp enough.
- Choose the minimum data needed. Decide what you must collect to prove segments differ in size, growth, price power, and stickiness: realized price ranges, discount ladders, contract length and termination rights, purchase cadence, expansion propensity, channel economics, and switching steps. Map each field to a source (primary interviews, surveys, invoices, distributor checks, panels).
- Test differences with primary research. Design interview and survey instruments to disconfirm: “Do compliance‑driven buyers truly pay more for documented audit trails?” “Do price‑sensitive SMBs churn at higher rates when price increases exceed 3%?” If differences vanish in data, collapse segments.
- Size and price each segment independently. Use the Chapter 5 methods to create segment TAM/SAM and realized price corridors. Apply gross‑to‑net by channel. For recurring models, compute GRR/NRR by cohort and segment.
- Quantify unit economics by segment. Estimate CAC, payback, and LTV/CAC using channel‑specific conversion and discounting. Where access is thin, triangulate with proxy benchmarks and widen ranges with clear confidence labels.
- Select target segments and implications. Rank segments by value at stake and achievability. Document “what would have to be true” to win each: capacity adds, partner listings, product hardening, or regulatory milestones. Translate residual uncertainty into terms (earnouts tied to NRR or price realization).
- Freeze the taxonomy and publish the rules. Lock the segment list and assignment criteria by the mid‑sprint gate. Put them in the data book so every exhibit, model input, and quote is segment‑coded the same way.
The Segmentation Canvas (copy and fill one per segment)
- Name and rule: label plus 2–3 binary assignment rules.
- Size and growth: segment SAM today; growth range with drivers.
- Buyers and DMU: economic buyer, technical approver, user; decision cycle length.
- Value drivers and risks: outcomes that matter; failure modes; compliance or ecosystem constraints.
- Price corridor: realized price range; discount drivers; elasticity notes.
- Stickiness: contract terms; switching steps; NRR/GRR with drivers of expansion and churn.
- Channel route and cost‑to‑serve: direct vs. partner; rebate/margin norms; service intensity.
- Competitors and likely responses: who shows up, where they win, and how they fight.
- “Would‑have‑to‑be‑true”: conditions to win value during the hold period.
- Confidence rating: high/medium/low and what would upgrade it.
Common segmentation axes that actually change decisions
- Job‑to‑be‑done: compliance audit vs. productivity boost vs. revenue generation.
- Switching cost profile: data lock‑in, training burden, regulatory recertification.
- Procurement regime: centralized with MFNs vs. local buyer discretion.
- Channel path: marketplace, distributor, VAR, or direct; each implies different gross‑to‑net.
- Usage intensity: light vs. heavy users; ties to usage‑based pricing and support loads.
- Risk posture: mission‑critical vs. nice‑to‑have; determines price corridors and churn risk.
- Ecosystem dependency: needs integration with platform X or certification Y.
- Geography and regulation: reimbursement codes, data residency, or local content rules.
B2B, consumer, and platform nuances
- B2B: Segment the decision‑making unit. A product can be “user‑loved” and “procurement‑blocked.” Include procurement as its own cohort when pricing power is central to the thesis.
- Consumer: Segment by need state and channel, not just age or income. Occasion, frequency, and promo sensitivity are stronger predictors of realized price and repeat rates.
- Marketplaces/platforms: Segment both sides. On the supply side, sort sellers by GMV, return/fraud rates, and take‑rate sensitivity. On the demand side, segment buyers by frequency, AOV, and category basket overlap.
Rapid 72‑hour segmentation build
- Day 0: Draft hypothesis segments, assignment rules, and a one‑page purpose statement.
- Day 1: Tag early interviews and secondary sources by segment; pull two top‑down series that can be split by your chosen axes; start bottom‑up counts for two representative segments.
- Day 2: Land first price corridors and NRR/GRR differences for at least two segments; collapse or refine hypotheses based on evidence; publish the initial Segmentation Canvas set.
- Day 3: Re‑size segments with refined rules; push segment‑coded inputs into the model; run an initial tornado to see which segment variables swing value most.
Quality checks that separate usable segments from slide art
- Measurability: can a stranger classify an account with the rule set in under a minute.
- Distinct economics: visible differences in realized price, CAC/payback, or GRR/NRR.
- Sizing tie‑out: segment totals sum to category totals; no orphan demand.
- Channel realism: gross‑to‑net and cost‑to‑serve differ by segment and are modeled explicitly.
- Stability: segments don’t flip labels with small changes in features or roadmap timing.
- Auditability: every segmented claim is footnoted and replicated by a verifier.
Red flags—and what to do about them
- Vague segments (“innovators,” “traditionalists”) with no assignment rule. Replace with observable criteria.
- Over‑segmentation that dilutes decision value. Collapse to the smallest set that explains price and stickiness differences.
- Segment labels that mirror org charts, not buyers. Rebuild around needs and behavior.
- Invisible channel effects. If realized price or CAC doesn’t change by segment, your segmentation is missing the channel lens.
- NRR averages hiding decay in one segment. Recompute cohorts by segment and re‑cut growth assumptions.
- Geography ignored in regulated markets. Create regulatory segments; move unapproved indications to option TAM.
How segmentation feeds the model and the term sheet
- Price: apply segment‑specific realized price and increase corridors; tie earnouts to price realization where uncertainty remains.
- Retention: set GRR/NRR by segment with explicit drivers; translate low‑confidence segments into wider ranges and covenant triggers.
- Share and SOM: concentrate share gains in segments where the right‑to‑win is evidenced; keep others in upside only.
- Channel: adjust gross‑to‑net and cost‑to‑serve by segment; reflect partner ramp times and rebates in timing.
- Post‑close priorities: staff coverage, partner listings, and product hardening aligned to target segments; measure leading indicators (win rate vs. named competitors, premium attach, discount depth) by segment monthly.
When you segment this way—rules first, economics second, evidence always—you stop arguing narratives and start comparing choices. The output is not a persona wall; it is a small set of segments you can price, win, and serve differently, with clear value at stake and a plan to capture it.
6.2 Voice-of-Customer Research – Step-by-Step Guide
Voice‑of‑Customer (VoC) turns market abstraction into evidence you can price, forecast, and bank on. In diligence, the goal is not to collect anecdotes; it’s to pressure‑test the growth story by listening to the buyers and users who create—or destroy—revenue. The method here is fast and falsifiable: anchor on the few “would‑have‑to‑be‑true” statements that move value, design instruments that try to break them, and translate every finding into model inputs with explicit confidence.
Begin by tying VoC to segments, not to the “average customer.” You defined segments in 6.1 because economics differ by need, channel, and switching cost. Your VoC plan should mirror those differences: decision‑makers vs. end users, loyalists vs. defectors, direct vs. channel buyers, heavy vs. light users, and procurement vs. budget owners. Aim to cover both sides of any contested story (e.g., “renewals are strong” and “pricing power is real”) with cohorts designed to refute as well as confirm.
Step 1: Translate hypotheses into interviewable questions
Rewrite each hypothesis as a question about recent, specific behavior. “Realized price can rise 2–4% annually without elevating churn” becomes “What happened the last time price increased—who approved it, what pushback occurred, and did anyone reduce scope or leave?” VoC lives on concrete episodes—last evaluation, last renewal, last loss—not opinions about the future.
Step 2: Select cohorts that mirror revenue and risk
Sample where dollars and exposure sit. For a two‑to‑three‑week sprint, a high‑yield mix looks like: current customers in top‑value segments; churned customers from the last 12–18 months; active prospects and recent lost deals; procurement and finance approvers; and channel partners/distributors that shape gross‑to‑net. When value hinges on a few strategic accounts, treat them as their own cohort and include executive‑level voices and front‑line users separately.
Step 3: Define quotas and booking rules that protect validity
Allocate quotas by segment, role, and region so your cuts match the model. Over‑recruit churned/lost cohorts relative to their population—they are informationally dense. Balance “fans” and “frustrated.” Time‑box recruiting with daily book‑rate targets, and pre‑authorize incentives and substitutes (e.g., channel checks when end‑customers are inaccessible) so fieldwork never stalls.
Step 4: Craft instruments that test, not lead
Use semi‑structured guides focused on recent decisions. Ask respondents to replay the last evaluation: trigger, shortlist, decision criteria, and veto points. Ladder from features to outcomes: “What did that feature let you do that you couldn’t before?” Probe price with acceptance, not affection: “At last renewal, which concessions were requested and granted?” For stickiness, map the switching path step‑by‑step—data migration, retraining, re‑certification, integration rework—and quantify effort in hours, dollars, and risk.
Step 5: Add targeted modules for pricing, retention, and competition
Insert a quick willingness‑to‑pay task (Van Westendorp or Gabor‑Granger) when you need corridors; keep it short and segment‑specific. For retention, run a “cohort autopsy” with churned customers: reason hierarchy, tipping point, and counterfactual (“What would have kept you?”). For competition, run win/loss prompts that force rank the three claims that moved the decision and the price differential at which the outcome would have flipped.
Step 6: Secure compliance and reduce bias up front
Operate under clean‑team rules when required: restrict raw PII and sensitive notes to named individuals; share only anonymized, aggregated outputs beyond the clean team. Disclose purpose unless counsel approves a blinded approach. Avoid coaching or leading questions; ban future pricing coordination topics. Record with consent and state retention/deletion timing for notes and audio.
Step 7: Field fast and monitor mix continuously
Launch within 24–48 hours. Track completes by cohort vs. quota, book rate, and segment balance daily. If a pivotal segment is under‑represented by Day 2, shift outreach and incentives the same day. Tag each interview at booking with segment, role, ACV band, and region so analysis cuts are ready as calls land.
Step 8: Code rigorously—turn stories into structured evidence
Create a simple codebook tied to your hypotheses: price pushback type, discount ladder step, reason for churn, competitor claim that resonated, switching steps required, and perceived differentiation. Code each transcript within 24 hours. Quantify frequency by segment and role, and log notable quotes with anonymized IDs that include segment, role, region, and revenue band so they can be traced during Q&A.
Step 9: Triangulate stated answers with observed behavior
Elevate evidence that is closer to behavior: invoices and renewal letters over recollections; win/loss outcomes over brand affinity; discount ladders over list prices; deployment telemetry over claimed usage. Where you only have stated intent, carry wider ranges and lower confidence. Pair VoC with secondary signals (scanner data, marketplace prices, channel margins) to bound realism.
Step 10: Convert findings into model‑ready inputs within 24 hours
Every material VoC result should land in the model or risk ledger quickly: realized price corridors by segment; approval hurdles for price increases; contract terms that change churn risk; cohort‑level expansion propensities; conversion rates and ramp times in the funnel; channel gross‑to‑net differences. When a claim remains uncertain, widen sensitivity ranges and state the “would‑have‑to‑be‑true” conditions explicitly.
Step 11: Surface early‑warning indicators for post‑close tracking
Translate big VoC themes into leading signals: discount depth, rebate accruals, procurement escalations, competitor trial frequency, premium attach rates, price‑related churn mentions, and renewal‑stage slippage. Tie each to a monitoring cadence so management can act before lagging KPIs move.
Step 12: Present VoC credibly—quotes with provenance and math with ranges
Use a small number of high‑signal quotes, each footnoted with an anonymous segment/role and month of interview. Summarize themes with counts (“11 of 16 procurement leaders require CFO sign‑off above $X ACV”) and show ranges where appropriate (“realized price in Segment B: $14–$19 per seat per month”). Keep opinions out; show what customers did and what they will likely do next.
Advanced moves that raise the signal‑to‑noise ratio
Work from “critical incidents”—the last renewal, a painful outage, a pivotal pilot—rather than generic satisfaction. Ask “what almost broke the deal?” to surface silent churn risks. Use contrast questions to reveal elasticity: “If a rival offered 12% below your current rate, what would you change first—scope, service level, or vendor?” When buyers describe integration or compliance hurdles, map and cost each step to build a Switching Cost Index by segment. For usage‑priced models, separate “active rate” (share of customers doing anything) from “intensity” (units per active) and “collectability” (what actually gets billed and paid).
Common failure modes—and the remedy
Happy‑talk bias appears when you speak only to champions; fix by oversampling churned/lost cohorts and procurement. Leading questions inflate willingness to pay; fix by anchoring in the last negotiation and by using acceptance questions at discrete prices. Thin evidence becomes false precision when ranges are collapsed into point estimates; fix by labeling confidence and showing corridors. Vendor‑sourced lists skew results; fix with independent recruiting and channel checks. And beware NPS masquerading as loyalty—tie “advocacy” back to renewal behavior and price acceptance before you lean on it.
VoC sprint checklist (copy, fill, enforce)
- Hypotheses rewritten as behavior‑based questions tied to value at stake.
- Cohorts and quotas aligned to segments, roles, and regions; churned/lost oversampled.
- Semi‑structured guides written to disconfirm; pricing and retention modules inserted only where needed.
- Clean‑team and compliance memo approved; consent and data handling defined.
- Daily dashboard tracking completes, mix, and book rate; plan B substitutes pre‑authorized.
- Codebook defined; transcripts coded within 24 hours; quotes logged with anonymized IDs.
- Triangulation rule applied: invoices/contracts and win/loss data prioritized over recollection.
- Findings translated into model lines (price, GRR/NRR, conversion, ramp times, gross‑to‑net) with ranges and confidence.
- Early‑warning indicators documented for post‑close monitoring.
- Exhibits verified, footnoted, and replicable; “Draft—Not Verified” labeled when QA is pending.
Run this play and your VoC will do the two things diligence needs most: break weak theses quickly and strengthen believable ones with evidence that ties cleanly to price, retention, and achievable growth—by segment, not in aggregate.
6.3 Cohort and Retention Analysis Template
Cohort and retention analysis is how you turn “sticky revenue” from a claim into math. In commercial due diligence, this work answers three valuation‑critical questions: how durable is the revenue base, where does expansion really come from, and what would have to be true for net retention to hold (or improve) through the hold period. The template below is built for speed and scrutiny. It gives you standard definitions, a clean workbook structure, and step‑by‑step instructions to produce numbers an Investment Committee can trust—replicable, reconciled to the ledger, and cut by the segments that move value.
Start with definitions you will not change mid‑sprint. Dollar‑based Gross Revenue Retention (GRR) measures how much starting recurring revenue you kept, excluding any expansion. Dollar‑based Net Revenue Retention (NRR) measures how much you kept after adding expansion (and reactivation, if you include it). Logo retention is the count‑based analog; it detects “silent churn” masked by expansion. Write the exact formulas into your workbook and align them with Quality of Earnings (QofE) on day one so no one argues definitions at T‑3.
- GRR over period t:
GRR = (Start ARR − Churn ARR − Contraction ARR) ÷ Start ARR. - NRR over period t:
NRR = (Start ARR − Churn ARR − Contraction ARR + Expansion ARR + Reactivation ARR) ÷ Start ARR.
Treat reactivation consistently: include it in NRR only if it meets your “new logo vs. resurrected logo” rule.
Data you need before you begin (minimum viable set)
Customer‑level recurring revenue ledger with monthly ARR/MRR by product/module, contract start/renewal/term dates, list vs. realized price or rate card, discounts/rebates/credits, channel (direct vs. partner), region, segment assignment (from 6.1), logo IDs (parent/child hierarchy), status flags (new, active, churned, reactivated), and reason codes for churn/contraction if available. Capture FX currency, invoice timing, one‑off adjustments, and pass‑through lines you will exclude from ARR. If usage‑based, pull units and rate plan per period; if hardware + service, split device vs. annuity.
Cohort design—make three choices and freeze them
- Cohort key. Use acquisition or go‑live month as default; for renewals analysis, also cut by initial term start. Document why.
- Time grain and horizon. Monthly is standard for sprints; 24–36 months gives you decay shape. Quarterly is acceptable in low‑volume B2B.
- Segmentation. Apply the same segment rules you locked in 6.1 (vertical, size band, channel, region, product family, ACV band). If a segment can’t be assigned quickly with observable fields, refine the rule before you compute anything.
Build the cohort matrix (the triangle) once, then reuse it everywhere
Create one sheet per segmentation view (e.g., “All,” “By Segment,” “By Channel”). Rows are cohorts by start month; columns are months‑since‑start (0, 1, 2 … n). For each cell, store:
- Start Logos, Start ARR (at the beginning of the period)
- Expansion ARR, Contraction ARR, Churn ARR, Reactivation ARR (during the period)
- End Logos, End ARR (at period end)
- GRR and NRR for that cell and the cumulative GRR/NRR for the cohort at age n
Compute roll‑ups at the top: weighted GRR, weighted NRR, logo retention, and the distribution of cohorts by age (to avoid averaging young and old cohorts indiscriminately).
Decompose retention into what actually changed
Retention is not a monolith; split it so you can manage it.
- Price vs. quantity (seats/units). For any change in ARR, attribute the share due to realized price change (rate increase, discount rollback, rebate change) vs. quantity change (seats, modules, usage). This prevents “price‑led expansion” from being mistaken for product‑led growth.
- Product mix. Tag expansion and contraction to modules/SKUs so you can see whether the annuity (consumables/services) is carrying the day while the core product decays—or vice versa.
- Channel effect. Separate direct vs. partner. Gross‑to‑net and renewal mechanics differ; partner attrition can hide inside NRR if you don’t isolate it.
- Contract mechanics. Map term length, auto‑renew clauses, termination/assignment rights, and notice windows to the month where churn can actually happen; long terms create retention optics that aren’t the same as customer love.
Diagnostics you should run every time
- Survival and hazard view. Plot cohort survival (ARR and logos) and monthly hazard (churn probability) to find cliff months (e.g., Month 12 for annual, Month 3 post‑implementation).
- Retention heatmap. Visualize GRR/NRR by cohort age and segment; hotspots usually align with a product gap, a price increase, or a channel change.
- Renewal waterfall (next 4–8 quarters). Build a schedule of ARR up for renewal, expected GRR, expected expansion, and “ARR at risk” by segment; this feeds both valuation and term‑sheet protections.
- Concentration overlay. Compute GRR/NRR with and without the top 10 accounts; concentrated expansion can mask broad‑based weakness.
- Elasticity signals. Cross‑tab price increase events with churn/contraction in the following two cycles; look for rising hazards in price‑sensitive segments.
- Cohort mix shift. If young cohorts retain worse than older ones, growth may be seeding future churn; treat as a red flag and widen downside ranges.
Template—fields to copy into your workbook
- Sheet A: Cohort Matrix
Cohort Month; Segment; Channel; Start Logos; Start ARR; Expansion ARR (price); Expansion ARR (quantity); Contraction ARR (price); Contraction ARR (quantity); Churn ARR; Reactivation ARR; End Logos; End ARR; Period GRR; Period NRR; Cumulative GRR (Age n); Cumulative NRR (Age n). - Sheet B: Drivers & Reasons
Cohort; Reason Code; Driver Type (Product gap, Service, Price, Competition, Procurement mandate, Budget); Notes; Dollar impact; Segment; Channel. - Sheet C: Renewal Waterfall
Quarter; ARR up for renewal; Expected GRR; Expected Expansion; ARR at Risk; Top Accounts; Contracts with adverse terms (MFN, termination, assignment). - Sheet D: Price/Quantity Bridge
Period; Segment; Seats/Units change; Realized rate change; Dollar effect from rate; Dollar effect from seats; Net effect; Elasticity notes. - Sheet E: Concentration & Outliers
Account; ARR; % of ARR; Retention history; Notes; Treatment (winsorize, exclude, disclose).
Usage‑based, marketplace, and hardware+service nuances
- Usage‑based. Track active‑rate (% of logos generating any usage), intensity (units per active), realized rate, and collectability (what is actually billed and paid). NRR can look strong on a usage surge that won’t persist; tie scenarios to observable drivers (seasonality, policy, incentives).
- Marketplaces/platforms. Retention is GMV‑based for sellers and purchase‑frequency‑based for buyers; revenue retention adds take‑rate stability. Watch return/fraud offsets that erode net take.
- Hardware + service. Separate device replacement (cyclical) from the annuity (consumables/services). Many narratives rely on the annuity; prove it with cohorted service revenue per installed base.
Quality controls that keep the math honest
- Align ARR/MRR definitions with QofE; exclude one‑offs, pass‑throughs, and services that are not recurring.
- Reconcile cohort totals to the general ledger; your triangle must roll up to reported ARR.
- De‑duplicate parent/child logos; decide once how you treat multi‑site accounts and acquisitions.
- Lock FX policy and present retention in both reported and constant currency if FX volatility is material.
- Label the treatment of pauses, credits, and partial churn; don’t hide them in contraction unless that is your standard.
- Verify that price and quantity bridges add to the observed delta; if they don’t, you have a mapping error.
- Keep a verifier separate from the author; nothing leaves the workbench without replication and footnotes.
Red flags (and immediate responses)
- High logo churn with “healthy” NRR. Expansion from a few large accounts is masking decay; split reporting with/without top accounts and cut price/mix by segment.
- Step‑function price increases precede churn. Down‑weight price‑led expansion in the forecast, widen ranges, and propose an earnout tied to price realization.
- Young cohorts underperform older ones. Move growth into upside only until you see stabilization; prioritize product hardening or onboarding fixes in the post‑close plan.
- Channel‑driven leakage. If partner‑sold cohorts renew worse or net less after rebates, adjust gross‑to‑net and cost‑to‑serve; push for channel terms or coverage changes.
- Contract optics. Three‑year terms with weak end‑user adoption will unwind later; treat GRR as artificially inflated and scenario‑ize the cliff.
How to push results into the model within 24 hours
- Replace single‑line “retention” with segment‑level GRR/NRR ranges and confidence labels.
- For subscriptions, forecast using a cohort roll‑forward (Start ARR → churn/contraction → expansion) rather than a flat NRR scalar.
- For usage models, forecast active‑rate, intensity, and realized rate separately; add collectability and an “incentive drag” where promos are material.
- Pull the renewal waterfall into the base/downside scenarios and tie ARR‑at‑risk to specific drivers (price, product gap, competitor).
- Translate residual uncertainty into terms: earnouts linked to NRR or price realization, covenants on promo intensity, closing conditions on key contract assignments.
Acceptance criteria for a decision‑grade retention analysis
- Definitions frozen and aligned with QofE; treatment of reactivation, pauses, and pass‑throughs documented.
- Cohort triangle built, verified, and reconciled to ledger totals; segments and channels applied consistently.
- GRR/NRR presented as ranges by segment with drivers; price vs. quantity bridge validated.
- Survival/hazard views and renewal waterfall produced; concentration impact shown with/without top accounts.
- Findings translated to model lines and term‑sheet levers; confidence and early‑warning indicators stated on the page.
Quick checklist you can copy into your workspace
- Cohort key, time grain, horizon frozen.
- Segment and channel tags applied consistently.
- Triangle computed with Start/End, Expansions, Contractions, Churn, Reactivations.
- GRR/NRR and logo retention calculated; survival and hazard charts reviewed.
- Price vs. quantity bridge completed for top segments.
- Renewal waterfall and ARR‑at‑risk built for next 4–8 quarters.
- Reconciliation to ledger done; verifier sign‑off logged.
- Model updated; ranges and confidence labeled; residual risk tied to price/structure levers.
Use this template and you’ll replace generic “strong retention” claims with a precise view of revenue durability—by segment, by channel, and by driver—so sponsors can price risk, shape terms, and prioritize post‑close actions with confidence.
6.4 Demand Elasticity Stress-Test Checklist
Elasticity work tells you how much revenue and unit volume will move when price, promotion, or product mix changes. In diligence, it protects you from two costly mistakes: paying for price upside that customers won’t accept, and overlooking price power that is hiding behind discounting habits or poorly framed increases. This stress‑test converts opinions about “pricing headroom” into evidence by segment, SKU, and channel—and pushes those findings directly into the model, retention forecast, and term sheet.
Approach this as a short, disciplined experiment rather than an academic exercise. You will reconstruct realized prices, isolate promotion and mix effects, and estimate how demand reacts in the short run (1–3 months after a change) and long run (renewal cycles or replenishment windows). Then you will pressure‑test a handful of decisions‑relevant scenarios: permanent price increases, removal of promotions, competitor price aggression, macro slowdowns, and policy or reimbursement resets. Throughout, keep segmentation front and center; averages hide the truth.
Step‑by‑step elasticity stress‑test
- Define the question and the materiality bar.
Write the single statement you are trying to prove or break (e.g., “We can raise realized price 200–400 bps in Segment A without pushing churn above 8% or downgrades above 5%”). Quantify value at stake and set stop/go thresholds before analysis. - Select segments, SKUs, and channels that actually move value.
Prioritize “key value items” and top revenue segments. Tag direct vs. partner routes separately; gross‑to‑net and pass‑through differ by channel and can flip the answer. - Reconstruct realized price and volume at a transactional level.
Start with the price waterfall: list → standard discounts → promo → rebates/chargebacks → returns/credits → channel margins → vendor‑realized price. Align periods, deduplicate, and separate base from promo volume. - Build a clean baseline.
Create a history of price, volume/units, and mix by segment and channel, controlling for seasonality, inventory constraints, outages, and one‑off deals. If supply constraints or stockouts occurred, flag those months rather than letting them distort elasticity. - Estimate own‑price elasticity two ways.
Method A: event study around natural experiments (price changes, promo withdrawals, pack size shifts). Method B: multivariate controls using simple regression or matched‑pair comparisons that strip out seasonality, promo, and mix. Report short‑run and long‑run effects. If the two methods disagree materially, widen ranges and lower confidence. - Separate price from promotion, mix, and quantity effects.
Quantify how much of past “price” movement came from mix up‑tiering, temporary discounts, or pack architecture rather than true net price. Do not claim price power when the math shows mix or promo doing the work. - Tie elasticity to retention and win rates.
For subscriptions, test whether cohorts exposed to price increases show higher churn, downgrades, or delayed renewals. For transactional models, examine repeat‑purchase frequency and basket size after price moves. Add the hazard to your retention model if you see a reliable pattern. - Add competitive and procurement lenses.
Use VoC findings to identify procurement veto points and competitor undercut patterns. Convert these into response curves (e.g., “a 10% price gap vs. Competitor X cuts win rate by 12–18 points in Mid‑Market”). - Run decision‑relevant scenarios.
Model permanent price increases, promo reduction, KVI price holds with tail monetization, competitor price war, macro slowdown, reimbursement change, and FX or input‑cost shocks. Present ranges and confidence; tie each to early‑warning indicators. - Translate to the model and term sheet within 24 hours.
Update realized price by segment and channel, adjust demand and retention where elasticities bite, and push the new contribution margins through unit economics. Where uncertainty remains, encode it in sensitivities and deal structure (e.g., earnouts on price realization, covenants on promo intensity).
Data pack you should assemble before you start
- Transaction‑level sales with list price, net price, discounts, rebates, credits, returns, pack size, and channel tags.
- Units or usage counts by SKU and segment; attach rates for bundles and options.
- Promotion calendar with funding source and duration.
- Inventory, availability, and service‑level data to detect stockouts or capacity caps.
- Competitor prices from panels, marketplaces, or distributor quotes.
- Contract terms for renewal cadence, MFNs, and termination rights.
- Macro or policy indicators that plausibly move demand in your category.
Analytical checks that keep you honest
- Use realized price, not list.
- Control for promo and mix explicitly; do not attribute their effects to price.
- Estimate short‑run and long‑run elasticities; label the window used.
- Validate with two methods or sources for any thesis‑critical estimate.
- Winsorize outliers and report with/without top customers to expose concentration.
- Ensure add‑up: price × volume bridges to revenue after gross‑to‑net.
- Present ranges with confidence labels; avoid point “precision” where evidence is thin.
Scenario menu you can lift directly into your model
- Permanent price increase of 2–4% in top segments; test with and without KVI exemptions.
- Removal of promotional funding by 30–50%; assume partial volume loss and partial mix shift.
- Competitor discounting of 10–15% in one pivotal segment; overlay win‑rate impact.
- Recession shock with category contraction of 5–10% and increased price sensitivity.
- Reimbursement cut or policy delay; model an immediate step‑down and slower recovery.
- FX or input‑cost spike; test pass‑through speed and demand loss from necessary increases.
B2B, consumer, and platform nuances
- B2B: Elasticity is mediated by procurement policy and switching costs. Expect low short‑run elasticity and step‑changes at renewal. Model approval thresholds and MFNs explicitly.
- Consumer: KVI items anchor value perception; hold or limit increases there while monetizing the tail with pack architecture and premium tiers.
- Usage‑based pricing: Break demand into active rate, intensity, and realized rate; elasticity often shows up as downgrades in intensity rather than logo churn.
- Marketplaces: Test GMV sensitivity on both buyer and seller sides and net it through take‑rate compression and returns/fraud.
Early‑warning indicators to monitor during and after the test
- Discount depth and promo reliance by segment.
- Price‑related churn or downgrade mentions in support and renewal notes.
- Win rate versus named competitors at comparable prices.
- Premium attach rates and up‑tier conversion.
- Channel pushback, rebate accruals, and delistings.
- Renewal slippage and extended approval cycles in procurement.
Compliance and guardrails
- Do not discuss future pricing or market allocation with competitors; pre‑close coordination is prohibited.
- Operate under clean‑team rules for invoice‑level analyses and raw customer data; share only aggregated, anonymized outputs with non‑clean members.
- For primary research, disclose purpose (unless counsel approves a blinded approach), obtain consent to record, and avoid eliciting material nonpublic information.
Red flags—and immediate responses
- Strong NRR but rising price‑related churn in a specific segment. Lower price uplift assumptions there; tighten ranges; consider an earnout tied to NRR.
- Elasticity estimates flip sign after controlling for promo or mix. Rebuild realized price and rerun; do not rely on list‑price analytics.
- Channel leakage overwhelms vendor price moves. Adjust gross‑to‑net and cost‑to‑serve; revisit channel strategy in post‑close planning.
- Capacity or service constraints drive volume loss after price increases. Treat as an operational bottleneck, not elasticity; coordinate with ops diligence.
Acceptance criteria for a decision‑grade stress‑test
- Two independent estimation methods or sources for the elasticity of each thesis‑critical segment.
- Realized price reconstructed with a visible waterfall; promo and mix accounted for.
- Short‑run and long‑run effects estimated and labeled.
- Scenarios built on the few variables that swing value and tied to early‑warning indicators.
- Model updated with ranges and confidence; term‑sheet levers proposed where uncertainty remains.
Run this checklist and you will replace generic “pricing headroom” claims with a quantified, segment‑specific view of what customers will accept, what they will resist, and how that translates—directly—into revenue, retention, and enterprise value.
