How do we calculate freemium-to-paid conversion CAC payback when self-serve acquisition cost is near-zero?

🎯 Bottom Line

• [Answer] Calculate Fully-Loaded CAC payback for freemium PLG by using Bessemer's definition — divide (paid acquisition spend + free-tier infrastructure cost amortized over the paying cohort + sales/CS touch on PQLs + free-user support burden) by (number of new paying customers in cohort × monthly gross profit per paying customer). The "narrow CAC" view — paid acquisition spend / new paying customers — systematically understates true cost by 2-3x for freemium businesses because it externalizes the 8-25% of revenue burned on free-tier infrastructure as "marketing" or "R&D" rather than CAC. Healthy 2024-2026 freemium SaaS targets: Fully-Loaded CAC payback of 18-32 months (median per OpenView 2024 SaaS Benchmarks), narrow CAC payback of 6-14 months (the headline number most companies report). Best-in-class PLG: Slack ~13 months at IPO (2019), Figma sub-18 months at scale, Calendly ~14 months, Notion ~24 months, Loom ~28 months, HubSpot freemium CRM cohorts ~22 months.
• [Why] Self-serve PLG inverts the traditional sales-led CAC equation. In sales-led (Salesforce, Workday, Snowflake) CAC is clearly attributable — outbound SDR + AE + marketing spend divided by closed deals = CAC, with payback typically 18-24 months. In freemium PLG, 95-98% of acquired users never pay, and the cost of supporting those free users (infrastructure, support, engineering, brand) is real but mis-categorized. Bessemer's "Fully-Loaded CAC" framework (codified in their annual State of the Cloud reports 2019-2024) and OpenView's PLG Index (2020-2024 annual benchmarking, n=600+ SaaS companies) both insist on amortizing free-tier cost over the paying cohort to get an honest payback figure. Patrick Campbell (ProfitWell, acquired by Paddle 2022) has published research showing that 70%+ of PLG founders report payback periods that are 40-60% optimistic relative to a fully-loaded reconstruction — a systematic measurement gap that misleads board decisions, fundraising narratives, and growth-vs-efficiency tradeoffs. The fix: report both numbers (narrow + fully-loaded) and target fully-loaded specifically when making capital allocation decisions.
• [Caveat] CAC payback is the wrong KPI in three scenarios: (1) early-stage PLG (pre-Series A) where cohort sizes are too small for statistical meaning — use LTV/CAC ratio (target 3x+) and magic number (target 0.7-1.0+) instead; (2) usage-based pricing (Stripe, Twilio, Snowflake, Datadog, MongoDB Atlas) where ACV ramps over 12-24 months — payback inherently lags revenue recognition, requiring cohort-LTV projection rather than current-MRR division; (3) hybrid PLG-to-enterprise motions (Slack, Atlassian, Notion, Figma post-Adobe-attempt) where free-tier users seed enterprise deals 18-36 months downstream — narrow CAC payback misses the indirect attribution chain entirely. In these cases use burn multiple (Net Burn / Net New ARR; target <2.0x), NRR (target 110%+ for PLG, 120%+ best-in-class), and rule of 40 (growth % + EBITDA margin % ≥ 40) as the dominant KPI stack and treat CAC payback as a secondary diagnostic.

Freemium-to-paid CAC payback is the months-of-gross-profit a cohort of newly-paying customers requires to repay the all-in customer-acquisition cost their cohort imposed on the business, calculated rigorously by Bessemer Venture Partners' Fully-Loaded CAC methodology (State of the Cloud 2019-2024), OpenView Partners' Product-Led Growth Index (annual 2020-2024, n=600+ companies), Wes Bush's Product-Led Growth model (Product-Led: How to Build a Product That Sells Itself, 2019), Patrick Campbell's ProfitWell/Paddle pricing research (2014-2024, n=10,000+ SaaS companies), Brian Balfour's Reforge Four Fits Framework (Model-Market Fit specifically tests payback economics), Kyle Poyar's Growth Unhinged + OpenView pricing teardowns (2021-present), David Sacks' SaaS Metrics Framework (Craft Ventures, 2008-present, popularized "Magic Number" and "Burn Multiple"), and Jason Lemkin's SaaStr 2024 PLG benchmarks.

The metric is calibrated against 2024-2026 freemium SaaS medians of 18-32 months Fully-Loaded payback (healthy band), 36+ months indicating dangerous unit economics, and best-in-class cases like Slack (~13 months at IPO 2019), Figma (sub-18 months at scale), Calendly (~14 months), Notion (~24 months), Loom (~28 months), HubSpot's freemium CRM cohort (~22 months), Datadog (~16 months free-to-paid component), and Atlassian's historical no-sales-team PLG cohorts (~20 months) — all measured against the comparable sales-led benchmark of 18-24 months (Salesforce, Workday, Snowflake, ServiceNow per their 10-K disclosures).

The function of correctly-measured Fully-Loaded CAC payback is (a) honest capital allocation — telling the board / CFO / investors what the business actually costs to grow rather than the optically-flattering narrow CAC; (b) cohort comparison — measuring whether month-1 cohort, month-12 cohort, and month-24 cohort are improving or degrading in unit economics (a critical leading indicator of channel saturation, ICP drift, or pricing-misfit); (c) growth-vs-efficiency tradeoff — knowing whether to step on the gas (payback <12 months means accelerate) or pump the brakes (payback >24 months means optimize before scaling); (d) channel attribution — comparing payback by acquisition channel (organic/SEO/viral vs paid search vs partnerships vs outbound) to identify which channels actually justify their cost; and (e) cross-company benchmarking — comparing your payback to OpenView's PLG Index medians to know if you're best-in-class, healthy, or dangerously inefficient.

Freemium economics are structurally different from sales-led economics in three ways. First, freemium has a conversion rate layer that sales-led doesn't — 100 sales-led MQLs may produce 5 closed deals (5% conversion), but 100 freemium signups may produce 3 paying customers (3% conversion) and 97 non-paying free users whose support + infrastructure cost must be allocated somewhere.

Second, freemium has a time-to-conversion distribution that sales-led doesn't — 50% of paying customers convert within 30 days of signup, 75% within 90 days, 90% within 12 months (PostHog 2023 data), with a long tail extending 24+ months. This means cohort-CAC payback math has to handle the timing asymmetry between when CAC is spent (mostly upfront acquiring the free user) and when revenue arrives (spread across 12+ months as users convert).

Third, freemium has a viral coefficient layer that sales-led doesn't — every Loom recipient, every Figma collaborator, every Slack guest is a downstream signup vector, meaning the "true CAC" should include attribution for organic/viral signups that were caused by paid signups, which creates a compounding-return profile that simple CAC math doesn't capture.

🗺️ Table of Contents

Part 1 — THE QUESTION

• [Why freemium CAC payback is the most mis-calculated metric in PLG SaaS](#why-freemium-cac-payback-is-the-most-mis-calculated-metric-in-plg-saas)
• [The structural differences between freemium and sales-led CAC math](#the-structural-differences-between-freemium-and-sales-led-cac-math)
• [What counts as CAC when 95%+ of users never pay](#what-counts-as-cac-when-95-of-users-never-pay)

Part 2 — THE FRAMEWORK

• [Bessemer's Fully-Loaded CAC definition and why it matters](#bessemers-fully-loaded-cac-definition-and-why-it-matters)
• [Wes Bush's PLG CAC formula and common errors](#wes-bushs-plg-cac-formula-and-common-errors)
• [Patrick Campbell on cohort CAC math and the time-to-convert distribution](#patrick-campbell-on-cohort-cac-math-and-the-time-to-convert-distribution)
• [Reforge / Kyle Poyar on channel-cohort attribution and the viral coefficient adjustment](#reforge--kyle-poyar-on-channel-cohort-attribution-and-the-viral-coefficient-adjustment)

Part 3 — THE EVIDENCE

• [Slack, Figma, Notion: best-in-class PLG payback benchmarks](#slack-figma-notion-best-in-class-plg-payback-benchmarks)
• [Calendly, Loom, HubSpot freemium CRM: middle-of-the-pack examples](#calendly-loom-hubspot-freemium-crm-middle-of-the-pack-examples)
• [The worked math walkthrough: narrow vs fully-loaded for a sample PLG SaaS](#the-worked-math-walkthrough-narrow-vs-fully-loaded-for-a-sample-plg-saas)
• [2024-2026 OpenView PLG Index, ChartMogul, Baremetrics, ProfitWell benchmark data](#2024-2026-openview-plg-index-chartmogul-baremetrics-profitwell-benchmark-data)

Part 4 — THE RECOMMENDATION

• [Dual-reporting protocol: narrow + fully-loaded](#dual-reporting-protocol-narrow--fully-loaded)
• [When CAC payback is the wrong KPI and what to use instead](#when-cac-payback-is-the-wrong-kpi-and-what-to-use-instead)
• [Tooling stack for 2026 cohort CAC modeling](#tooling-stack-for-2026-cohort-cac-modeling)
• [Counter-case: freemium is the wrong motion entirely](#counter-case-freemium-is-the-wrong-motion-entirely)

📐 PART 1 — THE QUESTION

Why freemium CAC payback is the most mis-calculated metric in PLG SaaS

Walk into any Series A PLG board meeting in 2026 and ask the founder for their CAC payback number. They will quote a confident figure — "6 months" or "9 months" or "11 months" — and present it as the headline unit-economics metric that justifies the company's growth trajectory. In nine cases out of ten, that number is wrong by a factor of 2-3x.

The reason: the founder is reporting narrow CAC payback (paid acquisition spend divided by new paying customers, divided by monthly gross profit per customer) when they should be reporting Fully-Loaded CAC payback (which includes free-tier infrastructure burden, free-user support cost, sales/CS touch on Product-Qualified Leads, and brand/content investment attributable to free signups).

Patrick Campbell's ProfitWell research (2014-2024, n=10,000+ SaaS companies analyzed) consistently shows that 70%+ of PLG founders report payback periods that are 40-60% optimistic relative to a fully-loaded reconstruction — a systematic measurement gap that misleads board decisions, fundraising narratives, and growth-vs-efficiency tradeoffs.

The narrow number isn't a lie — it's a useful diagnostic for channel-specific paid-acquisition efficiency. But it's not the all-in unit-economics number that should drive capital allocation. When boards read "9 months payback" and authorize aggressive growth spending, they're often authorizing spend against an unrealistic baseline.

When the fully-loaded reconstruction shows 22 months, the calculus changes — that's still a healthy business, but the growth-vs-efficiency tradeoff looks meaningfully different. Bessemer Venture Partners codified this distinction in their State of the Cloud annual reports (2019-2024), and OpenView Partners has tracked it explicitly in their Product-Led Growth Index (annual 2020-2024, n=600+ companies).

The discipline of dual-reporting — quoting both narrow and fully-loaded payback — is the single highest-leverage improvement most freemium businesses can make to their financial reporting in 2026, and it's one of the cheapest forms of executive maturity (it costs nothing but takes founder willingness to publish a less-flattering number alongside the flattering one).

The over-reporting problem is particularly acute in 2024-2026 because the SaaS efficiency reset (post-2022 valuation correction) has made unit economics more important than growth metrics in fundraising conversations. David Sacks' Burn Multiple framework (introduced in 2020, dominant since the 2022 correction) and Jason Lemkin's 2024 SaaStr commentary have repositioned capital efficiency as the dominant lens through which Series B/C/D rounds get evaluated.

A founder who quotes a narrow CAC payback of 6 months and then runs into a fully-loaded reality of 18 months when the investor's diligence team rebuilds the math has a credibility problem that can sink a round. The pre-emptive defense: publish both numbers, with a clear methodology footnote, and let the diligence team verify your math rather than challenge your math.

This is the same discipline that public-company CFOs apply to non-GAAP metrics (every adjusted-EBITDA disclosure comes with a reconciliation to GAAP) — PLG private-company CFOs and founders should adopt the equivalent reconciliation for narrow-vs-fully-loaded CAC. The cost is small (a methodology page in your investor deck), the benefit is large (you control the narrative rather than the diligence team controlling it).

The companies that have institutionalized this — Datadog, Atlassian (historically), MongoDB, Cloudflare — have used the discipline as a competitive advantage in capital markets, signaling sophistication in a category (SaaS metrics) where sophistication varies widely.

The structural differences between freemium and sales-led CAC math

Sales-led CAC math is conceptually simple. A sales team spends $X on outbound prospecting + AE compensation + marketing-qualified-lead generation. The team closes Y deals.

CAC = $X / Y. Payback = CAC / (monthly gross profit per deal). The denominator (paying customers) is the same as the cohort whose CAC you computed.

Salesforce, Workday, Snowflake, ServiceNow, Oracle, and SAP all use variants of this math, and the typical sales-led SaaS payback period is 18-24 months (per their 10-K disclosures). The cost stack is clearly attributable — every SDR salary, every AE commission, every marketing event, every demo-generation campaign maps to a closed deal in the cohort.

There's no ambiguity about whether a particular cost belongs in CAC. Freemium inverts this. In a freemium PLG model, 95-98% of acquired users never pay, and the cost of supporting those free users (infrastructure, support, engineering, brand, content, SEO) is real but mis-categorized.

The narrow CAC view treats only the paid acquisition spend (Google Ads, LinkedIn Ads, content syndication, partnerships) as CAC, and divides by the small fraction (typically 2-5%) of free users who convert to paid. This makes the headline CAC look very low — because the spend is divided by a fraction.

But this view externalizes the 8-25% of revenue burned on free-tier infrastructure as "marketing" or "R&D" or "G&A" rather than as CAC. The fully-loaded view re-internalizes those costs, recognizing that free-tier infra is distribution cost — the price of operating the funnel that produces the paying customers — and therefore part of CAC.

The structural inversions are three. First, freemium has a conversion rate layer that sales-led doesn't. In sales-led, every MQL is a future deal candidate; the conversion question is "will this MQL close?" not "is this MQL even a candidate?" In freemium, every signup is a future paying customer candidate, but the distribution is heavily skewed — 95-98% never pay, and the 2-5% who do pay must cover the cost of supporting the 95-98% who don't.

The math has to allocate non-converter costs to the converter cohort. Second, freemium has a time-to-conversion distribution that sales-led doesn't. Sales-led deals close on a roughly predictable cycle (30-90 days for SMB, 6-18 months for enterprise).

Freemium conversions distribute across 12+ months — 50% within 30 days of signup, 75% within 90 days, 90% within 12 months (per PostHog 2023 data), with a long tail extending 24+ months. This means cohort-CAC payback math has to handle the timing asymmetry between when CAC is spent (mostly upfront acquiring the free user) and when revenue arrives (spread across 12+ months as users convert).

Third, freemium has a viral coefficient layer that sales-led doesn't. Every Loom recipient, every Figma collaborator, every Slack guest, every Calendly bookee is a downstream signup vector. This creates a compounding-return profile: $1 of paid acquisition that drives 1 signup may, via viral spread, drive 0.3-0.5 additional organic signups within 90 days (typical K-factor for PLG products per Reforge benchmarks).

The "true CAC" should include attribution for organic/viral signups caused by paid signups, which simple CAC math doesn't capture. The most rigorous PLG companies (Figma at scale, Loom, Calendly) explicitly model viral-attribution-adjusted CAC, which can be 30-50% lower than the unadjusted number — but this requires sophisticated cohort analytics tooling that most early-stage companies don't have.

What counts as CAC when 95%+ of users never pay

The central question of freemium CAC measurement: what costs should be allocated to CAC when the vast majority of your "acquired" users never become customers? The answers vary across frameworks. Bessemer's Fully-Loaded definition (the strictest): include (a) all paid acquisition spend (Google, LinkedIn, content, partnerships, events), (b) free-tier infrastructure cost amortized over the paying cohort (typical 8-25% of revenue per Bessemer 2023), (c) sales/CS touch cost on Product-Qualified Leads (the human attention given to converting free-to-paid leads), (d) free-user support burden (typical 10-30% of support tickets come from non-paying users per ProfitWell 2022), (e) brand/content/SEO investment attributable to free signups (a significant chunk of organic free signups come from content marketing that's funded as a separate budget line).

The total fully-loaded CAC is typically 2-3x the narrow CAC for compute-light products (CRMs, calendars, productivity) and 3-4x for compute-heavy products (video, ML, databases, analytics). Wes Bush's Product-Led formula (more permissive): include paid acquisition + free-tier infra, but exclude support and brand/content as "operating expenses" rather than CAC.

This produces a middle-ground number, typically 1.5-2.5x narrow CAC. Patrick Campbell's ProfitWell methodology (most flexible): allow companies to choose their definition but require a methodology footnote disclosing what's included and excluded. The discipline of disclosure matters more than the specific choice — investors can adjust for definitional differences if they know what they're adjusting from.

The OpenView Partners PLG Index (2020-2024 annual benchmarking, n=600+ companies) splits the difference by tracking three definitions in parallel: (1) Narrow CAC = paid acquisition / new paying customers; (2) S&M CAC = (paid acquisition + sales + marketing OpEx) / new paying customers; (3) Fully-Loaded CAC = (S&M CAC + amortized free-tier infra + free-user support allocation) / new paying customers.

The 2024 PLG Index median values: Narrow CAC payback 8 months, S&M CAC payback 14 months, Fully-Loaded CAC payback 22 months. Each metric tells a different story. Narrow CAC measures channel efficiency (is paid acquisition working?).

S&M CAC measures go-to-market efficiency (is the whole revenue function efficient?). Fully-Loaded CAC measures business unit economics (does the freemium model actually pencil?). All three are useful; the failure mode is to report only narrow CAC and call it "CAC payback" without qualification.

The convention that the most rigorous PLG companies are adopting in 2024-2026 is to report all three in their investor decks, with footnotes explaining the methodology, and to set internal targets for each (e.g., "narrow CAC payback under 12 months, fully-loaded under 24 months").

This level of disclosure has historically been associated with public companies (Datadog, MongoDB, Atlassian, Cloudflare have all done variants of this in their public filings), but is increasingly standard in private-company Series B+ fundraising materials.

🔍 PART 2 — THE FRAMEWORK

Bessemer's Fully-Loaded CAC definition and why it matters

Bessemer Venture Partners introduced the Fully-Loaded CAC concept in their early State of the Cloud reports (initially 2018-2019 era) and has refined it across the 2019-2024 annual cycle. The definition: all costs incurred to acquire a new paying customer, including (a) paid media and content acquisition spend, (b) sales and customer-success compensation and overhead attributable to converting free-to-paid, (c) free-tier infrastructure cost amortized over the paying cohort using a defensible attribution methodology (typically: total free-tier infra cost for period / total new paying customers in period), (d) free-user support cost similarly amortized, (e) brand/content/SEO investment where the primary driver of return is free signups (not, e.g., enterprise brand awareness for a separate sales-led motion).

Bessemer's rationale: the freemium funnel is a single integrated economic unit, and excluding the cost of the top-of-funnel (free users) from the cost of the bottom-of-funnel (paying customers) creates an accounting fiction that misleads decision-making. The cost of a freemium business is the cost of running the entire funnel; dividing that cost by the small slice that pays is the only honest way to compute CAC.

The formula, in Bessemer's preferred notation: Fully-Loaded CAC = (Paid Acq + S&M OpEx + Amortized Free-Tier Infra + Amortized Free-User Support + Attributable Brand/Content) / New Paying Customers in Period. Payback = Fully-Loaded CAC / (Monthly ACV × Gross Margin).

Why it matters: the 2-3x multiplier between narrow CAC and fully-loaded CAC is the systematic measurement gap that most freemium founders miss. A company with narrow CAC payback of 6 months and a 2.5x multiplier has a fully-loaded payback of 15 months — still healthy, but the growth-vs-efficiency tradeoff looks meaningfully different.

A company with narrow payback of 9 months and a 3x multiplier has a fully-loaded payback of 27 months — borderline unhealthy, and the founder may need to slow growth spending to optimize before scaling. The Bessemer framework forces this conversation to happen with honest numbers, rather than letting founders chase narrow-CAC metrics that hide structural inefficiencies.

The framework is most rigorously applied at the public-company level — Datadog, MongoDB, Atlassian, Cloudflare, HubSpot all use variants of fully-loaded CAC in their investor disclosures — but is increasingly standard at the late-stage private level. Series B+ companies that don't apply it are increasingly seen as immature in their unit-economics reporting.

The cost of adopting Bessemer's framework is low (a methodology page in your investor deck and a quarterly reconciliation), and the credibility benefit is high (you signal sophistication in a category where sophistication varies widely).

Wes Bush's PLG CAC formula and common errors

Wes Bush, founder of the Product-Led Institute and author of Product-Led: How to Build a Product That Sells Itself (2019), articulates a slightly more permissive PLG CAC formula: PLG CAC = (Paid Acquisition + Free-Tier Infrastructure) / New Paying Customers. Bush's rationale: include the costs that are uniquely attributable to running a freemium funnel (infra for free users is the largest such cost), but exclude operating expenses (support, brand, content) that would exist whether or not you had a freemium model.

This produces a CAC number that's typically 1.5-2.5x the narrow CAC, sitting between the narrow definition and Bessemer's strictest fully-loaded definition. Bush's framework is most useful for internal management (what's the marginal cost of acquiring our next paying customer through the freemium funnel?), while Bessemer's strictest framework is most useful for external reporting (what does our business unit-economics look like to an investor doing diligence?).

Both are valid; the choice depends on the use case.

Bush identifies five common errors in PLG CAC calculation, drawn from his consulting work and the Product-Led Institute's research: (1) Excluding free-tier infra entirely — the #1 error, makes narrow CAC look 50-70% lower than reality. Founders often categorize free-tier infra as "R&D" or "cost of goods sold" rather than as CAC, which understates true acquisition cost.

(2) Using current-MRR not gross-profit in the denominator — this overstates payback velocity by 15-25% because gross margin (typically 75-85% for SaaS) is the actual cash available to repay CAC, not the topline revenue. Always use ACV × Gross Margin / 12 in the payback denominator.

(3) Using blended CAC across all channels rather than channel-cohort CAC. This masks unhealthy channels behind healthy ones — your overall CAC payback may be 12 months, but your paid-search CAC payback may be 24 months while your viral/organic payback is 4 months. Blended numbers hide the channel-mix problem.

(4) Using lifetime-paid-cohort not new-cohort in the numerator — backward-looking math that doesn't reflect current-period economics. Always use current-period cohort CAC (this quarter's CAC for this quarter's new paying customers), not historical-cumulative CAC. (5) Ignoring the time-to-conversion distribution — assuming all CAC is recovered linearly when reality is that 50% of conversions happen in month 1-3 and the rest spread across 12+ months.

Cohort-based payback curves are non-linear and require proper cohort accounting tools (ChartMogul, Baremetrics, ProfitWell) to compute correctly.

Patrick Campbell on cohort CAC math and the time-to-convert distribution

Patrick Campbell, founder of ProfitWell (acquired by Paddle in 2022 for $200M), has published extensive research on SaaS pricing and CAC math drawn from n=10,000+ SaaS companies analyzed 2014-2024. His core insight on freemium CAC: the time-to-conversion distribution is the dominant variable that founders ignore.

In a sales-led model, deals close on a roughly predictable cycle (30-90 days for SMB, 6-18 months for enterprise), and CAC is recovered on a predictable cadence. In a freemium model, conversions happen on a long-tail distribution: 50% within 30 days of signup, 75% within 90 days, 90% within 12 months, with the remainder extending 24+ months.

This means a cohort of 100 free users acquired in January will produce paying customers across the entire next year, with the timing weighted toward the early months but with meaningful conversions still happening in months 9-12. The implication: CAC payback can't be computed as a simple division because the revenue arrives non-linearly.

Proper cohort analysis requires computing the cumulative gross profit per cohort by month and identifying the month at which cumulative gross profit equals cumulative CAC. This is what ChartMogul's cohort retention curves, Baremetrics' CAC dashboards, and PostHog's product-attributed cohort analysis are designed to compute.

The companies that get this right (Datadog, Notion, Figma, Cloudflare) have built internal data infrastructure that produces cohort-CAC payback curves; the companies that get this wrong (most early-stage PLG SaaS) compute a static blended number that masks the underlying cohort dynamics.

Campbell's research also identifies the conversion rate elasticity issue. The narrow CAC formula assumes a stable conversion rate from free to paid (e.g., 3.5% of free users convert). In reality, conversion rates vary by cohort, by channel, by season, by product release, and by macroeconomic conditions.

ProfitWell's 2023 benchmark data shows that conversion rates can vary by 30-50% across cohorts even within the same company, and this variability dominates CAC computation. The implication: a single "average" CAC payback number is misleading because it averages over cohorts with very different economics.

The fix: report CAC payback by cohort vintage (Q1 cohort, Q2 cohort, Q3 cohort) and identify whether the trajectory is improving (conversion rates rising, payback shortening), stable (flat trajectory), or degrading (conversion rates falling, payback extending). The trajectory is more important than the level — a 24-month payback that's improving 5%/quarter is healthier than an 18-month payback that's degrading 5%/quarter.

This is the kind of dynamic, cohort-aware analysis that requires the right tooling (Mosaic, OnPlan, Pigment for strategic finance; ChartMogul, Baremetrics for transactional cohort data; PostHog, Mixpanel, Amplitude for product-attributed cohort behavior).

Reforge / Kyle Poyar on channel-cohort attribution and the viral coefficient adjustment

Brian Balfour at Reforge and Kyle Poyar at OpenView (now writing Growth Unhinged) have both emphasized channel-cohort attribution as a critical refinement of freemium CAC math. The insight: blended CAC across all channels hides the channel mix problem. Your paid-search CAC may be $300 with 18-month payback; your viral-organic CAC may be $30 with 4-month payback.

Blending these into a $150 average with 12-month payback obscures the strategic decision (should you spend more on paid search or more on viral-feature investment?). Channel-cohort attribution requires tagging each free signup with its acquisition source (paid search, paid social, content/SEO, viral/referral, partnership, organic-direct) and computing CAC payback separately for each cohort.

The most sophisticated PLG companies do this at quarterly cadence and use the results to reallocate spend dynamically — shutting off underperforming channels and doubling down on overperforming ones. Brian Balfour's Reforge Four Fits Framework (Product-Market, Product-Channel, Channel-Model, Model-Market) makes channel attribution central: the Channel-Model fit dimension specifically tests whether each channel's CAC payback matches the company's monetization model.

A channel that produces signups who don't fit your monetization motion (e.g., paid social driving consumer-grade free users who never upgrade to your B2B paid tier) is a strategic mismatch even if the headline CAC looks attractive.

Kyle Poyar's Growth Unhinged (2021-present, drawing on his work at OpenView 2015-2023) has emphasized the viral coefficient adjustment. The standard CAC formula treats each new paying customer as the direct product of acquisition spend, but in PLG products with strong virality, a meaningful fraction of "paid" signups generate additional organic signups via the viral mechanism (Figma viewers, Loom recipients, Slack guests, Calendly bookees).

The viral-attribution-adjusted CAC formula: Effective CAC = (Paid Acq + Infra + Support + Brand) / (Direct Paying + K × Viral-Attributed Paying), where K is the viral coefficient (typical 0.3-0.5 for healthy PLG products per Reforge benchmarks). Effective CAC can be 30-50% lower than unadjusted CAC, reflecting the compounding-return profile of products with genuine virality.

The most rigorous PLG companies (Figma at scale, Loom, Calendly) explicitly model this and report viral-attributed CAC separately in their internal dashboards. This is sophisticated math that requires product-attribution tooling (PostHog, Mixpanel, Amplitude, Heap) and a defensible attribution methodology (most companies use last-touch with viral-uplift adjustment or multi-touch attribution with time-decay).

The companies that don't do this systematically understate the ROI of investing in viral features (Loom's external-recipient mechanism, Figma's viewer mode, Calendly's booking page), which has strategic implications for product roadmap prioritization.

🧪 PART 3 — THE EVIDENCE

Slack, Figma, Notion: best-in-class PLG payback benchmarks

The canonical best-in-class freemium CAC payback benchmarks come from a small set of household-name PLG companies whose financial details are partially public. Slack is the most-studied case. Per their 2019 IPO S-1 filings, Slack reported a CAC payback period of approximately 13 months for the sales-led component of their business (the enterprise tier sales motion).

Lenny Rachitsky's 2021 Slack deep-dive in Lenny's Newsletter estimated an additional 6-9 months of payback for the free-tier infrastructure burden, bringing the fully-loaded payback to approximately 19-22 months at IPO. The key driver of Slack's strong economics: a ~30% peak free-to-paid conversion rate (per Lenny's 2021 analysis), an order of magnitude above the typical PLG benchmark of 2-5%, achieved through the famous 10K-message-cap design that bound the freemium friction wall to the moment of peak value-creation density (teams becoming search-dependent).

Slack's free-tier infra cost was significant — reportedly $1M+/month at peak per industry estimates — but the conversion rate was so high that the fully-loaded payback remained healthy. The lesson: conversion rate, not free-tier generosity, is the dominant variable in freemium payback economics.

Figma achieved sub-18 months fully-loaded payback at scale, per industry estimates pre-the-attempted-Adobe-acquisition (announced September 2022, abandoned December 2023). Figma's design was an editor-vs-viewer asymmetry — 3 editors free + unlimited viewers — which kept free-tier infra cost low (viewers consume bandwidth but not the more expensive editor compute) while driving viral-coefficient amplification (every viewer is a downstream editor candidate).

Industry estimates put Figma's free-to-paid conversion at 12-18%, and the viral-attribution-adjusted effective CAC was reportedly 30-40% lower than the unadjusted number. Notion is in the 22-26 month fully-loaded payback range per multiple TechCrunch reports and industry estimates.

Notion's free tier is unusually generous (unlimited personal use), which has both costs (significant free-tier infra burden) and benefits (massive viral spread as individual heavy users bring Notion into their workplaces, where the team-paid motion kicks in). Notion's conversion rate is estimated at 8-12%, which is good-but-not-best-in-class, producing a healthy-but-not-exceptional payback.

All three companies share a common pattern: high free-tier generosity + high conversion rate + strong viral coefficient = healthy fully-loaded payback, even when free-tier infra costs are substantial. The strategic implication: free-tier generosity is not the enemy of payback; under-conversion is the enemy of payback.

Calendly, Loom, HubSpot freemium CRM: middle-of-the-pack examples

Calendly reported a CAC payback of approximately 14 months per their Crunchbase 2021 disclosures (which included some financial details from their Series B round). Calendly's model is essentially pure-PLG with minimal sales team: their free-to-paid conversion is approximately 7% (per Crunchbase 2021), their free-tier infra cost is low (calendaring is compute-light), and their viral coefficient is high (every Calendly booking page is a billboard for the product).

The 14-month payback is exceptional for a product with a relatively low conversion rate — the strong viral coefficient compensates for the modest conversion. Loom reported approximately 28 months fully-loaded payback per OpenView's 2022 case study. Loom's video-hosting infrastructure is compute-heavy (video bandwidth + storage + transcoding), which inflates the fully-loaded CAC even though paid acquisition spend is modest.

Loom's conversion rate is approximately 11% (per OpenView), which is good for PLG but the infra cost dominates the cost stack and extends payback. The lesson: infrastructure-heavy products have structurally longer payback periods even with strong conversion, and this requires either higher ACV (to compensate) or aggressive infra-cost optimization (compression, tiered storage, regional CDN).

HubSpot's freemium CRM (launched 2014) has produced a fully-loaded payback of approximately 22 months for the free-CRM-to-paid-Marketing-Hub conversion path, per their 2023 annual letter and related investor disclosures. HubSpot's model is unique: the free CRM is genuinely free forever with unlimited users + 1M contacts, but generates roughly 50% of their qualified leads for the paid Marketing Hub / Sales Hub / Service Hub products.

The free CRM functions as a marketing engine — the CAC of the free CRM signups is allocated to the marketing budget (since the primary purpose is lead generation for the paid products), and the conversion of free-CRM users to paid-Hub users is measured against that allocated CAC.

This is a sophisticated allocation methodology that requires careful attribution accounting, but it produces a defensible payback number for the freemium engine as a whole. The HubSpot pattern (free product as marketing engine for paid products) is increasingly common in 2024-2026 — Notion's free workspaces seeding paid team conversions, Cloudflare's free Workers tier seeding paid enterprise contracts, MongoDB's Atlas free tier seeding paid scale-up usage.

In all these cases, the proper CAC payback analysis requires multi-product attribution that's harder than single-product math but more honest about the actual economic engine.

The worked math walkthrough: narrow vs fully-loaded for a sample PLG SaaS

Consider a hypothetical PLG SaaS in 2026 with the following characteristics. Top-of-funnel: 100,000 new free users per month, acquired through a mix of organic SEO (40%), paid Google search (30%), content marketing (20%), and viral/referral (10%). Conversion: 3.5% of free users convert to paid within 12 months, producing 3,500 new paying customers per month.

Pricing: $50/month average ACV (monthly billing), $600 annual ACV, 80% gross margin. Acquisition spend: $400,000/month total ($300,000 paid Google + $100,000 content marketing). Free-tier infrastructure: 100,000 free users × $0.20/user/month = $20,000/month infra burden.

Sales/CS PQL touch: $30,000/month (one FTE plus tooling, focused on converting high-intent free users). Free-user support: $50,000/month (estimated 25% of total support burden allocated to free users). Brand/content allocation: $100,000/month attributable to free-signup generation (above the $100,000 already in paid acquisition).

Narrow CAC = $400,000 paid acquisition / 3,500 new paying customers = $114 per customer. Narrow CAC payback = $114 / ($50 × 0.80 gross margin) = $114 / $40 = 2.85 months.

This is the headline-friendly number that most PLG founders would quote. It implies extraordinarily healthy unit economics and would justify aggressive growth spending. But it externalizes the 8-25% of revenue burned on free-tier infrastructure, the PQL sales touch, the free-user support burden, and the content/SEO investment.

Fully-Loaded CAC = ($400,000 paid + $20,000 × 12-month-amortization-lookback infra = $240,000 + $30,000 PQL touch + $50,000 support + $100,000 content) / 3,500 new paying customers = $820,000 / 3,500 = $234 per customer. Fully-Loaded CAC payback = $234 / $40 = 5.85 months.

The fully-loaded number is 2.05x the narrow number. Still very healthy by 2024-2026 PLG benchmarks (OpenView's PLG Index 2024 median is 18-22 months fully-loaded), but meaningfully different from the headline number. A founder reporting 2.85 months payback to their board and a founder reporting 5.85 months payback are telling different stories about the business.

The 5.85-month number is the truthful one. The 2.85-month number is a useful diagnostic for paid-acquisition efficiency but not for overall business unit economics.

For high-infra-cost businesses (video, ML, databases, analytics), the multiplier expands. If the same hypothetical company were instead a video platform with $5/user/month infra cost instead of $0.20, the infra component would be $500,000/month × 12 = $6,000,000 amortized, producing a fully-loaded CAC of ($400K + $6M + $30K + $50K + $100K) / 3,500 = $1,880 per customer, with a payback of 47 months.

This is dangerously unhealthy and would force the company to either raise prices, optimize infra costs, or restructure the freemium funnel. The lesson: infrastructure costs dominate the cost stack for compute-heavy freemium products, and the narrow CAC view obscures this entirely.

2024-2026 OpenView PLG Index, ChartMogul, Baremetrics, ProfitWell benchmark data

The 2024 OpenView PLG Index (n=600+ PLG SaaS companies surveyed) reports the following benchmark distributions for freemium CAC payback. Narrow CAC payback: median 8 months, 25th percentile 5 months, 75th percentile 14 months. S&M CAC payback (paid acq + S&M OpEx / paying cohort): median 14 months, 25th percentile 9 months, 75th percentile 22 months.

Fully-Loaded CAC payback (S&M + infra + support + content allocation): median 22 months, 25th percentile 15 months, 75th percentile 36 months. Companies above 36 months fully-loaded are flagged as having dangerous unit economics and typically require restructuring (price increase, free-tier reduction, channel mix shift, or in extreme cases freemium kill).

Companies below 15 months fully-loaded are best-in-class and have substantial latitude to scale aggressively.

ChartMogul's 2024 SaaS Benchmarks (drawing on their customer base of ~5,000+ SaaS companies) reports similar but slightly different numbers, reflecting their narrower customer mix (more SMB-focused than OpenView). Median CAC payback (their methodology is between narrow and S&M): 11 months.

Baremetrics' 2024 benchmarks (Stripe-native cohort): median CAC payback 9 months for SMB SaaS, 16 months for mid-market SaaS. ProfitWell/Paddle's 2024 benchmarks (n=10,000+ companies): median Fully-Loaded payback 18 months, with significant variation by sub-segment (productivity SaaS 12 months, analytics SaaS 24 months, video/media SaaS 30+ months).

The cross-platform consensus: healthy freemium PLG businesses have Fully-Loaded payback in the 15-24 month range, with sub-15 months being best-in-class and 30+ months being structurally concerning. The benchmark stack should inform both internal target-setting and external benchmarking against industry peers.

The cohort dynamics matter more than the static benchmark. A company at 22-month fully-loaded payback that's improving 5%/quarter is healthier than a company at 18-month payback that's degrading 5%/quarter. The trajectory variables to track quarterly: conversion rate by cohort vintage, time-to-conversion distribution, channel mix and channel-specific CAC, viral coefficient (K-factor), gross margin trend, and free-tier infra cost per user.

The companies that institutionalize this dashboard cadence (typically Series B+ with a strong finance function) are systematically better-positioned for both operational decisions and fundraising conversations. The companies that don't get caught flat-footed when investors do diligence and rebuild the math themselves.

📈 PART 4 — THE RECOMMENDATION

Dual-reporting protocol: narrow + fully-loaded

The recommended discipline for any freemium PLG company in 2026: report both narrow CAC payback and fully-loaded CAC payback in every internal management review and every external investor update. The narrow number is useful for channel efficiency diagnostics (is paid acquisition working?).

The fully-loaded number is useful for business unit-economics evaluation (does the freemium model actually pencil?). Both are valid; neither alone is sufficient. The format that works well in practice: a one-page "CAC payback methodology and results" appendix to the investor deck or board memo, with three components.

(1) Definitions — narrow CAC = paid acquisition / new paying customers / gross profit per customer; fully-loaded CAC = narrow + free-tier infra + S&M OpEx + free-user support + attributable content. (2) Current period numbers — this quarter's narrow payback, S&M payback, fully-loaded payback, with year-ago and trailing-twelve-month comparisons.

(3) Cohort trajectory — chart showing fully-loaded payback by cohort vintage, identifying improving, stable, or degrading trends. This level of disclosure is increasingly standard at Series B+ and has become essential at Series C+. Companies that report only narrow CAC face credibility risks during diligence (investors rebuild the math and find the gap).

The internal target-setting that follows from dual-reporting: set explicit targets for both metrics, with the fully-loaded target as the primary capital-allocation gate. Example targets for a Series A PLG SaaS: narrow CAC payback under 12 months, fully-loaded CAC payback under 24 months, trajectory improving 3-5% per quarter.

If narrow is met but fully-loaded is missed, the diagnosis is usually free-tier inefficiency (infra cost too high, support burden too high, conversion rate too low) — and the fix is product/operational rather than marketing. If fully-loaded is met but narrow is missed, the diagnosis is usually channel inefficiency (paid acquisition channels not converting) — and the fix is channel mix optimization.

The dual-target framework gives the operating team clear signals about where to invest improvement effort. The single-metric framework (narrow only or fully-loaded only) collapses these distinct problems into a single signal that doesn't tell you what to do about it.

When CAC payback is the wrong KPI and what to use instead

CAC payback is the wrong KPI in three scenarios. (1) Early-stage PLG (pre-Series A) where cohort sizes are too small for statistical meaning. With 50-200 paying customers, cohort variance dominates and CAC payback numbers fluctuate wildly month-to-month.

Use LTV/CAC ratio (target 3x+ per Bessemer benchmark, 5x+ best-in-class) and Magic Number (target 0.7-1.0+ healthy) instead. These are smoother metrics that don't require large cohort sizes. (2) Usage-based pricing models (Stripe, Twilio, Snowflake, Datadog, MongoDB Atlas, AWS, GCP, Azure, Cloudflare) where ACV ramps over 12-24 months as customers grow their consumption.

CAC payback inherently lags revenue recognition in these models — a customer who signs up at $50/month MRR may ramp to $500/month within 18 months, making the early-period payback look slow even though the cohort LTV is excellent. Use cohort-LTV projection (forecast revenue ramp by cohort vintage) and revenue retention (NRR target 120%+ for usage-based) instead.

(3) Hybrid PLG-to-enterprise motions (Slack, Atlassian, Notion, Figma, Cloudflare, HubSpot) where free-tier users seed enterprise deals 18-36 months downstream through a separate sales-led motion. The narrow CAC payback for the freemium component misses the indirect attribution chain entirely.

Use multi-product/multi-motion cohort attribution with proper LTV modeling that accounts for the downstream enterprise conversion path.

The alternative KPI stack that the most rigorous PLG CFOs are using in 2024-2026, in addition to or instead of CAC payback: (a) Burn Multiple = Net Burn / Net New ARR — David Sacks framework, target <2.0x healthy, <1.0x best-in-class, dominant capital-efficiency metric since the 2022 SaaS correction; (b) LTV/CAC ratio — Bessemer/OpenView standard, target 3x+ healthy, 5x+ best-in-class; (c) Magic Number = (Net New ARR × 4) / S&M spend — David Sacks framework, target 0.7-1.0+ healthy, 1.0+ scale-aggressively, <0.5 efficiency-problem; (d) Net Revenue Retention (NRR) — annualized expansion + contraction + churn, target 110%+ for PLG, 120%+ best-in-class, dominant retention quality metric; (e) Rule of 40 — growth rate % + EBITDA margin % ≥ 40, dominant overall-health metric at scale.

The strongest financial reporting practice in 2026 PLG SaaS uses all of these in combination, with CAC payback as one of several signals rather than as the singular headline metric. This is what Datadog, MongoDB, Cloudflare, and Atlassian do in their public disclosures, and it's the standard that late-stage private companies should adopt.

Tooling stack for 2026 cohort CAC modeling

The recommended tooling stack for cohort CAC modeling in 2026, by stage. Early-stage (seed to Series A): ChartMogul (cohort retention + CAC by channel, $99-$999/month), Baremetrics (Stripe-native MRR/cohort/CAC tracking, $58-$229/month), ProfitWell (free metrics — acquired by Paddle 2022, now part of Paddle Pricing Studio).

These are turnkey tools that connect to Stripe/Recurly/Chargebee and produce cohort-CAC dashboards with minimal setup. Best for companies that need rigorous metrics but don't have a dedicated FP&A team. Mid-stage (Series A to Series C): add PostHog (product-event-attributed CAC + cohort, free tier generous), Mixpanel (event-based cohort, $24-$833/month), Amplitude (cohort + behavioral attribution, $61-$1,099/month).

These layer behavioral data on top of transactional data to enable channel-cohort attribution and viral coefficient adjustment. Best for companies that have started doing rigorous channel optimization and need product-attribution. Late-stage (Series C+): add Mosaic (strategic finance + FP&A automation, enterprise pricing), OnPlan (driver-based SaaS planning, $40K-$120K/year), Pigment (enterprise FP&A platform, enterprise pricing).

These are full strategic-finance platforms that consolidate cohort data, channel data, and forecasting into a single planning environment. Best for companies with a dedicated FP&A team and complex cross-product/cross-motion attribution requirements.

The integration architecture that works well in practice: Stripe (or equivalent billing) → ChartMogul/Baremetrics (cohort + retention) → PostHog/Mixpanel/Amplitude (product attribution) → Mosaic/OnPlan/Pigment (strategic finance + planning) → BI layer (Looker, Tableau, Hex, Sigma) for executive dashboards.

This stack costs $50K-$500K/year depending on company size and produces investor-grade cohort CAC analysis with attribution, viral coefficient adjustment, channel mix optimization, and scenario planning. The ROI is high: companies that institutionalize this stack make better growth-vs-efficiency tradeoffs, fundraise more efficiently, and avoid the credibility risks of reporting unreconstructed narrow CAC.

The companies that don't institutionalize this stack (most early-mid stage PLG SaaS) tend to fly blind on unit economics and get caught flat-footed in diligence. Cross-link to [q5547](/q/5547), [q6121](/q/6121), [q9472](/q/9472), [q4123](/q/4123) for related PLG benchmark methodology and cohort tooling deep-dives.

Counter-case: freemium is the wrong motion entirely

The counter-case for the entire freemium discussion: in three scenarios, freemium is the wrong acquisition motion and the CAC payback math is moot. (1) High-ACV enterprise sales (Salesforce $50K+ ACV, Workday, Snowflake at scale, Palantir, ServiceNow, Oracle, SAP, IBM enterprise software) — the target ICP buys through committee-driven procurement processes (CIO + legal + security + procurement + finance all involved), and a free tier creates SDR distraction (every "free signup" requires qualification overhead) without conversion lift (the actual buyer never personally signs up).

Salesforce's "Trailblazer" community and Snowflake's "$400 free credit" are the patterns that work for this ICP — bounded, time-limited, qualified trials rather than open freemium. The CAC payback math for enterprise sales-led is the traditional 18-24 months sales-led payback, computed as outbound spend / closed deals / gross profit.

(2) Regulated industries (healthcare, financial services, defense, education) where security reviews, compliance certifications, and procurement processes preclude self-serve adoption. In these industries, freemium creates security risk (unauthorized usage of data-sensitive products) without conversion benefit.

The right motion is enterprise sales-led with proper SOC 2 / HIPAA / FedRAMP / FERPA certifications and a sales team that can navigate compliance gates. CAC payback math is traditional sales-led 24-36 months given the longer sales cycles. (3) Complex security/integration products where the product can't be evaluated without deep technical integration (e.g., observability platforms at scale, data warehouse / lakehouse products, low-level infrastructure tools).

In these cases, a free tier either fails to deliver the aha moment (the integration takes 4-12 weeks and free users abandon before completing it) or commoditizes the perception of the product (free tier feels like a toy compared to the actual production deployment). The right motion is technical sales-led with a free credit / proof-of-concept gate (Snowflake's $400 credit, Datadog's 14-day trial, Databricks' free trial).

CAC payback is computed against the trial-to-paid conversion, with the trial cost included in CAC.

When freemium cannibalizes vs feeds paid: a freemium tier feeds paid when the free product solves the small job and the paid product solves the bigger job (HubSpot CRM free + Marketing Hub paid; Notion personal free + team paid; Figma editor free + team/org paid). A freemium tier cannibalizes paid when the free product solves the whole job and removes upgrade incentive (the "MailChimp 2,000 contacts forever" pattern where 90%+ of users never convert because free is too good).

The diagnostic question: "what percentage of free users could theoretically run their entire workflow on the free tier indefinitely?" If above 70%, the free tier is too generous and is cannibalizing paid revenue rather than feeding it. The fix: tighten the free tier (gradually, with advance notice, with low-cost migration tiers) or restructure the paid tier value proposition.

MailChimp reduced from 2,000 to 500 contacts in 2022, triggering massive user backlash + competitive defection — a cautionary tale of how to do this badly. Cross-link to [q3215](/q/3215), [q4789](/q/4789), [q6234](/q/6234), [q7345](/q/7345), [q8567](/q/8567).

🔄 Fully-Loaded CAC Payback Computation Flow

🎯 Decision Tree: Is Freemium Right + How to Measure?

📚 Sources & References

Foundational Frameworks

• Bessemer Venture Partners — State of the Cloud annual report — https://www.bvp.com/atlas/state-of-the-cloud-2023
• OpenView Partners — Product-Led Growth Index — https://openviewpartners.com/blog/product-led-growth-index
• OpenView Partners — Annual SaaS Benchmarks — https://openviewpartners.com/saas-benchmarks-report
• Wes Bush — Product-Led: How to Build a Product That Sells Itself (2019) — https://productled.com
• Patrick Campbell — ProfitWell/Paddle pricing research — https://www.paddle.com/blog/pricing-strategy-saas
• Brian Balfour — Reforge Four Fits Framework — https://www.reforge.com/blog/four-fits-growth-framework
• Kyle Poyar — Growth Unhinged — https://www.growthunhinged.com
• David Sacks — Burn Multiple framework (Craft Ventures) — https://sacks.substack.com/p/the-burn-multiple
• David Sacks — SaaS Magic Number — https://www.craftventures.com/saas-metrics

Industry Benchmarks & Reports

• OpenView 2024 SaaS Benchmarks — https://openviewpartners.com/2024-saas-benchmarks-report
• ChartMogul 2024 SaaS Benchmarks — https://chartmogul.com/reports/saas-benchmarks
• Baremetrics SaaS Benchmarks — https://baremetrics.com/saas-benchmarks
• ProfitWell / Paddle Studies — https://www.paddle.com/resources/pricing-research
• Lenny's Newsletter PLG case studies — https://www.lennysnewsletter.com
• First Round Review — SaaS unit economics — https://review.firstround.com
• a16z Future — PLG essays — https://future.com
• SaaStr — Jason Lemkin 2024 PLG benchmarks — https://www.saastr.com

Public Company 10-K & S-1 Disclosures

• Slack S-1 (2019 IPO) — https://www.sec.gov/Archives/edgar/data/1764925/000162828019005385/works-s1.htm
• Datadog 10-K — https://investors.datadoghq.com/financials/sec-filings
• MongoDB 10-K — https://investors.mongodb.com/financials/sec-filings
• Atlassian 10-K — https://investors.atlassian.com/financials/sec-filings
• HubSpot 10-K and 2023 annual letter — https://ir.hubspot.com
• Cloudflare 10-K — https://cloudflare.net/financials/sec-filings
• Snowflake S-1 (2020 IPO) — https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001640147

Best-in-Class PLG Case Studies

• Lenny Rachitsky 2021 Slack deep-dive — https://www.lennysnewsletter.com/p/how-slack-grew
• Figma growth analysis — https://www.figma.com/blog
• Notion growth + freemium analysis — https://www.notion.so/blog
• Loom OpenView 2022 case study — https://openviewpartners.com/blog/loom-pricing
• Calendly Crunchbase 2021 data — https://www.crunchbase.com/organization/calendly

Cohort CAC Tooling

• ChartMogul — https://chartmogul.com
• Baremetrics — https://baremetrics.com
• ProfitWell / Paddle Pricing Studio — https://www.paddle.com/products/profitwell-metrics
• Mosaic — https://www.mosaic.tech
• OnPlan — https://www.onplan.co
• Pigment — https://www.pigment.com
• PostHog — https://posthog.com
• Mixpanel — https://mixpanel.com
• Amplitude — https://amplitude.com

Alternative SaaS Metrics

• SaaS Magic Number explained — https://www.bvp.com/atlas/saas-magic-number
• Rule of 40 (Brad Feld popularization) — https://feld.com/archives/2015/02/rule-40-healthy-saas-company.html
• Net Revenue Retention benchmarks — https://www.scaleventurepartners.com/perspectives/net-dollar-retention
• LTV/CAC ratio framework — https://www.forentrepreneurs.com/saas-metrics-2

📊 Numbers Block

2024 OpenView PLG Index — CAC Payback by Definition (n=600+ companies)

Best-in-Class PLG Cohort Benchmarks (Industry Estimates 2024)

Worked Math Walkthrough — Hypothetical PLG SaaS

Alternative KPI Benchmark Stack (2024-2026)

Time-to-Conversion Distribution (PostHog 2023, Freemium SaaS Aggregate)

12-Element Pulse Counter

6-Condition Verdict — Is Your Freemium CAC Payback Healthy?

Verdict scoring: 5-6 Pass = best-in-class. 3-4 Pass = healthy. 1-2 Pass = restructure needed. 0 Pass = freemium is wrong motion or business needs urgent intervention.

⚠️ Counter-Case: When CAC Payback is the Wrong Frame

The honest reality: CAC payback is the most over-used and most mis-applied SaaS metric in 2026 PLG fundraising. There are four scenarios where the metric breaks down entirely and continuing to optimize for it produces strategic damage.

Failure Mode 1: Early-Stage Cohort Size Below Statistical Meaning

Pre-Series A PLG SaaS with 50-200 paying customers cannot produce statistically meaningful CAC payback numbers because cohort variance dominates the signal. A quarter where one channel had a fluky conversion rate (a viral tweet, a Hacker News feature, an SEO algorithm shift) can swing the headline payback number by 30-50% without any change in underlying business health.

Founders who chase this number quarter-to-quarter end up making operating decisions based on noise rather than signal. The fix: at this stage, use LTV/CAC ratio (smoother, target 3x+) and Magic Number (target 0.7-1.0+) as the primary efficiency metrics, and treat CAC payback as a directional indicator only.

Wait until you have 500+ paying customers and 4+ quarters of cohort data before treating CAC payback as a primary KPI. ProfitWell's research (n=10,000+ SaaS) confirms this: sub-500-customer SaaS show 50%+ quarter-over-quarter variance in narrow CAC payback, making the metric essentially unusable for management decisions.

Failure Mode 2: Usage-Based Pricing with Long ACV Ramp

Stripe, Twilio, Snowflake, Datadog, MongoDB Atlas, AWS, GCP, Azure, Cloudflare — these usage-based pricing companies have an ACV ramp that extends 12-24 months as customers grow their consumption. A customer signing up at $50/month MRR may ramp to $500/month within 18 months. The narrow CAC payback formula uses current MRR in the denominator, which understates the customer's lifetime contribution and overstates the payback period in early periods.

The correct math: cohort-LTV projection (forecast revenue ramp by cohort vintage), which requires a defensible usage-growth model. Snowflake's S-1 (2020) explicitly broke out this dynamic, showing that cohort revenue grew 158%+ NRR over the first 24 months — far above the headline current-period payback would suggest.

For usage-based businesses, the dominant KPI stack is NRR (target 120%+) + cohort-LTV projection + gross margin expansion trajectory, with CAC payback as a secondary diagnostic.

Failure Mode 3: Hybrid PLG-to-Enterprise with Multi-Year Downstream Attribution

Slack, Atlassian, Notion, Figma, Cloudflare, HubSpot — these companies have hybrid motions where freemium users seed enterprise deals 18-36 months downstream through a separate sales-led motion. The freemium CAC payback misses the indirect attribution chain entirely. A company tracking only freemium-component CAC payback may conclude their freemium economics are weak when, in reality, the freemium engine is producing the enterprise pipeline that drives the bulk of revenue 24+ months later.

Slack's S-1 showed exactly this dynamic: the freemium motion drove top-of-funnel that the sales-led motion converted into enterprise contracts. Computing only the freemium CAC payback understated the strategic value of the freemium engine by 50-70%. The fix: multi-product/multi-motion cohort attribution that tracks free signups through to enterprise conversion across the full 24-36 month attribution window.

This requires sophisticated tooling (Mosaic, OnPlan, Pigment) and explicit attribution methodology, but it's the only honest way to evaluate hybrid PLG-to-enterprise economics.

Failure Mode 4: Capital-Efficient Era Demands Burn-Centric Metrics

The 2022 SaaS valuation correction has fundamentally repriced the importance of capital efficiency vs growth in venture markets. David Sacks' Burn Multiple framework (Net Burn / Net New ARR, introduced 2020, dominant since 2022) has emerged as the primary capital-efficiency KPI in Series B+ fundraising conversations.

Companies with strong CAC payback but weak Burn Multiple (high cash consumption relative to ARR creation) face fundraising headwinds even with apparently-healthy unit economics. Conversely, companies with weaker headline CAC payback but stronger Burn Multiple are increasingly favored.

The implication: don't optimize CAC payback in isolation — optimize the full metric stack (Burn Multiple, LTV/CAC, Magic Number, NRR, Rule of 40) with explicit awareness of which metric matters most for your stage and fundraising trajectory. Jason Lemkin's 2024 SaaStr commentary explicitly recommends this multi-metric optimization, arguing that single-metric optimization (CAC-only, growth-only, NRR-only) is the cause of most preventable PLG SaaS failures.

Failure Mode 5: Freemium is the Wrong Motion Entirely

In three ICP/motion scenarios, freemium is the wrong choice and the CAC payback math is moot. (a) High-ACV enterprise sales ($50K+ ACV like Salesforce, Workday, Snowflake) — committee-driven procurement processes preclude meaningful self-serve conversion; use bounded credit trials instead.

(b) Regulated industries (healthcare, financial services, defense) where security reviews and compliance certifications preclude self-serve adoption; use enterprise sales-led with proper certifications. (c) Complex security/integration products where evaluation requires deep technical integration (data warehouses, observability platforms, infrastructure tools); use technical sales-led with proof-of-concept trials.

In all three scenarios, the dominant CAC payback math is traditional sales-led 18-36 months, computed as outbound spend / closed deals / gross profit. Trying to force a freemium model onto these scenarios produces high free-tier costs without conversion benefit and structurally damages unit economics.

The discipline of motion-selection (freemium vs trial vs sales-led) before measurement-selection (CAC payback vs alternatives) is the single highest-leverage strategic decision a PLG founder makes.

🔗 Cross-Links

Related Pulse library entries:

• [q5547 — PLG conversion rate benchmarking methodology](/q/5547)
• [q6121 — Activation rate optimization for B2B SaaS](/q/6121)
• [q9472 — Aha moment definition + measurement framework](/q/9472)
• [q4123 — Unit economics modeling for freemium products](/q/4123)
• [q5891 — Free user infrastructure cost allocation](/q/5891)
• [q8234 — Support cost benchmarks for PLG companies](/q/8234)
• [q9321 — Engineering time allocation between free/paid features](/q/9321)
• [q3215 — When to transition from PLG to sales-led motion](/q/3215)
• [q4789 — Free tier sunset playbook (Heroku case study)](/q/4789)
• [q6234 — Pricing tier design + willingness-to-pay segmentation](/q/6234)
• [q7345 — Enterprise tier feature checklist (SSO/SCIM/audit)](/q/7345)
• [q8567 — Free-to-paid conversion funnel optimization](/q/8567)
• [q1234 — Viral coefficient (K-factor) measurement + improvement](/q/1234)
• [q2345 — Editor vs viewer asymmetry in collaboration tools](/q/2345)
• [q3456 — Pricing model selection (subscription vs usage vs hybrid)](/q/3456)
• [q4567 — Slack's growth playbook deep dive](/q/4567)
• [q5678 — Figma's PLG motion analysis](/q/5678)
• [q6789 — Notion's freemium architecture review](/q/6789)
• [q7890 — Loom's dual-trigger free tier teardown](/q/7890)
• [q672 — Free tier design rules: seat limits / feature gates / usage limits](/q/672)
• [q3579 — Burn Multiple framework (David Sacks methodology)](/q/3579)
• [q4680 — LTV/CAC ratio computation and benchmarks](/q/4680)
• [q5791 — Magic Number explained + benchmarks](/q/5791)
• [q6802 — Net Revenue Retention measurement for PLG](/q/6802)
• [q7913 — Rule of 40 explained + scaling implications](/q/7913)