How do we calculate freemium-to-paid conversion CAC payback when self-serve acquisition cost is near-zero in 2027?

The outcome you should expect when you price the free tier honestly

The first thing that happens when a RevOps team runs this method properly is that the reported number gets worse — often dramatically. That is the intended outcome, and it is worth setting expectations before anyone starts, because a team that expects the exercise to validate their existing deck will abandon it halfway through when the arithmetic turns hostile.

Here is the concrete shape of the change. Take a freemium product-led-growth motion that acquired 10,000 free signups in a month and converted 600 of them to paid over the following 180 days — a 6% blended free-to-paid rate, which sits at the healthy end for a permanent free tier with no time limit. If the only cost you count is attributable paid acquisition spend of $48,000, your CAC reads $80 per paying customer. Divide that by $90 of monthly ARPU and you get a 0.89-month payback. Twenty-seven days. That number goes into a board deck and nobody in the room asks a hard question, because a sub-one-month payback sounds like a rocket ship.

Now add the costs the standard formula has no slot for. Free-tier infrastructure for those 10,000 users across the six-month conversion window: $61,000. Free-user support plus abuse and fraud handling: $9,400. Onboarding, lifecycle email, and in-product education tooling: $7,200. Human-assist touch from CS and AE time spent on the roughly 1,400 product-qualified leads the cohort flagged: $22,000. Payment processing and involuntary-churn recovery on the 600 converters: $4,100. Total fully-loaded acquisition cost of $151,700, producing 600 payers, which is $252.83 per paying customer.

Then fix the denominator. You do not recover acquisition cost out of revenue — you recover it out of gross profit, because the cloud bill for serving that paying customer comes out of the same dollar. At a 78% gross margin, $90 of ARPU returns $70.20 per month toward CAC recovery. The honest payback is $252.83 ÷ $70.20 = 3.60 months.

Same business, same month, same raw data. The naive method said 27 days. The honest method says 3.6 months — a four-fold difference produced entirely by measurement choice, before a single business decision changed. And the 3.6-month figure is still the optimistic end of the honest range, because it has not yet been corrected for the conversion-lag time-shift.

What you should expect operationally is a three-stage reaction. Week one, disbelief — someone will insist the free-infrastructure cost belongs in R&D. Week two, a productive argument about allocation percentage, which is the right argument to be having. Week three onward, a genuinely different set of product decisions: seat caps get revisited, usage quotas get tightened, and the question "should we make the free tier more generous?" stops being a pure growth question and starts being a unit-economics question with a number attached.

The adjacent effect worth naming is on hiring and budget. Teams that have never priced the free tier tend to over-invest in top-of-funnel acquisition, because acquisition looks nearly costless per marginal signup. Once free-infrastructure cost-of-carry is a visible monthly line, the marginal signup has a price, and growth spend gets reallocated toward activation and conversion — which is almost always where the leverage was hiding anyway. That reallocation is the real return on doing this work; the corrected metric is just the mechanism.

What drives the number: the six cost streams and the events that carry them

The formula is short. Getting honest inputs into it is the entire job.

Fully-loaded CAC payback (months) = fully-loaded CAC per paying customer ÷ (monthly ARPU of the paying cohort × gross margin %)

Where the numerator expands to: paid acquisition spend + free-tier infrastructure attributable to the converting cohort + free-user support and abuse handling + onboarding and lifecycle tooling + human-assist touch on PQLs + payment processing and fraud cost, all divided by the number of paying customers the cohort produced.

Each of those six streams hides in a different place in the general ledger, and that is precisely why the standard CAC formula — sales and marketing spend divided by new customers — never sees them.

Free-tier infrastructure lands in COGS or R&D hosting. It is the dominant hidden line and, for many products, larger than the entire paid-acquisition budget. Compute, storage, bandwidth, and increasingly model inference all scale with free-user engagement, which is exactly what a freemium funnel is designed to maximize. For AI-native products this has inverted the old cost structure: a heavily engaged free user running large-context inference calls can cost more per month than a low-tier paying customer generates. The gap between marketing-only CAC and fully-loaded CAC used to run two to three times for a conventional SaaS product; for inference-heavy products it can run far wider, and the error concentrates in the most engaged — most promising — part of the funnel.

Free-user support and abuse handling gets booked as support OpEx, categorized as a cost of running the product rather than a cost of acquiring customers. Fraud and abuse tooling on a free tier is pure acquisition cost: it exists because you let strangers in the door.

Onboarding, lifecycle email, and in-product education often sits in marketing ops or gets buried in product engineering. It is unambiguously acquisition cost — its entire purpose is converting a free user into a payer.

Human-assist touch on PQLs is the tricky one, because it *is* in the sales and marketing line — but only for the deals that closed. Every minute a CS rep or AE spends on a PQL who never converts gets silently dropped from the CAC calculation. If you flagged 1,400 PQLs and 600 converted, you spent human time on 800 people who never paid you, and that time is a real cost of producing the 600 who did.

Payment processing and involuntary-churn recovery is treated as a flat margin haircut rather than a cohort-traceable acquisition cost. At roughly 2.9% plus $0.30 per transaction, it is negligible at enterprise ARPU and meaningful at the $15–30 monthly price points where a lot of freemium conversion happens.

Every box in that diagram is an event you must log with a timestamp, because the allocation decisions downstream depend on knowing *when* things happened, not just that they happened.

The allocation question — how much of $61,000 in free-infrastructure cost should be charged to 600 payers when 10,000 people used the free tier — has three defensible answers and one indefensible one. Full-cohort amortization charges all of it to the payers, on the theory that the free tier is a channel whose purpose is producing buyers; this is the reporting default. Pre-conversion-only counts only the infrastructure each user consumed before their conversion date, which is more accurate but requires per-user usage timestamps; in the worked example it lands around $44,800 charged, giving a $226 CAC and a 3.22-month payback. Converter-attributable counts only the infrastructure of users who eventually converted, which is biased low because it pretends you could have run the funnel without the 9,400 who didn't. And exclude entirely — leave it in R&D — is the bug this whole method exists to fix.

Report a range, not a point. Charging 0% of free-infrastructure gives $151 CAC and a 2.15-month payback. Charging 50% gives $202 and 2.88 months. Charging 70% gives $222 and 3.16 months. Charging 100% gives $253 and 3.60 months. Publishing the 100% number as the headline with the pre-conversion refinement shown as the precision estimate — and the 0% row visible so everyone can see what the wrong method would have claimed — is the difference between a model a sophisticated investor trusts and one they discount on sight.

The last driver, and the one that survives even after teams adopt fully-loaded CAC, is conversion lag. Freemium users do not convert the month they sign up. They convert on a distribution with a median somewhere between 30 and 180 days for a true freemium product. If you divide this month's spend by this month's conversions, you are matching April's marketing budget against people who signed up in December and January. In a company growing spend 8% month over month with a 90-day median lag, the conversions arriving today were produced by spend roughly 22% lower than today's — and the resulting distortion runs well into double-digit percentages in either direction depending on whether you anchor on spend or on conversions.

The fix is trigger-and-trace: define the cohort by the signup event, trace forward to find who converted, then trace spend backward to the signup window that produced them. Numerator and denominator finally describe the same population. It requires a clean event log, and it is the only method that survives a diligence room where somebody rebuilds your number from the raw data.

Benchmarks, realistic ranges, and why the benchmark is the least useful number here

Reference points first, then the caveat that matters more than any of them.

Fully-loaded CAC payback by motion. Pure self-serve freemium selling to SMB generally lands in the 4–10 month band, with anything past roughly 14 months worth investigating. Freemium with a PQL-assist layer selling to mid-market runs longer — call it 8–15 months, cautionary past 20. Blended PLG plus a real sales-led motion sits around 12–20 months. Enterprise-heavy businesses that use PLG only as top-of-funnel commonly run 15–24 months and can be healthy there, because the contracts are large and durable. The published benchmark sets from Bessemer, OpenView, KeyBanc, and ICONIQ all cluster in these neighborhoods, though each uses slightly different definitions.

A payback under about four months on a genuine freemium product is, in my experience, usually a measurement artifact rather than an achievement — it almost always means free-infrastructure cost is sitting in R&D or the denominator is revenue instead of gross margin. Payback past 18–20 months means the free tier is subsidizing people who will never buy, and the fix is product design (seat caps, feature gates, usage quotas), not marketing.

Conversion-rate ranges. A true freemium tier with no time limit converts somewhere in the 2–5% range; the very large free bases that anchor the famous PLG case studies live at the low end of that and make it up on volume. A reverse trial — full premium access for 14–30 days, then automatic downgrade to a limited free tier rather than a paywall — consistently outperforms classic freemium, typically landing in the 8–25% band. A conventional time-limited free trial with no permanent free tier converts in the 15–25% range, but that is a different motion with different economics and should never be cohorted together with freemium users. PQL-to-paid is the number that actually predicts revenue, and it runs 25–50%; in the worked example, 600 of 1,400 PQLs converted, or 42.9%.

Gross margin. The 80%+ margins that defined the 2015–2021 SaaS era are compressing. A PLG business with meaningful third-party API or inference pass-through cost is realistically operating at 68–78% today, and that compression flows directly into payback: at 78% margin, $90 ARPU returns $70.20 monthly; at 68% it returns $61.20, which pushes the same $253 CAC from 3.60 months to 4.13. A ten-point margin move is a half-month payback move, and margin is moving.

The companion metrics. Payback reported alone is gameable and near-meaningless. Pair it with net dollar retention (105–130% for healthy PLG), burn multiple (under 1.5 is good, under 1.0 is elite), and gross margin. The combination classifies the business in a way no single number can. Fast payback with high NRR and a low burn multiple is the elite cell — the flywheel works, go raise money and spend it. Fast payback with sub-100% NRR is a leaky bucket, where quick re-acquisition is masking churn and the right move is fixing retention before spending another dollar on growth. Slow payback with high NRR and low burn is expansion-led and perfectly acceptable if you have the cash to fund the gap. Slow payback with low NRR and a burn multiple above 2.0 is distressed: every cohort loses money and never recovers it.

Now the caveat. Every range above is a blend of businesses that measured CAC five different ways, with different free-tier generosity, different price points, and different definitions of the word "freemium." A published median PLG payback of, say, eleven months is an average across companies where some counted free infrastructure and most didn't. Comparing your trigger-and-trace, fully-loaded, gross-margin-denominated number against that median is comparing two different quantities that happen to share a name — and you will lose that comparison every time, precisely because you did the work honestly.

Use benchmarks as a sanity bracket and nothing more. If you calculate a one-month payback or a forty-month payback, that is a signal to re-examine your method before you re-examine your business. Inside the range, the benchmark tells you nothing actionable. The only comparison that controls for method — and therefore the only one worth optimizing against — is your own cohort against your own prior cohorts.

That cohort-over-cohort read is where the real signal lives. Three consecutive monthly cohorts from the same product might show signups growing 8,400 → 9,100 → 10,000, conversion drifting 6.4% → 6.2% → 6.0%, ARPU flat at roughly $68 → $69 → $70, and fully-loaded CAC climbing $228 → $241 → $253, for paybacks of 3.35 → 3.49 → 3.60 months. The interesting row is free-infrastructure cost-of-carry per payer: $89 → $96 → $102. Payback is degrading not because monetization weakened but because the free tier is getting more expensive per buyer it produces. Without the three-cohort comparison, you'd see "3.6 months, still healthy" and miss a trend that becomes a real problem in two quarters.

Risks, edge cases, and the ways this metric gets quietly gamed

The method has real failure modes, and a practitioner should know them before defending the number in a room full of skeptics.

The strongest legitimate objection: amortizing free infrastructure over payers over-charges them. A sharp CFO will argue that the free tier exists for reasons beyond conversion — brand, network effects, competitive moat, recruiting, developer mindshare — and that charging 100% of its cost to 600 payers overstates what those specific customers cost to acquire. The objection has genuine merit at the margin. If your free tier demonstrably drives word-of-mouth that lowers your *paid* acquisition cost, a partial allocation is defensible. The reasonable position: default to full-cohort amortization for the reported number, and show sensitivity scenarios at 70% and 50% so the range is visible. What is never defensible is 0%. The free tier produces buyers; some fraction of its cost is a cost of producing buyers; that fraction is not zero. The debate is whether the right number is 60% or 100%.

Trigger-and-trace is too heavy for an early-stage company, and that's fine. A seed-stage business with 300 signups a month and a part-time finance function cannot stand up a seven-event log this quarter. For them, trigger-and-trace is the right destination and the wrong starting point. A trailing-three-month spend-to-conversion ratio that includes free-infrastructure cost as a flat estimate is good enough to avoid the catastrophic error, which is counting free as zero. Precision matters when you are raising a priced round or making real budget-allocation decisions — not before. Do not let the perfect method prevent the adequate one.

Some PLG investors argue payback is the wrong primary metric entirely, holding that NRR and burn multiple are what matter for a genuine bottoms-up motion and that CAC payback is a sales-led import founders over-index on. Partly right. If your free-to-paid motion truly runs at low marginal cost and NRR is 130%, then arguing about a three-versus-five-month payback is a distraction from expansion economics. But "less interesting" is not "free," and the teams who say payback doesn't matter for PLG are usually the same teams that never measured free-infrastructure cost in the first place. Use payback as a guardrail against the caution zones above; use NRR and burn multiple as the primary growth-quality signals.

Cohort construction failure modes are where most of the damage happens, and every one of them produces a number that *looks* like a freemium payback and isn't:

• *Cohorting by conversion month rather than signup month.* Spend gets misattributed, and payback mysteriously "improves" the month you cut marketing spend — a tell that should terrify anyone who sees it.
• *Computing payback across the full signup base.* Dividing acquisition cost by all 10,000 signups instead of the 600 payers collapses blended ARPU toward zero and produces a meaningless number.
• *Mixing trial users and freemium users in one cohort.* Trials convert fast and high; freemium converts slow and low. The blend describes neither motion and hides both.
• *Redefining "converted" mid-stream.* Counting a plan upgrade as a new conversion double-counts and drifts the rate upward artificially.
• *Counting reactivated churned customers as new conversions.* This inflates apparent acquisition efficiency with what is really re-acquisition of revenue you already lost.
• *Quietly extending the conversion window.* This is the subtlest one. A signup cohort is never truly finished converting — someone always upgrades 14 months later. Pick fixed checkpoints, typically 90 days for fast experiment feedback and 180 days for board reporting and payback, freeze those definitions, and never move them to flatter a quarter. Anything converting past the window is real revenue but belongs in expansion, not in new-conversion CAC.

The survivorship trap in the denominator. Monthly ARPU of the paying cohort is not constant across the payback window. Customers churn, and the survivors skew toward the happier, higher-ARPU accounts — so ARPU measured from customers still paying today overstates what the original cohort actually produced. For a per-customer payback figure, weight ARPU by the months each customer actually paid. For cash planning, model the churn curve explicitly: in the worked example, per-customer payback is 3.60 months, but full cohort cash recovery — accounting for the customers who churn before finishing their repayment — runs closer to 4.9 months. Both numbers belong in the model, the per-customer figure for benchmarking and the cohort figure for cash.

The instrumentation failures that kill these projects are all upstream of any analysis. A cloud bill that isn't tagged free-tier versus paying-tier makes the free-infrastructure split a guess and collapses the whole method — fix tagging at the resource, namespace, or tenant level and reconcile to the total bill within a couple of percent. Signup events missing acquisition source make spend back-tracing impossible; capture UTM and referrer at signup and persist them to the user record. A PQL flag that's computed but never logged with a timestamp makes pre-conversion allocation impossible. Untracked human-touch minutes make the self-serve-versus-assisted split impossible. Ambiguous first-payment events that don't record the prior plan state mix trial and freemium motions permanently. Every one of those remediations is engineering work — schema and tagging — not analysis. A RevOps leader who spends the first week of this project debating the formula has misallocated the week.

The gaming patterns worth watching for, each with its tell: moving free-infrastructure cost to R&D shows up as a CAC line that stays suspiciously flat while the free base balloons. Cohorting by conversion month shows up as payback improving right when spend is cut. Reporting revenue payback and calling it CAC payback shows up as a sub-two-month figure on a real freemium product. Excluding assisted conversions as "those were sales-led" shows up as an implausibly low self-serve CAC that won't reconcile against total payments. Each of these is survivable in a board meeting and fatal in a diligence room, which is the whole argument for doing it right the first time.

One edge case worth naming separately: the segment split. Splitting the 600 payers into self-serve and assisted is the most actionable cut in the entire analysis. In the worked cohort, 430 converted with no human touch at roughly $171 CAC and $61 monthly gross-margin dollars, for a 2.80-month payback — genuinely efficient PLG. The 170 assisted conversions cost about $460 each but generated $112 in monthly gross margin, for a 4.11-month payback. That pattern is exactly what you want: the human touch costs more *and* lands on accounts that can carry it. If the assisted segment showed *lower* ARPU than self-serve, the touch would be destroying value and you'd pull it immediately. The blended 3.60-month figure is the reportable number, but the split is what tells you what to do on Monday.

A practical rollout plan across 90 days

The sequencing matters more than the sophistication. Here is a realistic path from "our CAC is near zero" to a model that survives external scrutiny.

Days 1–30 — stop the bleeding. Tag the cloud bill to separate free-tier from paying-tier infrastructure. This is the single highest-leverage action in the entire program and it is pure engineering hygiene, not analytics. Done when the two numbers reconcile to the total bill within about 2%. In parallel, calculate one fully-loaded CAC for your most recent fully-converted cohort using rough estimates — you are after the *shape* of the number, not precision. Switch every internal report to a gross-margin denominator; this alone corrects the largest single distortion and costs nothing but a formula change. And stop matching same-month spend to same-month conversions; a trailing-three-month window is a crude but adequate interim measure while you build something better.

Days 31–60 — build the cohort engine. Stand up the seven-event log: signup with acquisition source, activation with the specific core action, usage events granular enough to compute per-user infrastructure cost and fire the PQL threshold, the PQL flag event with its score, human-touch events with owner and minutes, first-payment with prior plan state and MRR, and plan-change plus churn events for the survival curve. Trigger your first proper signup cohort and trace it forward to conversion. Produce the self-serve versus assisted split and actually make the decision it implies: is the human touch paying for itself? Add free-infrastructure cost divided by payers produced as a standing monthly metric — that ratio is your free-tier generosity gauge, and watching it drift is how you catch the problem two quarters before it hurts.

Days 61–90 — make it board-grade. Assemble the unit-economics page showing payback alongside NRR, burn multiple, and gross margin on one slide; never let payback travel alone. Run the allocation sensitivity at 100%, 70%, and 50% so the range is visible rather than hidden. Write the one-paragraph narrative that says which segment is efficient, which is subsidized, and what you are changing about it. Then document the method thoroughly enough that a diligence reader could rebuild your number without you in the room — a reproducible method is a fundraising asset in its own right.

Cohort granularity depends on volume. Weekly cohorts give fast feedback on funnel experiments but are noisy and produce small payer counts. Monthly is the standard and works for most freemium businesses. Quarterly is stable and board-friendly but hides intra-quarter swings. A useful floor: a triggered paying sub-cohort needs roughly 100 or more converters before payback is decision-grade. Below that, report it with an explicit confidence caveat and lean on the trailing trend instead.

If you have no data team, this is still achievable. Product analytics covers signup, activation, usage, and PQL events. Your billing platform exports first-payment, plan-change, and churn. Your cloud provider's cost console supports allocation tags for the free-versus-paid split. A monthly spreadsheet joins the three on user ID and cohort month. It is inelegant and it will not scale past a few thousand conversions a month, but it is *correct* — and a correct rough model beats a precise wrong one every time. Graduate to a warehouse pipeline when the spreadsheet join takes longer than the insight is worth, not before.

Assign it cross-functionally on day one or it stalls. Cloud-bill tagging is engineering's deliverable. The gross-margin denominator is finance's. Only the cohort logic is genuinely a RevOps deliverable. Freemium economics fails most often when it gets labeled "the RevOps project" and the tagging work never gets prioritized against a product roadmap.

What changes after day 90. The plan builds the model once; keeping it honest is a permanent operating habit. Three things become routine: every monthly operating review reads the newest mature cohort plus the cost-of-carry trend; every board meeting shows the four-number grid rather than a single figure; and every material free-tier change — a feature gated, a quota raised, a seat cap moved — triggers a re-forecast of free-infrastructure cost-of-carry *before* it ships rather than after. One warning on the transition: when you switch from calendar-month matching to trigger-and-trace, your reported payback will jump. Brief the board before that happens. An unexplained metric jump in a board deck destroys more credibility than the underlying number ever could — say plainly that the business didn't change, the measurement stopped lying.

Related questions

Does this method change for a reverse trial instead of classic freemium?

The formula is identical, but the inputs shift. Reverse trials compress conversion lag to weeks rather than months, so time-shifting matters less, and free-infrastructure cost is concentrated in a short premium-access window rather than spread across a permanent free base. Conversion rates run substantially higher, typically 8–25%.

How does usage-based pricing affect freemium CAC payback?

Usage-based pricing makes the denominator move. Initial ARPU is often low and grows as consumption ramps, so a static monthly gross-margin figure understates recovery. Model the expansion curve explicitly and compute payback against a ramped gross-margin schedule rather than a flat first-month number.

Should expansion revenue count toward CAC payback?

No — keep payback strictly about recovering acquisition cost from the initially purchased subscription. Expansion belongs in NRR and cohort LTV, which are separate metrics. Blending expansion into payback makes a slow-recovering cohort look fast and hides whether the initial conversion economics actually work.

What if we can't separate free-tier from paying-tier cloud cost?

Then you cannot calculate this honestly, and the fix is engineering, not analysis. As an interim estimate, allocate infrastructure by a usage proxy — total events, storage bytes, or inference tokens per user class — and label the result explicitly as an estimate until proper resource tagging lands.

How often should we recompute freemium payback?

Monthly for the operating review, using the newest cohort that has reached its 180-day maturity mark. Quarterly for the board. Recomputing more often than cohorts mature just adds noise, since a cohort under six months old hasn't finished converting and its payback figure is provisional.

FAQ

Why use gross-margin dollars instead of revenue in the denominator?

Because payback answers a cash question: how many months until this customer returns the money you spent acquiring them? You never see the full revenue dollar — a portion goes straight back out to hosting, payment processing, and support for that account. At a 78% gross margin, $90 of monthly revenue returns $70.20 toward CAC recovery. Using revenue overstates recovery speed by exactly the inverse of your margin, which for a compressing-margin AI-native product is an increasingly large lie.

Isn't charging all free-tier infrastructure cost to paying customers unfair to them?

It is a modeling convention, not a moral claim about individual customers. The free tier's purpose is producing buyers, so its cost is a cost of producing buyers, and you spread that cost across the buyers it produced. Reasonable people disagree about whether the right allocation is 60% or 100% — brand and network-effect value are real arguments for partial allocation. The one position that isn't defensible is 0%, which is precisely the assumption that generates the "near-zero acquisition cost" claim.

Our payback came out at 2 months. Is that good?

Almost certainly it means something is uncounted. On a genuine freemium product with a permanent free tier, a sub-three-month fully-loaded payback is rare enough that the first response should be auditing the method rather than celebrating. Check three things in order: is free-tier infrastructure cost in the numerator, is the denominator gross-margin dollars rather than revenue, and are you matching spend to the signup window rather than the conversion month. One of those three is usually the culprit.

How do we handle free users who never convert but generate real infrastructure cost?

Their cost stays in the numerator and their count stays out of the denominator. That asymmetry is intentional and is what makes the method honest — you had to run the whole funnel to get the converters, so the whole funnel's cost is a cost of acquisition. Watching free-infrastructure cost divided by payers produced as a standing monthly metric is how you catch a free tier drifting toward over-generosity before it shows up in payback.

Can we exclude sales-assisted conversions to report a clean self-serve number?

Report both, never just one. The self-serve segment is the real PLG efficiency read, the assisted segment tells you whether human touch is earning its cost, and the blend is the honest company-level figure. Excluding assisted conversions to publish a flattering self-serve CAC is a documented gaming pattern with an easy tell: total conversions stop reconciling against total payments.

What conversion window should we freeze on?

Two fixed checkpoints work well: 90 days as the fast-feedback metric for funnel experiments, and 180 days as the board-reporting and payback denominator. Document both, and never move them. Quietly extending the window is one of the easiest ways to manufacture a rising conversion rate from nothing, and it is one of the first things a careful diligence reviewer checks.

Sources

• Bessemer Venture Partners — State of the Cloud
• Bessemer Venture Partners — 10 Laws of Cloud
• OpenView Partners — Product Benchmarks
• David Sacks — The Burn Multiple
• Andreessen Horowitz — 16 Startup Metrics
• SaaS Capital — Research and Benchmarks
• ICONIQ Growth — Insights and Reports
• Stripe — Billing and Payments Documentation
• AWS — Cost Allocation Tags
• Google Cloud — Cloud Billing Cost Management

Related on PULSE

• How do we model blended CAC when PLG and sales-led motions run together?
• What does a 90-day self-serve expansion model look like?
• How does usage-based pricing change CAC and payback math?
• Which free-tier design levers control conversion — seat caps, feature gates, or quotas?
• How do we score product-qualified leads for a PLG funnel?
• What belongs on a board-ready unit-economics dashboard?