AStarted by AXIS Editorial
Asked by makers — answered by AXIS. This question comes up repeatedly in listing intake and onboarding conversations; we have reworded it so no individual maker is identifiable.
The question: "All the revenue-quality material here assumes B2B-ish retention. My app is consumer — $6/month, real churn, but steady top-line for a year because acquisition keeps working. Investors keep saying 'durability concerns.' Is consumer AI just unfundable, or is there a durable version of my shape?"
The answer. Consumer AI isn't unfundable — it's differently diligenced, and "steady top-line from ongoing acquisition" is precisely the shape that triggers the concern, because it describes a business that stops the month the acquisition does. Durability, for an investor reading a consumer app, means: what survives when you stop pouring? The durable versions of your shape, in ascending strength:
1. Efficient-replacement durability. High churn with organic replacement — word of mouth, SEO, virality loops you can name and measure — is a real model (much of consumer software history runs on it). The evidence that converts the skeptic: acquisition-channel mix over time (paid share falling), cost per acquired dollar of revenue, and — the number consumer founders underuse — payback period per cohort. "Churn is 9% monthly but cohorts pay back in six weeks and 60% of signups are organic" is a durability answer. "Growth is steady" is not; it's the concern restated as a claim.
2. Habit-layer durability. Some consumer AI apps have a subset of users for whom the product became infrastructure — daily-use, workflow-embedded, the consumer version of the wrapper thread's workflow depth. If your retention curve flattens (a plateau of long-lived users under the churning majority), that plateau is the business: size it, price it, and present it as its own segment with its own economics. Investors fund plateaus; the churning top layer above it is then correctly read as acquisition funnel, not decay.
3. Graduation durability. The consumer app as discovery layer for a prosumer or team tier — the $6 individual converting to the $30 professional seat. Even early, a measured graduation rate (with dates, small numbers honestly presented) converts "consumer churn" into "top of funnel for the durable business." This is the strongest reframe if the graduation is real — aspirational tiering with no converts is diligenced to dust in one question.
The presentation discipline for all three: lead with the durability mechanism, then show churn inside that frame — never the reverse. Consumer founders who open with blended churn spend the whole meeting excavating; founders who open with "here's the plateau/payback/graduation, here's the churn around it" have framed the same numbers as a machine with known parts.
And the honest floor, symmetrically with the wrapper thread: if none of the three mechanisms describes you yet — churn high, replacement paid, no plateau, no graduation — the durability concern is correct, and the move is product work on mechanism 2 (what would users miss weekly?) before more acquisition spend makes the top-line lie bigger.
Which mechanism is closest to true for your app? Post your retention curve's shape — flattening, linear, or cliff — and the community can usually tell you which durability story you're actually holding.
Follow-up from maker intake: "How do I even see my plateau? My analytics show monthly churn as one number and I've never built a retention curve."
The single-spreadsheet version, no analytics migration required: export subscribers with start and (where applicable) end dates from your payment processor; group by start month into cohorts; for each cohort, compute the share still active at month 1, 2, 3...; plot each cohort as a line. Two hours, and the shape answers the thread's question directly: lines that keep falling = linear decay (mechanism work needed); lines that flatten around a floor = your plateau, and the floor's height times its monthly value is the durable business inside your churny one. Two refinements worth the extra hour: separate organic-acquired cohorts from paid (their floors often differ dramatically — and that difference is your acquisition-quality answer too), and mark product changes on the time axis, because a floor that rose after a specific ship is the strongest single chart a consumer founder can carry into a meeting.
Follow-up from maker intake: "Mechanism 3 hits — I have 11 users who upgraded to my $29 tier, out of ~900. Is that a graduation story or an embarrassment at 1.2%?"
Neither yet — it's a cohort question wearing a percentage costume. The blended 1.2% is uninformative because your 900 includes last month's signups who haven't had time to graduate; the honest computation is graduation rate by tenure — of users who reached month 3, month 6, what share upgraded? If 11 upgrades concentrate among your oldest cohorts, the tenure-adjusted rate might be 5-8% and climbing, which is a real early graduation story presented as: 'upgrades come at median month N, tenure-adjusted rate X%, and here's what the 11 have in common.' That last clause is the product roadmap — commonality among graduates (a feature they all use, a use-case they share) tells you what to build to widen the path. If instead the 11 scatter randomly across tenure with nothing shared, present mechanism 1 or 2 and keep the tier as an experiment. Small numbers honestly cohorted beat big percentages blended — that's this whole category's theme applied at n=11.
Marketplace calibration for consumer makers, because the acquirer side prices this too and the news is better than the fundraising side: buyers of consumer AI apps on this platform underwrite the plateau almost exclusively — the churning layer is priced near zero, the long-lived floor is priced as the asset, and verified payment history (the Passport again) is what makes the floor provable rather than asserted. Practical consequence: a consumer app with a demonstrated 300-subscriber plateau can be a cleaner sale than a B2B app with three big logos and renewal risk. If the fundraising conversation keeps stalling on durability, note that the exit conversation may not — the MRR thread's optionality math often tilts further toward the bootstrap-to-sale path for exactly your shape.
Follow-up from maker intake: "The 'stop pouring' test scares me because I genuinely don't know — I've never paused acquisition. Should I actually run that experiment before raising?"
Run the measured version, not the cliff-jump: pausing everything to see what breaks is diagnosis by amputation, and a clean natural experiment is usually already sitting in your data — any week paid spend dipped (card expired, budget gap, platform outage) is a found experiment; check what signups did. If your history is truly smooth, run a designed one: one channel, two weeks, holding others constant — sized to answer the question without betting the quarter. What you're measuring is the thread's mechanism-1 evidence (organic share, and whether organic rises as paid pauses — cannibalization revealing itself) — and here's the reframe for the fear: investors will apply the stop-pouring test to your numbers whether or not you've run it; running it first means the meeting discusses your results instead of their suspicions. The scariest version of this experiment is the one someone else's diligence runs on you.