The most dangerous stage of a project is sometimes not “nobody uses it,” but “plenty of people use it,” while you still cannot prove it deserves more of your time and cash.

ueCalc, a financial-modeling tool, ran for about nine years. Public figures: ~10,947 registered users, ~90 who ever paid, ~€6,356 total revenue disclosed, and only ~3 active paying customers at the time of writing. Personal investment of about €320,000 is the founder’s published estimate.

It was not a product with zero need: people signed up, bought templates, and some subscribed.

The problem was different: a low-frequency job kept being packaged as a subscription business, and revenue never covered basic fixed costs.

So this is not “why ueCalc failed.” It is a stop-loss question:

When users keep growing, but the current way of charging still cannot support the product, when do you stop proving the same assumption?

This piece only covers that slice: signup growth, free acquisition, low-frequency demand, subscription mismatch, and which stop-loss moves to trigger. It does not teach how to build the product, and it does not predict whether the founder’s later public experiment will work.

The chain to keep:

Signup growth → willingness to pay unproven → free users ≠ target buyers → low-frequency need forced into subscriptions → revenue cannot cover fixed costs for long → stop treating signup count as progress; reprice or shrink burn

~90 / 10,947 is not a universal fail line. What transfers is when this chain trips—the structure low-frequency use × free acquisition × subscription = long mismatch—not “few payers means kill the product.”

1. The ledger: what nine years proved—and what they never did

Daniel Khanin teaches startups unit economics (book Unit Economics, foreword by Ash Maurya) and spent nine years showing this product line did not clear the bar. ueCalc targets founders: unit economics, P&L forecasts, metrics trees—less forty-tab Excel. Numbers are from his Indie Hackers post and blog autopsy.

Read the timeline as separate signals:

Stage	What happened	What it proves	What it does not
2017 launch + trial	~600 signups in first two months, almost no sales	People will try	Trial turns into pay
Trial removed	Signups fall to ~10–15/month; three quiet years	Growth depends on the gate	“Nobody needs this”
2021 free	Hundreds of signups/month for ~15 months; ~5,000 in the free window	Free can fill a funnel	The funnel is full of buyers
Monetize from Nov 2023	One-time templates + subscription; then ~3,133 signups, ~90 payers	Some will pay once for a template	Subscriptions can fund the company

After November 2023 (earlier sales data incomplete, per the founder): ~3% conversion; ~69 template buyers (~2.2%), ~21 subscribers (~0.67%). ~€6,356 over ~29 months ≈ €219/month; minimum fixed cost to keep the company alive ≈ €500/month—before paying himself full-time.

On subscriptions: average LTV ≈ €70; median subscription life ≈ 2 months, then churn.

Stop-loss starts here: not “has anyone heard of you,” but “can cash under the current sell cover fixed cost.” Ten thousand signups only mean the first link can run long and loud.

2. Failure chain: four links stacked

1. Signups as progress

In the free period he told people “we have 5,000 users”—knowing it was a vanity metric, yet still believing pay would come once he figured out how to charge.

For a small team the danger is simple: while the signup curve rises, it is hard to put “payers, margin, can we cover €500?” at the top of the weekly agenda. Rising users delay the stop; they do not prove the business.

2. Free users ≠ target buyers

His write-up: people came because it was free, not because they needed financial modeling. When price appeared, real need was missing. Free built habit, not intent.

Same pattern as “grow free, monetize later,” but longer—years of free. While willingness to pay is unproven, more features and more free acquisition only postpone the test.

3. Forcing subscriptions on low-frequency need

As a consultant he already knew: unit economics is not a daily task. Founders need it for a first model, a pitch, a strategy reset, a business plan—roughly 3–5 times a year.

He still sold classic SaaS monthly (public figures around €40–50/month). A tool used only a few times a year rarely sustains long monthly subscriptions. The split matches: one-time buyers clearly outnumbered subscribers.

The stop is not “ship a daily habit feature.” It is: when the use window is short and you still narrate subscription, change the charge shape before you add features.

4. Revenue cannot cover fixed costs

Marginal cost near zero can make unit economics look positive; contribution still failed to cover fixed cost—€219 vs €500. From January 2025 he left full-time work on ueCalc and returned to consulting / a job.

For indie builders: “one deal pencils out” ≠ “this business feeds you.” The second gate decides whether to double down.

There is also an expectations gap: many wanted a magic button; the tool still needed their data and their execution. People who want the problem solved are not automatically people who will pay for your solution.

3. Structure: low-frequency need vs subscription KPIs

ueCalc’s structural issue: the job happens a few times a year, yet expansion was measured with subscription KPIs.

Consulting work makes “low frequency → don’t force subscription” obvious; the urge to build scalable SaaS pushes the same fact aside. What is missing is usually not another feature—it is admitting:

This is not a daily-active SaaS; subscription KPIs will systematically misread it.

Consultants stuck with Excel; investors complained about bad models but almost never paid for the service—buyer and packaging may also be wrong. This column’s spine stays: signup metrics, free funnels, subscription mismatch, fixed cost.

4. When to trigger stop-loss

Do not double down because one month’s signups look good, and do not wait until “one more year” just because the site is still up. The real danger is rarely one metric alone—it is several signals showing up together for a long stretch:

Signal	What you see	First move
A. Signups up, pay flat	Users dominate the dashboard; pay/revenue rarely do	Change dashboard: payers, revenue, fixed-cost cover
B. Free drives growth	Growth is mostly “free”; paywall kills the curve	Stop scaling free acquisition; prove pay on a small stream
C. Short use window, still selling sub	Few uses/year, heavy MRR story	Prefer buyout / per-use / packaged advice; stop “retention features” as self-deception
D. Revenue < fixed cost for long	Contribution cannot cover minimum burn	Stop full-time doubling down; cut cost, change sell, or shrink time
E. Very short subscription life	Median 1–2 months then leave	Don’t call short subs PMF; rebuild the pricing window
F. Interviews close, self-serve doesn’t	1:1 sells; funnel doesn’t	Drop “product sells itself”; fix sales/landing/activation—or stay manual

5. Stop-loss ≠ shut down: three exits

The founder says he will not shut down yet—he will change the model in public. For readers here, keep three executable options:

1. Stop treating signup count as progress
Weekly: payers, revenue, fixed cost, this week’s pay experiment. Signups may be watched; they must not prove “we’re fine.”

2. Rebuild how you charge
For low-frequency tools prefer buyout / per delivery (templates, model files) / tool + teaching. Leave subscriptions for real weekly/daily workflows. Same axis as the pricing guide: price by how long users need you.

3. Change buyer or packaging
From “all registrants” to “needs an investor model this week”; or from pure software to software + office hours. A consultation link with zero clicks still needs a real test—not just a link.

Later public experiments and lifetime-deal hypotheses do not prove that pushing subscriptions harder back then was right. This postmortem only asks which step should have stopped forcing subscriptions on a low-frequency tool.

6. If you are in a “ueCalc stage”

When signups look great, pay is thin, use is a few times a year, and you sell monthly:

1. How many times a year do users truly need the core outcome?
2. Is subscription the main charge?
3. Can the last 90 days’ revenue cover minimum fixed cost (including your time)?

If the answers are “a few times a year + subscription-led + cannot cover fixed cost,” the next move is not more features. It is: pause subscription as the primary charge, and test only one non-subscription offer (buyout or a single template is enough) to see whether pay appears.

Change one primary variable at a time: pricing window first, then acquisition, then wedge.

When this applies

Best fit: low-frequency, high-cognition tools (finance, filings, quotes, contracts); indie products grown on free signups; teams that treat signup count as funding or ego.

Do not copy-paste onto daily collaboration SaaS, true weekly workflows, or enterprise deals where software is an attachment to a long sales cycle—those need pipeline metrics, not a “90/10,947” consumer conversion reflex.

Again: ~10,947 / 90 / €6,356 are stage figures from one public autopsy, not an industry death line. The chain triggers the stop—especially whether you are ignoring low frequency × free acquisition × subscription.