This is a stop-loss postmortem, not a ClientHunter product tour, and not a success sequel about what came later.

Take one chain home:

Chart looks good → celebrate signups, ignore the signup-to-pay cliff → feature interviews replace “why pay now” → pause the feature list; run purchase-trigger interviews; if intent stays empty, stop the current ICP

On Indie Hackers, Jack Builds recounted a B2B AI lead-gen tool: after posting on Twitter, Hacker News, and Slack communities, he hit ~500 signups in about 30 days—the charts looked exciting. Closer look: about 3 paid, ~0.6% conversion (the title said “0 paying”; the body uses three). He later admitted chasing vanity metrics, treating free signups as warm leads, waiting until week 8 to talk seriously with users, and aiming at “lead gen for B2B companies”—an ocean, not a niche.

He later rebuilt as clienthunter.ai, conversations first, features second. The site confirms current positioning. This piece only covers the signup surge / near-zero pay stage—not rewriting later marketing as “validation already happened.”

~0.6% is not a universal B2B SaaS failure line. What’s worth reusing is judging, while the curve still rises, whether “more acquisition / more features” buys validation—or only delays facing the cliff.

1. The ledger: what held—and what didn’t

From the Indie Hackers write-up (name/positioning cross-checkable with the site):

Item	Public claim
Stage product	B2B AI lead generation tool
Acquisition	Twitter, Hacker News, Slack, etc.
~30-day signups	~500
Paid	3 (0.6%)
User talks	Serious talks around week 8
Positioning flaw	“Lead gen for B2B companies”—author: too wide
After	Rebuilt as clienthunter.ai: conversations before features

Signal	What it proves	What it doesn’t
500 signups / rising curve	Distribution and hook work	Willingness to pay
Eager free trials	Many “I’ll try when I have time”	Warm sales leads
Ongoing feature shipping	The team can ship	Purchase triggers are solved
3 payments	A few people did pay	Current ICP and packaging work

Stop-loss starts at “are signups treated as progress,” not at “did anyone show up.”

Versus BrandingStudio: PH spike + extreme activation; here a community cold-start curve looks good. Same family—volume without pay—with extra weight on asking “will they card?” last.

2. Failure chain: three early signals postponed

1. Signups treated as validation; the pay cliff treated as noise

The author is blunt: each new user felt like a win; in B2B a free signup is often “I’ll try later,” not a warm lead.

Signal one: does signup-to-pay (or strong paid intent) keep up with the curve inside an observation window?

When the curve rises and pay stays near zero, optimizing for more signups amplifies an unproven front of the funnel.

2. The cut is an ocean, not an acceptable ICP

“Lead gen for B2B companies” is not a niche. Without who / under what trigger / for which outcome they pay, feature interviews scatter: everyone wants something different; you think you’re validating demand while collecting a wish list.

Signal two: is the named paying object getting narrower and more concrete—or more “for everyone”?

3. Week-8 talks; easy to ask for features, not card triggers

His biggest lesson: talk to ten users before shipping; the real question is “what would make you pull out your credit card right now,” not “what features do you want” or “would you use this.”

Signal three: are purchase-trigger interviews happening, or only feature talks and growth moves?

Feature interviews can always add one more item; purchase-trigger interviews surface faster that nobody will pay—or that payers aren’t in your current traffic.

The real issue is often not “not enough features,” but:

You haven’t started testing the purchase decision—yet you’re already using the growth chart as proof you’re right.

3. What to copy / what not to

Copy

• Split “signups / curve” from “pay / paid intent” on the weekly dashboard.
• Ask “why card now” before “what feature next.”
• Narrowing the cut is validation, not stinginess.
• Empty paid intent across an observation window → change ICP assumptions before adding features.

Don’t

• Don’t turn ~0.6% into the industry failure line for B2B lead-gen tools.
• Don’t ignore the body’s “3 paid” because the title said “0”—near-zero and absolute zero share the gate; use the body figures.
• Don’t rewrite the early cliff away with later site/marketing numbers.
• Don’t treat “talk to ten users” as ritual; interviews without purchase-trigger questions can still be wish collection.

4. When to trigger stop-loss

Low conversion alone isn’t a shutdown order. The danger is the chart still rising, so next week’s budget keeps going to acquisition and features.

Signal	What you see	First move
A. Long signup-to-pay cliff	Signups at scale; pay near zero	Stop celebrating signups; weekly focus on pay and intent
B. Persistent no paid intent inside the window	Talks/trials can’t name “why pay now”	Stop expanding current ICP; narrow or change buyer hypothesis
C. Feature interviews only	Wish list grows; no card scenarios	Switch to purchase triggers; demote feature asks
D. Cut keeps widening	“One more customer type / one more use case”	Stop being everything; nail one willingness-to-pay sentence
E. Growth crowds out commerce	Time goes to posts, launches, features	Pause one acquisition cycle; fill pay validation first
F. “They’ll pay later” excuses	“Once they know the product they’ll buy”	Cap the window; no intent → change the hypothesis

As a small team’s own rule, pre-set a 2-week observation window: if A holds and either B or C holds, enter a formal decision—pause the next feature, finish a set of purchase-trigger interviews; if talks still show no “pay now” signal, stop the current ICP instead of defaulting to “another acquisition round.” That is not an industry standard that B2B must hit a conversion rate in two weeks.

Versus ThinkAny: there unit economics under large ongoing traffic; here earlier—as soon as the growth story forms, the pay cliff should be the main dashboard. Versus RedChecker: trial never starts; here people already signed up—the gate is willingness to pay. Versus BrandingStudio: spike window vs community-curve window; both show volume doesn’t prove purchase.

5. Two exits—not a denial of later rebuilds

1. Pause the feature list; ask “why pay now” first
His later direction: conversations before features. For this column, the main move inside the window isn’t another release—it’s booking talks about purchase triggers, even when the answers sting.

2. Empty paid intent → stop the current ICP, not “stop founding”
When the cut is an ocean, more acquisition repeats “try, don’t pay.” Narrow to one object with a sayable card reason; if that fails, change the hypothesis—don’t fish the same ocean another thirty days.