Metric fraud is real, and prosecuted
HeadSpin’s founder went to prison for overstating ARR ~8×. A company marked at $1.16B fire-sold for $28M once the fraud surfaced. Numbers in a deck are assertions until someone re-derives them.
Independent forensic diligence
Trust,
but verify.
We independently verify a target’s revenue, users, web presence, and founders — re-derived from raw signals, not the numbers in the data room.
Cross-validation
live
Illustrative. Real engagements re-derive every figure from source data.
Signals in:
- Stripe
- SerpAPI
- Wayback
- CourtListener
- OpenSanctions
- OpenCorporates
The problem
The data room is full of claims, not facts.
Agents are polluting the funnel
Automated browser agents now generate a fast-growing share of "user" traffic. User counts built on bot sessions — and signup lists padded with undeliverable emails — look healthy and mean nothing.
+7,851% agent traffic, YoY1
Mid-tier funds have no forensic step
Large funds run six-week quality-of-earnings reviews. Growth-stage funds move faster — and rarely get an independent, forensic read on whether the metrics are real.
The deliverable
A signed verdict, not a vibe.
-
Signed evidence pack
Every finding ships in an Ed25519-signed pack with a content hash — tamper-evident and independently verifiable.
-
Claimed-vs-verified verdict
A line-by-line reconciliation of every headline metric, plus a single red-flag score you can drop into an IC memo.
-
72-hour turnaround
Forensic depth on a diligence clock. Not a six-week QoE — a fast, independent read before you wire.
Heron · evidence pack
v1
Verdict
Material discrepancy RED FLAG
Red-flag score
78/100
- Revenue8× claimed
- Real users−66%
- Founder screen2 flags
Ed25519 · 9f3a…c1d4
signed · verified bytes
How it works
Five stages from raw signal to signed verdict.
-
01
Connect raw signals
The target installs a one-line snippet and connects Stripe. We pull independent ground-truth they can’t retroactively edit — plus their public web footprint.
-
02
Re-derive revenue
We reconstruct recognized revenue from raw Stripe events — independent of the ARR on the slide.
-
03
Authenticate the user base
We classify every session human vs. agent and test signups for deliverability and email-entropy. Bots and fake accounts never count as users.
-
04
Screen presence & founders
We screen web presence — Reddit, search, backlinks — and run founders against court records and sanctions lists.
-
05
Scan & sign
We check the data against known tampering signatures, then issue a claimed-vs-verified verdict, a red-flag score, and a signed evidence pack.
Presence & people
We map what’s not in the data room.
187 signals crawled2 flags
4 sources checked2 flags
Illustrative. Every node resolves to a cited source in the signed evidence pack.
Why independent
We work for you. Not the founder.
Friendly aggregators
Take the company’s own numbers, format them, and hand them back. The target is the customer — so the metrics are never adversarially tested.
Heron
Re-derives the numbers from raw signals the founder can’t retouch, then tells you exactly where claimed and verified diverge. Adversarial by design.
Before you commit
Know what’s real before you wire.
Forensic verification in 72 hours.