Whiteboard digitized — the funnel that should drive who we pick & how we test onboarding.
Perch · CRO cockpit · v0.3 · Jun 24 2026 · roles · step 8 · agent bar · HO=homeowner locked · amber = confirm my read
General Contractor 1 category
Not split by trade — one GC track, banded by revenue.
S< $10M revenueSELF-SERVE
M$10M – $500MWHITE GLOVE
L> $500MWHITE GLOVE
Subcontractor 34 trades
34 trades × bands. Lower revenue ceilings than GCs.
S< $10M revenueSELF-SERVE
M$10M – $100MWHITE GLOVE
L$100M+WHITE GLOVE
Pre-con roles primary users
Office side — where the bid/buyout decision lives.
Biz owner · champion / signerPRIMARY
C-suite · economic buyerPRIMARY
Estimator · core daily user — takeoff/bidsPRIMARY
PM / PE · exec — project setup, buyoutPRIMARY
Con roles secondary / expansion
Field side — delivery & execution.
Director · PM manager — oversightSECONDARY
Foreman · subs / field leadSECONDARY
Laborers · field roles / functionsSECONDARY
Self-serve journey
Entry: Landing Page / Home Page (LP/HP)
⤳ Shortcut: a known lead can skip direct to Init (front steps pre-filled).
0Book a demo? optional fork before signup
1Create org
2Verify account — type branchGCSUBHO
3Industry / or / trade
4Plan tied to size band → pricing
5Invites seed the team — by role (see §2)
6Onboard — freeform their real projects / bids
7Dumped into chat handoff to Tom (agent)
8Init !!! workspace stood up — first real value
White-glove journey
Same 8 steps — different operator
Perch does
all this
for them.
Sub-agent research must be 100%
Perch auto-researches the company to pre-fill steps 1–6. For white-glove it has to be airtight — the human never re-keys it.
⚑ Conflicts between the board and test-clients.html v0
A
Revenue bands are wrong. v0 used GC <$25M / $25–200M / $200M+. Board says GC <$10M / $10–500M / >$500M, Sub <$10M / $10–100M / $100M+. → Re-band all cards to the board.
B
The organizing axis is the JOURNEY, not the price tier. v0 cut by Core/Pro/Enterprise. The board cuts by self-serve vs white-glove — and that's the thing onboarding-testing exercises. Re-label every persona by journey; they need different test scripts (customer-driven vs Perch-driven).
C
We're under-indexed on self-serve. The 8-step flow IS the self-serve journey — yet v0 has 1 self-serve persona (Frank Nolte) and 13 white-glove. To test onboarding we need self-serve (small GC + small subs, <$10M) as first-class personas. Open: small/self-serve example per trade → 34×3, not 34×2?
D
HO = homeowner (resolved). It's only the third account-type option at step-2 verify — not a track we build. This exercise stays GC + Sub. No action.
E
Card fields should feed the flow. The 8 steps consume: type (2), industry/trade (3), plan (4), seats to invite (5), a real sample project (6). v0 cards carry Size/Does/Who/Angle/Source — close, but should explicitly carry trade, plan, seat count, and a sample bid so a persona can actually walk the flow.
F
Each persona needs an arrival channel. A real end-to-end test starts upstream: self-serve → PLG / Search / Direct; white-glove → Referral / SLG / Cold. Add a channel field so test runs start at the top of the funnel.
G
Revenue vs ACV — which gates the journey? Left of board gates by customer revenue; right gates by ACV ($60k/$120k). They can disagree (low-revenue shop, big contract). Need one rule — mirrors v0 flag #1 (revenue vs headcount).
H
NEW — personas need role-based seats, not just a company. Onboarding step 5 (invites) seeds the team by role, and Perch's surfaces are role-based (§2). Each test company should carry a realistic seat list — which Pre-con roles (owner, estimator, PM/PE) sign up first, who they invite — weighted to precon since that's the primary product. Also tests the white-glove sub-agent research (must pre-fill a real company at 100%).