Journey → Onboarding → Initialization
How a GC or sub goes from finding us to hooked. Companion to the funnel map.
Perch · CRO cockpit · v0 · Jun 24 2026 · ~/perch-cro/onboarding
The thesis. Customer journey, onboarding, and initialization are related but distinct. The journey gets them to the door. Onboarding's job is not to collect data — it's to get them using the tool. And per Brock, the last 4–5 onboarding steps are initialization: first real activity, the moment stickiness is created. So sub-agents should pre-fill the boring setup the instant we have an email — to push the customer into use as fast as possible.
1 · Who — the customer categories (unchanged)
Two types only for this exercise: GC (1 category) and Subcontractor (34 trades). Each banded S / M / L by revenue → size decides the journey: S = self-serve, M/L = white-glove. (HO = homeowner is just the 3rd account-type option at verify — not a track we build.) Full bands in the funnel map.
2 · The three layers — related but not the same thing
Customer Journey
Find us → at the door
"How do they get to Perch?"
ch6 acquisition channels
·LP / HP
0Book a demo?
Goal: arrival. Owned by marketing + PLG/SLG. Ends at signup intent.
Onboarding
Through the door → using the tool
The 8-step flow — two halves.
Setup half (1–4) — account exists & configured
1Create org
2Verify · GC/Sub/HO
3Industry / trade
4Plan
Initialization half (5–8) — first activity & stickiness
5Invites
6Onboard · freeform
7Dumped into chat
8Init !!!
Goal: get them using it — not data collection. The setup half should be near-invisible (pre-filled by agents) so they reach the init half fast.
3 · Initialization = the stickiness engine
Steps 5–8 · "using it → sticky"
Brock's point: the back half of onboarding is really initialization — the customer's first real activities in the tool. This is where the habit forms and retention is won or lost. Every choice here answers one question: how do we make them use it from the get-go?
Invite the team (5)Multiplayer = switching cost. Pull in the right roles early.
Load a real project (6)Freeform onboard with their actual bid/job — value on day one, not a sandbox.
Tom in chat (7)The agent does something useful immediately — first "wow."
Init (8)Workspace stood up and live. They've crossed into real use.
Integrate emailWire Perch into where work already happens.
Personalized + relevantName personalized, smarter questions — earned by the research pipeline below.
4 · The role overlay — each role uses Perch differently locked: detail later
Locked in: a company isn't one user — it's a set of roles, and each has a distinct day-to-day inside Perch. Onboarding & initialization must personalize per role (what they see first, what they do first). Pre-con roles are the primary daily users (Perch is a precon product); con roles are secondary / field. The day-to-day per role is the next thing to define — placeholder below.
| Role | Side | In app | Day-to-day in Perch — what they actually do |
|---|
| Business owner | PRE-CON | PRIMARY | TBD — champion/signer view |
| C-suite | PRE-CON | PRIMARY | TBD — economic-buyer / portfolio view |
| Estimator | PRE-CON | PRIMARY | TBD — core daily user: takeoff / bids |
| PM / PE (exec) | PRE-CON | PRIMARY | TBD — project setup, buyout |
| Director (PM mgr) | CON | SECONDARY | TBD — oversight across projects |
| Foreman | CON | SECONDARY | TBD — field lead |
| Laborers | CON | SECONDARY | TBD — field roles / functions |
5 · Sub-agent research — fire on first contact
Trigger: the moment we capture an email or website (as early as step 0–1). Agents research in the background so the setup half is pre-filled and the onboarding can steer itself by company + role.
TRIGGEREmail / website capturedThe earliest signal we have.
→
AGENT 1Scan email → find siteDomain → company web presence.
→
AGENT 2Crawl & understandWhat they build, who they serve, size, trade.
→
ENRICHPre-fill onboardingType (GC/Sub), trade, size→band→journey, customers served.
→
STEERDetect role → personalizeTitle/signature → tailor the flow per role.
Reliability bar: the research must be near 100% — in white-glove the human never re-keys it, and a wrong pre-fill is worse than a blank field. Payoff: less typing → reach real use (initialization) faster → stickier.
6 · Locked vs deferred
✓ Locked in now
- Three layers — journey ≠ onboarding ≠ initialization, but they chain.
- Init = the last 4–5 steps = first activity = where stickiness is made.
- Onboarding's job is use, not data — minimize setup friction.
- Roles each use Perch differently → personalize onboarding/init per role.
- Sub-agents fire on email/website capture to pre-fill & steer.
- HO = homeowner — 3rd account-type option only; exercise stays GC + Sub.
↦ Deferred — "we'll talk later"
- Day-to-day per role — fill the table in §4 (needs the real precon_web service/surface list).
- Personalized onboarding flows — the specific steered paths per company + role.
- Stickiness mechanics — which init levers we actually build first.
- Sub-agent design — how the agents run, what they write back, the 100% bar.