The Velo growth plan, updated to where we actually stand in July 2026 — five disciplines, one north star, and the twelve month target model behind it.
Staging passed end to end Jul 10: plans, metering, freezes, seats, roles. Banner SDK 12.3KB against a 30KB budget. Billing still stubbed.
The v2 banner runs on veloconsent.com itself. Our own product on our own site — displacement starts at home.
GA4 spine live through the first party /m endpoint (GTM-MBS9C4KL), verified Jul 16: scan_start, scan_complete, deep_scan_request, early_access_signup, newsletter_signup, contact_submit — leads carry hashed email for enhanced conversions.
GTM Community Template submitted and in gallery review — ahead of the plan. Site at 37 indexable pages, one AEO post per day, honest scan tool.
D3 closed Jul 21: production runs on the Amplio Cloudflare account. Snippet host and edge api unblocked.
Early access list: 0 signups. The capture machinery exists (deep scan invite, early access form) — what's missing is traffic to it and a nurture arc behind it. No spin: the list is what the next 30 days build.
Role, revenue share, IP. Gates the Stripe flip (D5). Due before real money flows.
Walter's call: ampliodata.io (replacing the Usercentrics we pay for, Jul 24) and Certify (first real client pilot, white glove, live by Aug 7). Cropple and Instalat stay as backups.
veloconsent.com still has no Search Console property — the AEO outcome loop stays blind until it exists.
The order matters — teams that run these out of sequence scale funnels that leak.
Never scale acquisition before the product is a must have. Roughly 40% or more must say they would be very disappointed to lose it.
The experience that turns a visitor into a devotee — then compress the time to reach it.
A single metric that captures value actually delivered, decomposed into an equation of levers.
Weekly growth meeting, ICE scored backlog, several experiments a week, every result logged.
Acquisition, activation, retention, revenue, referral. Language fit and channel fit are tested like product features.
Prerequisite for all five: instrumentation. Ours went live Jul 16 — the GA4 spine on the first party endpoint.
Velo has zero external users today. Until must have status is proven, everything is organic and compounding only. With single digit pilot counts the test is a conversation, not a survey — but we ask the literal question and log the literal answer.
The Cookiebot wedge does not expire in a quarter. A broken first impression with agencies would.
Across live Velo domains: the count of conversions and events that Consent Mode v2 plus the edge recovery layer captured which the previous setup would have dropped. It is the value the positioning promises, it only moves when a customer is live and configured and getting results, and it sums naturally across an agency portfolio. Domain count alone rewards shelfware; MRR lags value by a quarter.
Every factor is a lever with its own experiments. The agency channel multiplies domains per account — 5 to 200 sites per win — which is why it outranks every other channel on impact.
"Your old setup was dropping X% of your conversions; last week Velo recovered N of them." The moment the product stops being a compliance checkbox and becomes a revenue instrument. The free scan is the pre purchase proxy — one continuous narrative arc:
POST /api/scan fetches the homepage server side like a first time visitor — no JS. Reads CMP signatures, tags (GTM, GA4, Meta, TikTok…), a visible Consent Mode v2 default, and tracking cookies in Set-Cookie. Detection dictionaries live in packages/detect, shared with the product scanner.
Under every verdict: "does your banner actually hold tags back?" POST /api/deep-scan loads the site in real headless Chrome (Browser Rendering), clicks nothing, and records every tracker request and cookie that fires before consent. Email required — unless the visitor is already known, then we don't ask twice.
scan_start → scan_complete (with the verdict) → deep_scan_request → early_access_signup, pushed via the veloDL/veloLead helpers, through GTM to the first party /m endpoint into GA4. Leads carry hashed email as user_data for enhanced conversions. CTA branches: green → "Get the full audit"; amber and red → "Fix it with Velo".
1. Nurture arc on captured emails (plan 1.3): four Resend emails, scan verdict → what it costs → how recovery works → launch offer; trigger on deep_scan_request and early_access_signup. 2. Verify veloconsent.com in Resend — until then the branded sender falls back and sandbox delivery is owner only, so nurture cannot ship first. 3. Surface deepscan:* keys in the ops velo view next to lead:* so captures are visible daily. 4. Product side events (signup, banner_live, cmv2_verified, first_report_viewed) still to instrument in the console — the aha funnel is blind after the site handoff until they exist.
The Aug 2025 repricing is a documented migration wave. Our own CMP fingerprinting builds the list; the outreach engine sends it. Like for like swap, flat pricing.
A tool, not a form. The deep scan invite already captures email on every verdict — what remains is routing traffic to /scan and the nurture arc behind it.
One answer shaped post per day, already on autopilot — the compounding channel. Commercial and comparison pages come post launch.
Distribution inside the exact tool the ICP opens daily; every gallery search for "consent" is free placement. Submitted, in review.
Rapid first: portfolio pricing, white label, margin built in. Comparison portals route each agency to Velo vs the exact CMP they already license.
G2 and Capterra profiles prepared now, submitted when public. Launch post for backlinks and entity signal, not the ICP.
NEW — the wordpress.org directory is where Complianz and CookieYes built their whole funnel. Every CMP search inside wp-admin is free, permanent distribution.
NEW — session metered pricing fits app billing natively, and consent apps rank on daily "GDPR" searches. Second store, after WordPress proves the motion.
NEW — publish what our scans already see: which CMPs run where, by country and platform. The industry dataset nobody has; an AEO magnet and backlink engine.
Also scored, in the ledger backlog: LinkedIn engine 6.7 (point the Mon/Wed/Fri routine at Velo) · consent teardowns 6.0 (public behavior only, every claim verified).
Every plan opens with "here is what you were losing" — sets up the recovery aha, makes the weekly report a payoff.
Snippet paste → banner renders → CMv2 verified, with a visible checklist. The PLG floor everything stands on.
Import the incumbent config, side by side preview, cutover checklist. Built only if objection data demands it.
NEW — configure the banner live on the site (colors, layers, copy) and leave with your snippet. The demo card exists; make it an editor.
Value restated in the customer’s own numbers, every week. The retention trigger — the most important email in the product.
"Your site changed — here is what’s new." Already in the entitlements; the job is making the email good.
New law, new enforcement, new default. Velo watches so you don’t have to — cheap content, high perceived vigilance.
NEW — "your consent rate vs sites like yours" inside the weekly report, from our own aggregate once ~30 domains are live.
Every consented pageview on a Free domain is an impression in front of people who deploy CMPs. Paid removes it.
One click "send this report" to a client, a boss, a dev. B2B referral is forwarding a useful artifact.
The agency’s recurring cut makes every agency a commissioned sales force. Structural referral.
NEW — sites that pass get an embeddable "consent verified" badge linking back. The SSL Labs grade play: verified sites advertise us.
Flat multi domain pricing: adding a domain costs nothing at first, then metering grows the account. Domains per account is a first class metric.
Displacement copy and comparison pages anchor on the incumbent’s real renewal math — verified numbers only.
NEW — two months free on annual at the metering step. Cash up front funds the seed budget; churn drops a tier.
Session metered, flat across domains. Stripe flip held on D4.
One weekly growth review — Walter + Leon, 25 minutes, Claude runs the numbers beforehand. Given the scheduled task fleet, execution capacity is not the constraint. Decision speed is.
One row per experiment: hypothesis, ICE, dates, result in numbers, decision. Same discipline as the AEO learnings loop — the house pattern that already works.
home-b is the standing A/B vehicle. Recovery led versus price led versus compliance led heroes, judged on scan starts and captures. The winner writes the outreach subject lines too — language fit is one system across site and email.
GA4 event spine through the first party /m endpoint on the site Worker, verified Jul 16. Scan, lead capture, and signup events flow; the growth panel reads from here.
The dependency spine: G0 unblocks pilots → pilots unblock the must have gate → the gate unblocks wave 1 → wave 1 results unblock scale. Everything else runs in parallel and does not wait.
Instrument ✓ · hosting ✓ · dogfood 1 ✓ · scan capture, dogfood 2, pilots 1–2
Displacement list built (not sent) · directories ready · dry run ≤ 30 min · Stripe flip
Signup opens · directories + template submitted · one note to the warm list
Must have conversations with every pilot at ~30 days live; fix the leakiest step
Wave 1: ~50 Cookiebot refugees. Review against the outreach engine's standing bar
Rapid signed · comparison portals · badge loop measured · expansion experiments
Waves step 50 → 150 → 300/wk if the bar holds; second agency; the MoM curve on the next slide
| Month | Live domains | Signups | Paying | MRR | MoM |
|---|---|---|---|---|---|
| Jul 2026 | 2 | 0 | 0 | €0 | — |
| Aug 2026 · launch Aug 31 | 5 | 10 | 0 | €0 | +150% |
| Sep 2026 | 9 | 35 | 2 | €30 | +80% |
| Oct 2026 · G6 retro | 18 | 70 | 6 | €95 | +100% |
| Nov 2026 | 28 | 110 | 12 | €200 | +56% |
| Dec 2026 | 40 | 155 | 20 | €360 | +43% |
| Jan 2027 | 58 | 210 | 31 | €600 | +45% |
| Feb 2027 | 82 | 275 | 45 | €950 | +41% |
| Mar 2027 | 114 | 355 | 63 | €1,450 | +39% |
| Apr 2027 | 154 | 450 | 86 | €2,100 | +35% |
| May 2027 | 202 | 560 | 112 | €2,950 | +31% |
| Jun 2027 | 260 | 685 | 142 | €4,250 | +29% |
Targets, not actuals — model v2.1, rebuilt from real data at every Monday review from W31 on.
ARPU expands, ~€12 → ~€30. Session metering, the Server tier, and multi domain agency accounts do the expanding — land and expand, not price rises.
The tail is agencies, not luck. Monthly domain adds step 7 → 12 → 24 → 58: waves scale 50 → 500 sends a week only while the reply bar holds; Rapid from October, Adfluencers (14 clients) from November. December tempered.
Everything after launch assumes G4 (Fri Sep 4) passes; the north star baselines from the first pilot reports in W31.
The stage gate is also the budget gate: €0 paid until G4 passes. Until then the whole machine — scan, AEO, outreach, deep scans — runs on ~€50/mo of fixed infra (Workers paid plan incl. Browser Rendering, Resend, domains). Then paid unlocks stepwise — seed doubled for the aggressive curve — and from spring the cap is pegged to revenue so spend can never outrun proof.
monthly paid cap = max( €1,200 seed, 50% × last month MRR )
Seed unlocks stepwise after G4 — Sep €600 → Oct €1,000 → Nov+ €1,200 — each step only if the guards hold. The 50% of MRR term takes over from June.
Blended paid CAC ≤ €150 · payback ≤ 6 months (CAC ≤ 6 × ARPU) · activation ≥ 60% holds. Any guard fails → pause spend, fix the funnel, resume. Pausing is free; a leaky funnel with paid traffic is not.
| Month | Paid cap | MRR target | Rule applied |
|---|---|---|---|
| Jul – Aug 2026 | €0 | €0 | pre G4 — organic only |
| Sep 2026 | €600 | €30 | seed step 1 · after G4 (Sep 4) |
| Oct 2026 | €1,000 | €95 | seed step 2 · Sep guards held |
| Nov 2026 – May 2027 | €1,200 | €200 – 2,950 | full seed |
| Jun 2027 | €1,475 | €4,250 | 50% × May MRR · ~€11.5k paid, full year |
Search intent only, one channel at a time: "cookiebot alternative", "consent mode v2", CMP comparison queries — the buyers already in motion. No broad social. Max 2 concurrent paid experiments, each a ledger row with ICE and a pre registered kill metric before it spends a cent.
How the 260 live domains decompose. Agencies carry the tail; paid stays the smallest slice by design.
The pacing caps from the previous slide, distributed. One channel proves itself before the next opens.
Shares recalibrate every Monday from the install source tag · agencies + displacement carry ~64% of the curve — which is why D4 and the Rapid signature outrank any ad decision.
The fleet already runs half this plan unattended. Six builds put the rest on rails. Machines prepare — the Monday review decides. Nothing sends, spends, or migrates past a gate on autopilot.
AEO post daily (velo-aeo-daily) · outreach engine with claim verification and cadence caps built in · scan funnel fully evented · deploy cycle + ops dashboard refresh.
Pulls the GA4 funnel, LEADS KV counts, ledger state, and spend vs cap every morning; renders the growth panel on ops and auto prepares the Monday pack.
Resend sequences fired by deep_scan_request and early_access_signup — the four email arc from verdict to launch offer, then evergreen post launch.
Velo segment inside the existing outreach engine; wave size steps 50 → 150 → 300 → 500 a week automatically while the G5 reply bar holds, halts when it breaks.
The weekly email per live domain — consent rate, recovered signals, € range — generated from console data. Hand built for the pilots first; the format is proven before it is automated.
Daily spend vs cap with CAC guard alerts into the Monday pack; deepscan and lead keys auto staged into crm.veloconsent.com so no capture ever sits unseen.
≥ 20 scan runs a week · ≥ 35% capture rate on scan reports · dogfood 2 live · Certify install scheduled · first recovery report sent.
Warm list ≥ 200 · 4 live domains · ≥ 40 AEO posts with first citations logged · dry run ≤ 30 minutes.
≥ 20 signups since launch · activation ≥ 60% · median time to live ≤ 30 minutes · wave 1 out · G5 review Sep 21.
≥ 18 live domains · ≥ 6 paying accounts · backlog rescored · plan v2 written from the ledger, not from hope.
Every week, regardless of phase: recovered signals per week — up and to the right from the first report on. Wave reply bars are the outreach engine's standing thresholds, inherited.
D4 unsettled while revenue starts. Hard ordering: no Stripe flip before the Leon arrangement. Worst case, launch runs stub billing and manual invoices.
Pilot slips → gate slips → wave slips. Two dogfoods derisk the install path first. The wedge does not expire — hold the gate, never skip it.
Launch week swallows delivery time. Launch actions are Claude and Leon heavy by design; Walter's launch day is one note to the warm list plus monitoring.
Deliverability and claim holds on wave 1. The standard engine path has verification built in; claims are pre verified when copy is drafted, not at send time.
Verified claims only. No statement that a competitor or prospect lacks something without a resolved container and a rendered screenshot. Comparison pages meet the same bar as outreach.
Only the proven figures — 34% hidden sessions, 20–40% recoverable, +37% ROAS, 72h setup — always ranged, always qualified. Never an invented stat, customer, or testimonial.
Scan honesty is the moat. "Nothing detected" stays inconclusive under any CTA pressure. The scan converts because it is trustworthy.
Migration order — revised Jul 22. Certify pilots pre launch by Walter's call, with a white glove bar: zero downtime, Carme informed, rollback ready. Odisei still migrates only after the gate.
This deck restates knowledge/velo/growth-strategy.md and growth-plan.md as of 2026-07-22. The plan is rebuilt at the 90 day retro; the model is rebuilt every Monday. Growth is decided weekly — not in decks.
Verify veloconsent.com in Resend (nurture can't ship before it) and surface deepscan:* leads in the ops velo view — the capture exists, make it visible and followable.
Replace the Usercentrics we pay for. One fee gone, before and after consent rates into the ledger.
White glove install, Carme informed, baseline captured; first recovery report the week after. Backup if it slips: Cropple.
Four emails, scan verdict to launch offer. Humanized, every claim qualified.