Niche portfolio - the 7 spec pages¶
The single highest-leverage prep work. 50% of LP briefs explicitly ask for portfolio examples - without one in their vertical, we can't bid honestly.
Companion docs: packages.md (the tier specs that define each example's scope) | playbook.md (positioning rules) | lp-deep-mine.md (the demand data)
Last updated: 2026-05-26
The 7 spec pages¶
7 pages across 6 verticals × 3 tiers. First 4 cover ~64% of LP brief demand.
| # | Vertical | Tier shown | Fake brand | % of LP demand | Why |
|---|---|---|---|---|---|
| 1 | Local service | Tier 1 ($499) | (Kiwi - already live, our anchor) | 15% | Already built with real numbers (3.95x lift); zero cost to use as Tier 1 example |
| 2 | B2B SaaS pricing | Tier 2 ($1,499) | "Stridepay" - invoicing tool | 23% | Largest single LP vertical on Upwork |
| 3 | E-commerce DTC single-product | Tier 2 ($1,499) | "Saltcastle" - food brand | 15% | Paid-social use case, video hero, CRM-connected |
| 4 | Healthcare / wellness | Tier 2 ($1,499) | "Calmkit" - mental health app for SMEs | 11% | High demand, zero competition from us today |
| 5 | Course / coaching | Tier 1 ($499) | "Ship & Tell" - founder writing cohort | 5% | Creator economy; simple form + sales copy |
| 6 | Professional services lead-gen | Tier 1 ($499) | "Northfield Conveyancing" - UK solicitor | 5% | Common low-end brief; appointment-booking pattern |
| 7 | PPC system flagship | Tier 3 ($2,999) | "FlowDesk" - SaaS with quiz + 3 variants | n/a | Tier 3 needs a flagship that demonstrates the system |
Build order¶
1 (already done) → 2 (SaaS) → 3 (DTC) → 4 (healthcare) → 5 / 6 → 7
First 4 builds cover 64% of demand (49% from #2-4 + 15% Kiwi). Stop and start bidding after #4 if pressed for time; revisit 5-7 once first contracts land.
Verticals we are NOT building portfolio for¶
- Real estate (3% demand, niche and low-ROI for us)
- App download / mobile (1% demand)
- Crypto / Web3 (1% demand, hard-skip per criteria)
- Agency / marketing landing pages (1% demand, low priority)
- Creator personal brand (<1% demand)
Stack¶
All 7 spec pages built on Next.js + Vercel with Lovable as the build accelerator.
Rationale (corrected from earlier - see myths.md):
- 3x brief volume vs WordPress in our DB
- One stack = 30-min clone+customise (vs WP theme/plugin chaos that prevents templated builds)
- 48hr delivery promise realistic on Vercel, hard on WP
- 90+ Lighthouse out of the box
NOT because of a rate premium for stack - at landing-page level both stacks max around $60/hr median. The pivot reason is operational (speed, templatability), not rate.
WordPress capability stays in back pocket (Stormline). Don't pitch it.
Build standard for each spec page¶
Per page, ship: - Live URL (subdomain on Hustle or fresh Vercel project) - Mobile + desktop screenshots, watermarked - Lighthouse score screenshot (target 90+ mobile, all CWV green) - 30-second Loom of "the thinking behind this page" (banked for use in future bids) - 100-word case-study writeup (problem / approach / outcome metric, even if hypothetical) - Watermark removable in source (so if client wants to license, you've got the un-watermarked version)
Each page also exemplifies its tier: - Tier 1 examples (Kiwi, Ship & Tell, Northfield): 5 sections, single form, YouTube embed if relevant. Demonstrate Tier 1's 48hr feasibility. - Tier 2 examples (Stridepay, Saltcastle, Calmkit): hero video, full conversion stack, CRM connection, Slack notification, lead-magnet download. Demonstrate Tier 2's full integration. - Tier 3 example (FlowDesk): 3 variants, quiz, multi-channel routing, attribution dashboard. Demonstrate Tier 3's system thinking.
Deliberately NOT shipped: GitHub repo links. Giving the source lets clients clone and ship without paying. Watermarked screenshots + live URL only.
Vertical-specific notes for build¶
- Stridepay (B2B SaaS): pricing table is the centrepiece (3-tier comparison), trust = customer logos + integration logos, copy = benefits over features.
- Saltcastle (DTC food): hero video background, single-product focus, scarcity / "ships in 24h", payment trust badges, mobile-first.
- Calmkit (healthcare): therapist-tone, calming colour palette, trust = HIPAA mention + clinician advisory board (mocked), gentle CTAs, no aggressive scarcity.
- Ship & Tell (course): founder personal-brand voice, cohort dates + countdown, testimonials from "alumni", price tier comparison (Early bird / Standard / Premium).
- Northfield (legal services): trust-heavy (regulatory badges, partner photos), appointment-booking CTA, FAQ around legal process, no jargon.
- FlowDesk (Tier 3 flagship): the quiz routes to one of 3 page variants based on answer; each variant has different headline + CTA. Demonstrates personalisation at scale.
Prerequisites before building¶
The portfolio is a means to the end of better bids. Don't build until:
1. Upwork headline updated (Phase 1.1 in todo.md)
2. Reference docs finalised (landing-page-conversion-design.md, landing-page-copy.md, trust-signals.md, client-research-method.md)
3. Lovable templates set up for each vertical (so each new page = clone + customise, not from scratch)
Time estimate¶
- Tier 1 page from template: 60-90 min each (3 pages = 3-4 hours)
- Tier 2 page from template: 2-3 hours each (3 pages = 6-9 hours)
- Tier 3 flagship: 4-6 hours
- Total: ~15-20 hours focused work for all 7 (assuming templates exist)
- Without templates: 2-3 full days
Infinite payback (used on every future bid).