Decision: app-build pricing model + freelance delivery strategy
Set in a 2026-06-11 working session pricing the mitchells-and-butlers apps proposal (Toby Carvery + one more complex second brand app), but intended as the reusable reference for pricing ALL future app builds. The pricing Excel is the artifact to keep (to be sent to Elliot).
Current truth
- Build estimate (two apps): ~21 dev-days. Architecture/tooling design 5d (deliberate thinking time — acceptance criteria, tool descriptions, examples; “don’t rush this”), build 6d (second app raised from 5: confirmation email, DB writes, multiple API integrations), local testing 2d, post-hosting functional check 1d, eval-framework alignment 1d, running evals + iterating 3d, submission 1d (+ hosting decision work).
- Dev cost assumption: $80–110/hr (~$880/day top end) → ~$20k (~£15k) for 21 days. Price for the two apps: £32,500–35,000 (~100% markup + buffer for uncertainty/hosting/API access; even a 50% time overrun still leaves ~50% margin).
- Additional brand rollouts:
8 days ($8k/£6k) → priced £6,500 per rollout, strictly conditioned: same structure/booking flow/testing framework, NO additional development — any new API or requirement is billed hourly on top. Add a handoff-documentation day (our cost, not charged) so any developer can pick up future rollouts; consider a committing-client discount for multi-app deals. - Claude connector + performance testing deliberately NOT in the pricing table — positioned as “other services” with favorable terms for committed build partners.
- Delivery model: top-end freelancer over cheap. Critical criterion: founders must never reach “I’d be better off doing this myself” — pay whatever avoids that, accept lower margin. Long-term goal: ONE person who owns all app builds + MCP optimization work, grown via consistent work pipeline; every engagement is a chance to find them. “Fractal” (trusted) is the fallback if nobody is found in time. No full-time hire yet — pipeline consistency needed first.
- Status caveat: the very next day (06-12 M&B wrap-up) revealed M&B’s build is already underway with their tech partner Guestwise — so the two-app price was likely never pitched as a build; model stands as the pricing reference and the freelancer strategy stands company-wide.
Open questions
- Who is “Fractal” (trusted outsourcing fallback) — agency? past collaborators? Confirm name/spelling.
- Hosting line garbled: “Not going to be out pick… maybe it will be elic” — both garbles plausibly = Alpic (MCP hosting). Was the intent “host on Alpic” or “client’s environment”? (Granola summary says likely client’s environment.)
- “Violate tech” (garbled) — past freelance client Elliot used as a day-rate benchmark (> £1k/day) — what company?
- Who was the “$50/hr annoying guy” prior freelancer quote referenced?
- Transcript was logged as Elliot solo (no calendar event) but is clearly TWO voices; the second voice built the pricing Excel and was told “you don’t need to join” the M&B call — almost certainly Vincent. Confirm.
Timeline
- [2026-06-11] (working session — Elliot + second voice, presumed Vincent) Full pricing model built in Excel: 21-day two-app estimate, $80–110/hr dev assumption, £32.5–35k price, £6.5k per strict-copy rollout, hourly for anything beyond. Freelancer strategy agreed (top-end rate, “never think I’d do it myself”, one long-term owner of app builds + MCP optimization, Fractal fallback, no FT hire yet). Excel to be sent to Elliot as future pricing reference. Next-day M&B call (Ewa + Richard) confirmed Elliot-only.