Upwork strategy to-do list¶

Dependency-ordered. Each phase unlocks the next. GitHub issues mirror this list.

Last updated: 2026-05-26

Phase 0: Research foundation (do FIRST, lightweight)¶

0.1 Conversion-centred design reference — ✅ confirmed current 2026-05-26. 244 lines, cited, opinionated, ready to load as Lovable system prompt. No updates needed.
0.2 Writing-great-LP-copy reference — ✅ done 2026-05-26. references/landing-page-copy.md (~260 lines, opinionated, cited).
0.3 Trust signal checklist reference — ✅ done 2026-05-26. references/trust-signals.md (~290 lines, taxonomy + placement + per-vertical + anti-patterns + 10 cited sources).
0.4 Client-research methodology — ✅ done 2026-05-26. references/client-research-method.md (~225 lines, 6-step workflow + time budgets + output format + tools).
0.5 Vertical-specific patterns reference — ✅ done 2026-05-26. references/vertical-patterns.md (~440 lines, 7 verticals × brief patterns + features + trust signals + tone + objections + pricing + red flags + win patterns).

2.1 Build the client-research Claude skill — input: brief text + client URL. Output: research summary (trust signals, USPs, brand voice, CRO issues, competitor LPs, recommended approach) + ready-to-paste Lovable prompt. Cuts pre-bid research from 30 min → 5 min.
2.2 Slack alert filter change — currently 9+ pings Slack. Change to 7+ filtered to landing-page lane only (Hustle identity OR LP keywords from portfolio.md). Volume estimate: 2-3 alerts/day. Code change in notifications/slack.py.
2.3 Reject-keyword scoring — scorer should auto-cap WordPress/Webflow/Shopify/Framer/Wix/Squarespace/CF/Kajabi/GHL/Unbounce/Instapage briefs at 4 (already in portfolio.md, verify scorer picks it up post-restart)
2.4 Country filter audit — confirm scorer respects existing India/Pakistan/Bangladesh/Philippines/Israel/Nigeria cap-at-3 rule for client country
2.5 Auto-draft proposal generator (optional, ~2-3hr code, ~$3-6/day API cost) — triggers on every 7+ LP job, drafts proposal using playbook rules. Slack alert links to draft for one-click review/edit/send. Build only after Phase 1 ships.

3.1 Start bidding 5-10 tailored proposals per week using playbook.md rules
3.2 Bid-timing target: within 12-60 min of post (avoid the 5-10 min dead zone)
3.3 Log every bid in log/applications.md with: time-to-bid, format (TLDR+long vs short), Loom y/n, tier offered, outcome
3.4 Run active tests from tests.md — TLDR+long combo (10 bids min), Tier-2 pre-built spec landing (10 attempts max, kill if <30% win)
3.5 Sunday weekly review — reply rate, interview rate, contract win rate. Update playbook + myths as data lands.
3.6 First-30-days deadline tracker — first contract within 30 days = 3x more likely to build sustainable income. Track in dashboard counter.

4.1 Pitch $79/mo Care retainer to every $200-500 build client at handoff
4.2 Pitch $349/mo Growth retainer to clients who hit a result and want more
4.3 Pitch $999/mo Performance retainer only after a Growth client wants to scale
4.4 Track recurring revenue separately from project revenue

5.1 Build outcome dashboard — bids sent, replies, interviews, contracts, win rate by tier, win rate by vertical, by tactic
5.2 Win/loss analysis automation — for each closed bid, capture: what worked, what didn't, what client said
5.3 Lovable credit budget tracking — alert when monthly spend exceeds plan limit so we know to upgrade or throttle
5.4 Shopify + Replo learning sprint (Phase 2 expansion) — once Next.js system is operational and earning, invest a week in Replo to capture the 11% Shopify demand
5.5 Framer reassessment — revisit if/when Framer's AI-prompt mode matures (currently designer-first, not prompt-native)