Michael
966af8ef94
Tools shipped this batch (4 → 6 of 9 Ready):
04 Missing Value Handler src/core/missing.py + cli_missing.py + GUI
05 Column Mapper src/core/column_mapper.py + cli_column_map.py + GUI
09 Pipeline Runner src/core/pipeline.py + cli_pipeline.py + GUI
with soft tool-dependency graph (recommended,
not enforced) and JSON save/load for repeatable
weekly cleanups.
Format Standardizer reworked for 1 GB international files:
• Vectorised dispatch + LRU cache over phone/date/currency/boolean/email
• Per-row country / address columns drive parsing
• Audit cap (default 10 k rows, ~50 MB RAM)
• standardize_file(): chunked streaming entry point (~165 k rows/sec)
• currency_decimal="auto" for EU comma-decimal locales
• R$ / kr / zł multi-char currency prefixes
• cli_format.py with auto-stream above 100 MB inputs
Encoding detection arbiter + language-aware probe:
Closes the last 4 xfails (cp1250 / mac_iceland / shift_jis_2004 / lying-BOM)
via tied-confidence arbiter + Cyrillic / EE-Latin coverage probes.
Distribution-readiness assets:
• streamlit_app.py — Streamlit Community Cloud entry shim
• src/gui/app_demo.py — single-page demo, ?p=<persona> routing,
100-row cap + watermark, free-vs-paid boundary enforced at surface
• samples/demo/ — 3 niche datasets + pre-tuned pipeline JSONs
• landing/ — 4 static HTML pages (apex chooser + 3 niche),
shared CSS, deploy.py URL-substitution script,
auto-generated robots.txt + sitemap.xml + 404.html + favicon
• docs/PLAN.md, DEMO-PLAN.md, DEPLOYMENT.md, POST-LAUNCH.md, NEXT-STEPS.md
— full strategy + measurement + deployment + master checklist
Test counts:
before: 1,520 passed · 4 skipped · 17 xfailed
after: 1,729 passed · 0 skipped · 0 xfailed
Tier-1 corpora added:
• missing-corpus 3 use cases + 16 edge cases
• column-mapper-corpus 3 use cases + 5 edge cases
• format-cleaner intl 20-row 13-country stress fixture
Engine hardening flushed out by the corpora:
• interpolate guards against object-dtype columns
• mean/median skip all-NaN columns (silences numpy warning)
• fillna runs under future.no_silent_downcasting (silences pandas warning)
• mojibake test no longer skips when ftfy installed (monkeypatch path)
• drop-row threshold semantics: strict-greater (consistent across rows / cols)
• currency_decimal validator allow-set updated for "auto"
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
159 lines
8.7 KiB
Markdown
159 lines
8.7 KiB
Markdown
# Post-launch — 90-day measurement plan
|
||
|
||
> Creator-only. The other half of `PLAN.md`: PLAN tells you what to
|
||
> build, this tells you what to measure once it's live and which
|
||
> numbers trigger which actions.
|
||
> **Version**: 1.0 · **Adopted**: 2026-05-01 · **Owner**: Michael
|
||
|
||
This is a runnable monthly checklist, not analytics theatre. Every
|
||
metric below has a **threshold** and an **action**. If you're not
|
||
willing to execute the action when the threshold trips, drop the
|
||
metric — measuring without responding is busywork.
|
||
|
||
## 1. The five numbers that matter
|
||
|
||
Every other dashboard, chart, or vanity stat is downstream of these
|
||
five. The funnel is short on purpose; pre-PMF traffic doesn't have
|
||
the resolution to support more.
|
||
|
||
| # | Metric | How to compute | Threshold | When tripped |
|
||
|---|--------|----------------|-----------|--------------|
|
||
| 1 | **Persona engagement** | `demo.run_completed / demo.page_view` per persona | < 30 % for 4 consecutive weeks | Demo isn't running or BEFORE preview isn't compelling. **Action:** check iframe loads; widen BEFORE preview to show pollution clearly; move demo above the fold. |
|
||
| 2 | **Demo→CTA intent** | `demo.cta_clicked / demo.run_completed` per persona | < 5 % for 4 consecutive weeks | Demo is impressive but the CTA isn't earning trust. **Action:** add network-tab privacy screenshot; soften the price callout; A-B test eyebrow copy on the CTA card. |
|
||
| 3 | **Purchase rate** | `gumroad.purchase / demo.cta_clicked` per persona | < 30 % for 4 consecutive weeks | Visitors click through but don't pull the card out. **Action:** check Gumroad listing renders cleanly; verify refund-policy copy; check that the screenshot on the listing matches the demo they just ran. |
|
||
| 4 | **Refund rate** | `gumroad.refunds / gumroad.purchase` rolling 30 days | > 5 % | Buyer expectation mismatch. **Action:** read every refund email; determine if it's a feature gap (build it), a positioning lie (rewrite), or a personal-fit miss (fine, ignore). |
|
||
| 5 | **Support load** | email tickets / 100 sales rolling 30 days | > 10 | The product isn't self-serve enough at this price. **Action:** find the top 3 questions; add to in-app onboarding + landing-page FAQ + the persona's saved pipeline. |
|
||
|
||
These five also map to BUSINESS.md §12 — that doc names the metrics;
|
||
this doc operationalises them.
|
||
|
||
## 2. Monthly review — 30-minute checklist
|
||
|
||
Block 30 minutes on the first Monday of every month for the first six
|
||
months. After month 6 if numbers are stable, drop to 15 minutes
|
||
quarterly.
|
||
|
||
```
|
||
[ ] Pull last 30 days of demo events from Cloudflare Web Analytics
|
||
[ ] Pull last 30 days of Gumroad sales + refunds export
|
||
[ ] Compute the five numbers in §1 per persona
|
||
[ ] Note which thresholds are tripped (if any)
|
||
[ ] Read every refund email since last review
|
||
[ ] Read every support email since last review
|
||
[ ] Decide ONE thing to change this month (only one)
|
||
[ ] Update CHANGELOG with what was changed and why
|
||
[ ] Schedule next review
|
||
```
|
||
|
||
The "decide ONE thing" rule is load-bearing. Pre-PMF traffic doesn't
|
||
have the volume to A/B-test multiple changes in parallel — you'll just
|
||
confuse yourself about what moved the number.
|
||
|
||
## 3. Per-persona scoreboard (template)
|
||
|
||
Maintain in a single text file or spreadsheet. The shape that fits in
|
||
a notebook page is the shape you'll actually update.
|
||
|
||
```
|
||
Month: 2026-06
|
||
─────────────────────────────────────────────────────────────────
|
||
Shopify Bookkeeper RevOps Total
|
||
Page views 420 180 290 890
|
||
Demo runs 137 59 82 278
|
||
CTA clicks 9 7 6 22
|
||
Purchases 3 2 2 7
|
||
|
||
Metric 1 (engage) 33% 33% 28% 31%
|
||
Metric 2 (intent) 7% 12% 7% 8%
|
||
Metric 3 (purchase) 33% 29% 33% 32%
|
||
Metric 4 (refund) 0% 0% 0% 0%
|
||
Metric 5 (support) 3 tickets / 100 sales
|
||
|
||
Tripped thresholds: RevOps engagement (28% < 30%)
|
||
|
||
This-month change: Move demo embed above the fold on revops
|
||
page; reduce hero text by 40%.
|
||
|
||
Last-month change: Added network-tab screenshot to all 3
|
||
pages. Result: intent +1.5 percentage
|
||
points on Shopify, flat elsewhere.
|
||
```
|
||
|
||
## 4. Stage-gate triggers from PLAN.md
|
||
|
||
Reproduced here so the gate criteria sit beside the metrics that
|
||
fire them:
|
||
|
||
| Trigger | From | Action |
|
||
|---|------|--------|
|
||
| **First paying customer** | PLAN §4 | Continue. Plan is working. |
|
||
| **Zero paid in 90 days** | PLAN §4 | Audit the funnel. Don't add features. Run a small (1-week) outbound experiment to 30 niche-community contacts as a control, even though it stretches the no-touch constraint, to determine whether the bottleneck is reach or conversion. |
|
||
| **$5 k/mo MRR** | DECISIONS §8 | Re-evaluate async constraint. Add priority-support tier (PLAN §2.7). |
|
||
| **$10 k/mo MRR** | DECISIONS §8 | Revisit time-budget allocation. Decide on tools 06–08 vs. additional bundles. |
|
||
| **Marketplace shutdown** | PLAN §4 / DECISIONS §8 | Switch landing-page CTA to own-domain Stripe Checkout. Pre-built; one-line edit. |
|
||
| **Streamlit hard direction change** | DECISIONS §8 | Low-probability re-lock. Tk fallback documented. |
|
||
| **Burnout signal** | DECISIONS §8 | Stop. Triage. The constraint matters more than the revenue ramp. |
|
||
|
||
## 5. What we deliberately do NOT measure
|
||
|
||
These look productive but predict nothing pre-PMF. Don't add them.
|
||
|
||
- **Bounce rate** — single-page sites have artificially high bounce. Useless signal.
|
||
- **Time on page** — landing pages are *supposed* to be quick reads. Long time on page often means confusion, not engagement.
|
||
- **Heatmaps / scroll-depth** — no statistical resolution at <500 monthly visitors. Add when you cross 5 k/month.
|
||
- **Email open rates** — under §2.7 priority support is the only email channel; opens aren't a buying signal.
|
||
- **Social mentions** — vanity. The signal that matters is "did they buy" or "did they come back."
|
||
|
||
## 6. What we measure once, then trust
|
||
|
||
Do these once, then let them run for 6+ months without re-measuring:
|
||
|
||
- **Demo correctness** — once per pipeline release, run all 3 demos
|
||
end-to-end via `tests/test_pipeline.py` and check the output looks
|
||
reasonable. The CI pipeline already does this; nothing to add.
|
||
- **Cross-platform install** — once per release, verify the
|
||
PyInstaller bundle launches on Mac / Windows / Linux. After three
|
||
green releases, trust the build pipeline; spot-check on major OS
|
||
updates only.
|
||
- **Privacy claim integrity** — once at launch, capture the network
|
||
tab while running the cleaner and host that screenshot at a stable
|
||
URL. Re-capture only when a new tool or dependency is added.
|
||
|
||
## 7. Per-persona attribution
|
||
|
||
The buy buttons on every landing page carry `?from=<persona>` query
|
||
parameters. Gumroad propagates that into the order metadata. Use it
|
||
to attribute purchases:
|
||
|
||
| persona key | landing page URL | Gumroad query | Source |
|
||
|---|---|---|---|
|
||
| `shopify-pet` | `/shopify-pet/` | `?from=shopify-pet` | Shopify operator |
|
||
| `bookkeeper` | `/bookkeeper/` | `?from=bookkeeper` | Bookkeeper / freelance accountant |
|
||
| `revops` | `/revops/` | `?from=revops` | Marketing / RevOps agency |
|
||
| `apex` | `/` | (no query — use `unknown` bucket) | Generic discovery |
|
||
|
||
When `unknown` exceeds 30 % of total, add UTM tagging to community
|
||
posts and SEO blog backlinks so you can break the bucket apart.
|
||
|
||
## 8. The four months that decide whether the plan works
|
||
|
||
Reading PLAN.md §3 + this doc together, the rough script:
|
||
|
||
| Month | What's running | What we expect to learn |
|
||
|---|---|---|
|
||
| **M1** (June) | Installers · demo · 3 landing pages · Gumroad live | Whether the funnel mechanically works. Numbers will be noisy; just look for one purchase. |
|
||
| **M2** (July) | M1 + community posts in 3 niches + 1 SEO post | Which persona converts. Re-allocate effort to the highest-converting niche. |
|
||
| **M3** (August) | M2 + landing-page changes from M2 review | Whether intent-rate moved on the change. Decide tools 06–08 go/no-go. |
|
||
| **M4** (September) | M3 + first repeat-buyer signals | Whether the Pipeline Runner is producing retention as designed. |
|
||
|
||
By end of M4, the data tells you whether the plan is producing
|
||
$1k–3k/mo (BUSINESS.md §6 6-month target) — extrapolated from the
|
||
trajectory, not the absolute number.
|
||
|
||
## 9. The hardest part of the plan to execute
|
||
|
||
Not the metrics. Not the build. **The "decide ONE thing per month"
|
||
rule** — operators with engineering backgrounds chronically pick
|
||
three changes per month and conclude nothing because their signal
|
||
is muddled. This doc says one. It means one.
|