Eight products, locked before 96 challengers were scored

Eight products are what I am building right now: a WordPress plugin, two Shopify apps, two web apps, two iOS apps and a Mac app. None of that came from a flash of inspiration. It came from scoring, a decision on 4 September 2026 that locked the eight, and a full day after that spent trying to beat them. The order matters: the eight were locked on 4 September. The 96 challengers were scored on 5 September, against a portfolio that already existed — not the other way round.

How a candidate gets a number

Every product idea — the eight that made it and the ones that didn’t — gets scored on the same seven-part rubric, each part rated 0–10: distribution, willingness to pay, demand, how hard the core problem really is to copy, whether one person can build and run it solo, the risk that a platform builds it in or a live competitor already has it, and synergy with whatever else is already in the portfolio. The seven combine into one weighted score. Distribution carries more weight than any other single criterion, at 30%. How hard the real problem is to copy carries 15%. The remaining five criteria split what’s left.

Why distribution outweighs everything else

I’m a solo founder with no existing audience, and the rubric is built around exactly that constraint. It doesn’t ask whether the pain is real, or whether a well-known company could sell a solution — plenty of candidates clear those bars. It asks something narrower: is there a documented case of an unknown founder reaching real paying customers through this exact channel? Of the 16 challengers that got a full deep-dive, 14 were capped at 5 out of 10 on that question, because none had such a case for its own channel. WordPress.org has two on record, both for accessibility and security plugins: SSL Zen reportedly reached about $5,000 a month within five months back in 2020, and WP Activity Log grew from one developer working alone since 2013 to 300,000 installs. Those cases sit in the same category as the WordPress accessibility scanner, which scored highest in the portfolio. In the 5 September run, the only two products above 6.0 were that scanner and TrueMargin, which had a channel case of its own: fast-growing seller communities on Reddit, and a competitor that had just raised its entry price and left a gap underneath it.

Eight products, locked first

Product	Platform	What it does	Price
A11yLog	WordPress/WooCommerce plugin	Scans a site with axe-core, writes real fixes into its HTML, keeps a dated evidence log	Free scan, $79–$249/year
Settled	Shopify app	Payout calendar, hold/reserve flags, plain-English reason a payout doesn’t match sales	$19–$79/month
Settled Disputes	Shopify app	Assembles chargeback evidence — tracking, photos, messages, policy snapshot — into one reviewed PDF	$29–$79/month
TrueMargin	Web app	True profit per order, SKU and channel from CSVs sellers already export from Etsy, eBay, Amazon, Shopify	From $12/month
ReconMatch	Web app	Matches deposits from five sources to QuickBooks Online or Xero, explains what doesn’t match	$19–$79/month
Visual Schedule	iOS app	One-step-at-a-time schedule for autistic and ADHD kids, synced live across caregivers	Free, $6.99/mo family tier
EvidenceVault	iOS app	Timestamped, geotagged proof captured while packing an order or checking a guest in	Free, $7.99/mo Pro
TrueMargin Desktop	Mac app	Same profit math as TrueMargin, running locally, no account, no network call	$59.99 once

I locked this shape — this exact mix of platforms, all eight built in parallel — overriding my own earlier rule against launching products in parallel. The 30-hour monthly maintenance ceiling stayed, and the math against it is blunt: maintenance once all eight are live and supported is estimated at 42–61 hours a month, one-and-a-half to two times the ceiling. That doesn’t add up on its own. The decision names two ways out: kill gates that cut the portfolio to five or fewer, or raising the ceiling.

What a kill gate actually is

A kill gate is a number agreed on before anyone knows the answer — a bar set before the test runs, not a report card written after.

A11yLog’s pre-code bar is emailing 20 agencies and universities from a list of 30 sites already scanned for issues, cleared only if at least 5 agree to a pilot and 2 pay for one — short of that, a different WordPress candidate is meant to take the slot. After launch, a separate bar is set: fewer than 300 organic installs after 60 days, or under $500 in monthly revenue 90 days after the paid tier goes live, and the plan calls for killing it. TrueMargin’s pre-code bar is a free calculator plus a post in a seller community, needing at least 100 sign-ups and 10 real CSV uploads to clear. There’s a portfolio-wide version too: six months after every product is live, the plan is to keep at most five, the ones with revenue or growth, and close the rest or put them into maintenance — a direct answer to the maintenance math above, since five products against a 30-hour ceiling is a different proposition than eight.

Then, 96 more ideas

A day after locking the eight, the question became: is there something better I missed? Six channel memos and 782 new Reddit posts from the previous 30 days fed 96 new candidates, scored against the same weighted rubric. Sixteen of the strongest got a full deep-dive, and an independent verification pass then rechecked 91 of the numbers cited across those: 81 held up outright, 4 were directionally right, 3 were wrong and got corrected without changing any score, and 3 couldn’t be verified at all with the tools on hand.

The result: zero new candidates crossed 6.0 after the full deep-dive. One looked, briefly, like an exception. WooRecon — a tool to match WooCommerce payouts to bank deposits — got an initial quick score of 6.30 from one of the two candidate-generating passes; a second, independent pass scored the same idea at 5.30, because it inferred the new demand only from a payment-gateway signal rather than direct evidence. Quick scores aren’t the final word, though. WooRecon’s full deep-dive knocked it down to 5.20: real competitors already cover this ground, and the fresh evidence wasn’t strong enough to justify a ninth product. It didn’t disappear — it became a data-ingestion path inside the reconciliation engine that four of the eight products already share, rather than a product of its own.

So the portfolio held. The only two locked products that scored above 6.0 in this same run were the WordPress scanner (6.45) and TrueMargin (6.25) — the two with the channel evidence described above. What the day did produce: a lower risk score for Settled once the run found Shopify had shipped its own payout-to-bank matching, a documented list of exactly which piece of evidence is missing for every near-miss candidate, and six product features worth folding into the eight rather than building as separate apps.

Where the eight actually stand

Product	Status
A11yLog	In development
Settled	In development; Shopify App Store review next
Settled Disputes	In development
TrueMargin	In development
ReconMatch	In development
Visual Schedule	In development; not on the App Store yet
EvidenceVault	In development; not on the App Store yet
TrueMargin Desktop	In development; follows the web version

That’s as of September 2026. Every one of the eight has real code behind it. None has launched — no installs, no App Store listing, no paying customer. The scoring said which eight to build and which 96 not to. It has nothing to say about whether any of them will actually work. That part still gets tested against strangers, not against a rubric.