SponsorLab Roadmap Plan
Enter hub password to continue.
That's not right. Try again.
SponsorLab Roadmap Completion
Written 2026-07-15, updated 2026-07-15 after execution
This was the plan to close every still-open item from the SponsorLab site 100x roadmap, and most of it is done. Of the 20 roadmap items, 17 are closed and live on sponsorlab.ai, one is retired (the pink accent), and one is deferred (Review schema, it needs real reviews first). The logo wall is the last item still building. A new item turned up mid-execution, Task 10, and it's the most important thing left: the site's headline proof numbers are stale.
The most important thing left: your proof numbers are stale
The site publishes "136 creators, 365 brands, 685 verified payments" in the hero, the proof section, the footer, and on every vertical page. The real current numbers, verified against live D1 today, are 186 creators, 430 brands, 710 payment links.
So the site understates itself. That's the safe direction to be wrong in, but it still has to be fixed, because those three numbers are the site's entire differentiation from a generic AI sponsorship tool.
The creator count was already stale before today: 136 was true on July 13, and 20 women's-midlife creators were added on the 14th without the page being updated.
One caveat to fix before publishing a new brand figure: 6 brand rows carry fabricated placeholder domains (things like unknown-scott.example) left over from an old discovery pass. They need to be merged or removed, which puts the honest brand number closer to 424 than 430. None of them appear in any outreach draft, so nothing was ever going to be pitched to a fake domain, but they shouldn't be counted in a public number either.
There's also a data-quality gap that's silently undercounting evidence: creator_sponsors has four different spellings of the same brand ("Adidas", "Adidas Terrex", "adidas Terrex", "Adidas Five Ten"), but the brands table only has "Adidas Terrex", so three of the four don't join. That kind of spelling drift is worth a cleanup pass before the next number refresh.
Where all 20 roadmap items actually stand
| # | Roadmap item | Status | Note |
|---|---|---|---|
| 1 | Rewrite hero to brokerage positioning | done | Shipped in the redesign |
| 2 | Fix follower framing to 50k-500k | done | Zero "200K" or "5M" left on the page |
| 3 | Name the six verticals | done | The 01-06 numbered beats |
| 4 | Evidence-database proof section | done | Proof numbers shipped, now stale, see below |
| 5 | "For Brands" path | done | A real brand inquiry form shipped, replacing the mailto |
| 6 | Swap green accent for pink #d81b7a | retired | Will not do. Pink is retired globally. |
| 7 | Replace unsourced stat tiles | done | Now backed by the real database numbers |
| 8 | FAQ + FAQPage schema | done | "Questions we get asked" plus schema |
| 9 | Title/meta keywords | done | Rewritten to brokerage language |
| 10 | Six vertical landing pages + sitemap | done | Shipped as seven pages, not six. Sitemap went 1 URL to 8. |
| 11 | Real imagery (page has zero images) | building | Last item. The logo wall. |
| 12 | Creator/Brand split above the fold | done | Two CTAs live above the fold |
| 13 | Illustrative outcome story | done | The record card |
| 14 | Verity Agentic footer credit | done | Cut |
| 15 | Brand-side contact form | done | Backend built, deployed, and tested end to end before the form shipped |
| 16 | Citable aggregate stats sitewide | done | Shipped, now needs the number refresh below |
| 17 | Expand JSON-LD (Review/AggregateRating) | deferred | Needs real reviews. See below. |
| 18 | Sharpen beta urgency with a real number | done | "Limited spots open" cut. "Currently in private beta" kept, it's true. |
| 19 | Re-check Core Web Vitals after images | open | Pending, rides with the logo wall |
| 20 | Back the "preferred source" link with content | done | Cut the banner instead |
Two items I am deliberately not implementing
Item 6, the pink accent, is retired
The roadmap was written 2026-07-13 and told us to swap the forest green for pink #d81b7a. Your 2026-07-14 order retired pink and pastels globally, and you reconfirmed it today. The site already ships the correct palette: monochrome plus burnt orange #c2410c, with zero occurrences of the old pink. Every page this plan creates inherits that palette. Nothing in this plan makes sponsorlab.ai pink.
(This hub page keeps the hub's own pink chrome, per the standing rule not to surprise-repaint existing sites. That rule covers the hub and Rebuilding; sponsorlab.ai is on the new palette.)
Item 17, Review/AggregateRating schema, is deferred
Review schema requires real reviews. SponsorLab has closed zero deals and has no customers, so populating it would mean fabricating testimonials. That's both a Google structured-data policy violation and dishonest. We revisit it once real deals close. The FAQPage half of item 17 already shipped.
Your four new verticals: one earned a page
Four candidate verticals were tested for evidence density beyond the original six: pickleball, women's strength, HIIT, and sauna. Climbing, one of the original six, is in the table below for comparison. The numbers count evidence on PROSPECT-tier creators only, the pitchable tier, excluding benchmark creators who are learn-from-only and never pitched.
| Vertical | Prospects | Evidence links | Per creator |
|---|---|---|---|
| Climbing | 12 | 46 | 3.83 |
| Pickleball | 10 | 14 | 1.40 |
| Women's midlife | 27 | 7 | 0.26 |
| Women's strength | 8 | 2 | 0.25 |
| HIIT | 4 | 0 | 0 |
Pickleball shipped. It's real: 14 evidence links across 8 creators who have evidence, out of 10 prospects, all US-weighted, verified via Apify. Real sponsors include Selkirk Sport, Franklin Pickleball, Stealth Belt (explicit "paid partnership"), Tifosi Optics, and Medistik (#ad). Live at sponsorlab.ai/sponsorship/pickleball.
Women's strength did not ship a page. A research agent labeled it "strong, ship it," but that verdict counted two benchmark creators, Sohee Carpenter (667k) and Meg Gallagher (612k), whose sponsors don't prove the pitchable lane. Strip them and the entire evidence base is one creator, Alex Silver-Fagan, with two brands. It does have 8 real US creators though, so it's fixable with a deeper scrape, not a dead end.
HIIT did not ship. There are real US creators (Juice & Toya, Sydney Cummings, Danielle Pascente, Heather Robertson) but zero qualifying sponsors. Several had bare bio brand tags with no #ad, code, or partnership language, and those don't count as evidence.
Sauna did not ship, and this one is structural rather than a scraping miss. That space is dominated by brand accounts (Plunge, HigherDOSE, Sun Home Saunas) and studios, not individual mid-tier creators. Only 4 individual creators surfaced at any size, only 2 are US, and neither US one has qualifying evidence. The market to broker into doesn't really exist yet.
Shipping 2 strong verticals beats 4 hollow ones, because an empty vertical page contradicts the "we have 710 verified payments" pitch the whole site rests on.
Why the women's lane was never a brand problem
You asked why there aren't more brands for women's midlife, expecting the US telehealth menopause providers and women's activewear brands to already be in the database. They were. Midi Health, Alloy Women's Health, Ritual, Vuori, Athleta, Alo Yoga, and Lululemon were all there already, with verified emails and budget signals of 78 to 86.
The bottleneck is evidence, not brands. The matching engine only fires when a brand has proven payment evidence to a creator in that niche, which is the whole 710-payment value prop.
The lane has 33 creators, more than any other vertical, and produces just 10 evidence links. Outdoor/adventure has 11 creators and produces 91.
The reason is visible in the data: the creator set skews UK and Australian, so the evidence found is British. Sweaty Betty, Wild Nutrition, Ancient + Brave, and Specsavers, which is a UK optician. Specsavers and Massimo Dutti appearing as "sponsors" of a fitness creator means the extractor was catching casual brand mentions, not real sponsorships.
44 brands were added anyway (369 to 413), filling the genuine US telehealth gap: Evernow, Winona, Gennev, Elektra Health, Maven Clinic, Kindra, Womaness, Joylux, MyMenopauseRx, Stella, Peppy, Versalie, and Hers. Plus activewear: Gymshark, Fabletics, Girlfriend Collective, Outdoor Voices, Beyond Yoga, Oiselle, Halara, Carbon38, Varley, Under Armour, and Aerie. 16 of these have verified emails.
The real fix is US creators, a deeper scrape than 12 recent posts, and a tighter extractor.
Four candidates were rejected rather than faked: Pause Well-Aging (the domain is squatted, it now resolves to a Vietnamese football streaming site), Midday/Lisa Health (DNS returns SERVFAIL, it doesn't resolve), Rory (folded into Ro, no longer a distinct menopause brand), and Amberly (no such company exists, only "Amberen," which is unrelated).
The shared-footer email trap was also caught: the same address, [email protected], appeared on CRZ Yoga, Terry Bicycles, and Bala, three unrelated sites running one agency theme. All three were demoted from "verified" to "pattern-guess."
The nine tasks
The failing test for the whole plan. Assert the seven pages and both forms, watch it fail, then build until green.
Goes first and alone because it is the template for the other five. Real per-vertical numbers pulled from D1, not guessed. Stops for your review before cloning.
Gravel/MTB, backcountry ski, endurance, nutrition and recovery, women's midlife fitness. Each gets its own copy, its own D1 numbers, its own FAQ.
One URL becomes seven. The six verticals on the homepage become links. Sitemap submitted to Search Console (needs your Google session).
Blocked on decision 1. Licensed, optimized under 300KB, explicit dimensions so it doesn't tank CLS, and never captioned as proof.
Two CTAs under the hero, stacking below 640px, verified at six widths.
Blocked on decisions 2 to 5. Footer, banner, naming, urgency.
Backend first: migration, handler, deploy, end-to-end test, THEN the form. No broken forms ever ship.
Validator twice (transactional site), Core Web Vitals, and a regression check on the beacon and the leaked-file redirects.
The five decisions, all answered and implemented
You answered all five on 2026-07-15, and all five are now implemented: the Verity footer credit is cut, the preferred-source banner is cut, the site reads "SponsorLab," one word, sitewide, the beta scarcity line is cut, and the brand logos are cleared and captioned factually. Recording the reasoning here so it survives into future sessions.
1. Imagery
ANSWERED: brand logos are cleared, we have permission to use them. Task 5 uses a real brand logo wall, captioned factually (see the note below). Creator headshots were NOT cleared, so no creator faces go on the site until you say otherwise.
2. Verity Agentic footer
ANSWERED: CUT. Removing the credit, the outbound link, and the parentOrganization entry in the JSON-LD.
3. Preferred-source banner
ANSWERED: CUT. Item 20 closes as "cut" rather than "backed with content."
4. Naming
ANSWERED: SponsorLab, one word, sitewide.
5. Beta urgency
ANSWERED: there is no real cohort number, so the scarcity line is cut rather than given a fabricated figure. "Currently in private beta" is true and stays; "Limited spots open" is the unsourced claim and goes.
One honesty constraint worth flagging
The women's midlife fitness vertical page ships without a proven-brand list. Every other vertical has real proven payers in the data. This one's only proven payers are WonderBiotics and YES/Yes Organics, each evidenced by a single affiliate post with a discount code. That is not a credible public proof list. The brands with real money in that lane (Ritual, Momentous, Midi, Alloy, Vuori) produced zero warm matches because the scrape found no proven-payer evidence for them, so naming them on the page would be fabrication. The page leans on the 33-creator supply side instead until you decide how to fix the underlying evidence problem.
What this plan does not cover
- The women's midlife evidence problem itself: scrape deeper, add a cold-but-high-value pitchable tier, or accept the lane needs more evidence. Still your call.
- Verifying the 99 pattern-guess brand emails before any send.
- The per-brand-approval send workflow on the wired SES.
- Review/AggregateRating schema, deferred until real deals close.
Loose threads
- The brand-inquiry notification email to you could not be directly confirmed via server logs (a wrangler pages deployment tail produced no output in this environment). The synchronous confirmation email path succeeded on every test and shares identical code, so the evidence is strong but indirect. Worth one real submission from your side to be certain it reaches your inbox.
- A live Pexels API key (.pexelskey) was found untracked at the SponsorLab repo root, created by an imagery agent, with nothing in .gitignore covering it. It was never committed, so nothing leaked, but any blind
git add -Awould have published it. .gitignore was widened to cover bare key files. - The women's-midlife and women's-strength lanes both need a deeper scrape before they can carry a brand list.