Integration pages as a distribution system. Not a directory.

July 2024: the integration directory was a flat grid of logos nobody had touched since the original builds.
Twenty-one months later it was a funded Ecosystem Marketing program with CRO-tested LPs, paid coverage, marketplace listings on a weekly maintenance cadence,
and a personal Vercel staging environment because the production handoff was too slow to ship variants from.

HubSpot integration page: two refreshes, on record

The HubSpot page was the highest-traffic integration page in the catalog. I shipped two distinct rewrites. Both are in the Wayback Machine. Drag the divider to see what changed.

Refresh 1, Nov 2024 vs Jun 2025: the V2 structural rebuild lands. Before: November 2024, V2 just shipped (the +48% signups A/B test). After: June 2025, V2 settled into production.

Refresh 2, Jun 2025 vs Apr 2026: the AI-era restage. Before: June 2025. After: April 2026, AI Voice Agent and AIVA integrated into the page.

What it was

A logo wall. Pages built once and never revisited. No measurement, no attribution, no way to answer the CFO when she asked what the integrations team was actually shipping.

The diff

Before, jul 2024: a flat grid of partner logos. Pages thin enough that a buyer skimmed them and bounced.

After, mar 2026: an Ecosystem Marketing program. Four LPs on a personal Vercel staging environment (HubSpot and Salesforce, organic and paid variants). HubSpot marketplace listing rewritten for the first time since original publication. Salesforce CTI listing updated with a pricing flag. New Integration Hub with category filters replacing the flat grid. Paid keyword intel feeding copy decisions: HubSpot Dialer at 5.12% CVR, Salesforce Dialer at 3.70% CVR.

What I actually did

1. Phase 1 (aug–oct 2024): page-by-page VPC refresh on six pages. Outreach, HubSpot, Greenhouse, Bullhorn, Plecto, Pipedrive. HubSpot v1 A/B test landed +26.94% baseline (Wingify-measured). Signups 6.1% vs 4.8%. Demos 3% vs 2.4%.
2. Phase 2 (nov 2024): HubSpot v2 structural rebuild. New sidebar, tabbed content, section navigation, clearer pricing redirect, embedded activity log walkthrough. +48% signups. +25% SAL/Opp/Customer. +17% pricing clicks. +86% footer scrolls.
3. Phase 3 (dec 2024 → nov 2025): coverage expansion. Bitrix24, Attio, ServiceTitan, monday.com, Zoho, Jobber, CapsuleCRM, OnePageCRM. 123 demos and 309 signups via integration LPs in the jan–aug 2025 window (sep 2025 monthly report). Partner newsletter reached 591 partners at 28.79% open rate.
4. Phase 4 (mar 2026): finalized program strategy with the PPC, SEO, and performance leads. Deployed four LPs to a personal Vercel staging project. Paid variants shipped with ROI calculators, comparison configurators, and embedded product demo sections.
5. Marketplace work: HubSpot rewrite (3.8/5 with 26 reviews, targeting 4.5+) and Salesforce CTI pricing flag (JustCall $79 vs Salesforce native custom pricing).

What stayed honest

The HubSpot v1 +26.94% is Wingify-measured (third-party tool, not me reading my own dashboard). The v2 lifts (+48 / +25 / +17 / +86) are from the November 2024 monthly PMM report, self-reported. The 123 demos and 309 signups are an eight-month cumulative on a surface that previously had zero measurement, so the baseline was effectively a tenth of what the cumulative now shows. Phase 4 LPs lived on a personal Vercel project, not the production repo, because the production handoff would have eaten three weeks I didn't have.

What it became, in substrate

The page-by-page integration refresh method structured around the test that landed +26.94%, the marketplace-listing-as-channel pattern that treats listings as a maintenance cadence not a one-time publish, and the operator-owned Vercel staging environment that lets PMM ship variants without blocking on the production repo are three of the moves substrate carries forward. The skill, routine, and principle layers are open source for any client to clone and run.