Apply
Apply

Marketing & Sales

Give Notebook, skills, X, and tools at the top. Nurture. Convert. Conversion is a completed funding application in Funding Chat.

Updated 2026-10-31. Current version 1.0.0. The skill URL stays /skills/marketing. This lab reads /v1/funnel/summary.

Scanned clean | SkillSpector | v1.0.0 | 2026-11-01

This scan is not a certification.

Tests

DateWhat was comparedMetricWinner
2026-10-30Landing-page design bake-off on one mcafax.com draft: pbakaus/impeccable, Nutlope/hallmark, elayadesign/ai-design-skills, and emilkowalski/skills.Score out of 10.pbakaus/impeccable, 8.1. Hallmark 7.9 (runner-up). elaya 7.3. emilkowalski/skills 7.0. Only the winner's practice is the one the skills hub already follows. The others were not added.
2026-10-30DSPy headline and meta rewrite against the current headlines, on 20 held-out Notebook pages. Local 1.5B model.Score on the held-out pages.DSPy, 49.67 to 65.08. Recommended for a daily run. Not installed in the marketing skill.
2026-10-30Mem0 lesson recall against a keyword lookup.Recall at 3.Mem0, 0.59 versus keyword 0.36. Not installed in the marketing skill.
2026-10-30GEPA prompt rewrite against the starting prompt. Local 1.5B model.Prompt score.No winner. 0.1283 to 0.1284 (flat). Not added.
2026-10-30Agent Lightning topic agent against its starting accuracy.Accuracy.No winner. 0.55 to 0.475 (worse). Not added.
2026-10-27Self-hosted analytics: Plausible Community Edition against Umami.Weighted score out of 45, then the article-engagement tie-break.Plausible Community Edition. Both scored 34/45. Plausible won the engagement tie-break. Not installed on the site.
2026-10-27Instagram profile bio variants. Designed only.Follows per profile visit and link taps. Not measured.No winner. The test was not applied. Nothing on the profile was changed.

How we got here

Version 1.0.0, dated 2026-10-26.

What is being tested next

Visible updated date and a lead image for AI Mode

Queued 2026-10-28. A visible updated date and a strong lead image on Notebook pages, against the current page template, after a reported AI Mode test that lists organic links with a description, an image, and a date.

Not set beyond AI Mode clicks and citations. No result yet.

Micro skills

No micro skills meet the research bar yet.

Funnel experiments

Funnel experiments only. A row with an empty result is status running. No funnel experiment is logged yet. A row with an empty result has status running.

HypothesisChannelStageVariantMetricStartEndResultWinnerStatus
No experiments logged yet.

Machine-readable: /v1/marketing/experiments. Content-lab winners, when the record has one, are a different log: /v1/marketing/playbook.

Tactics research

Tactics research. Entries are added when supplied. None yet.

TacticCompanySource URLQueued as test
No tactics logged yet.

Machine-readable: /v1/marketing/tactics.

Conversion by source

Counts are from the application store. A row is a source token and two counts. Public routes omit names, emails, and phones.

apply_started 15. apply_completed 0.

Last touch

Sourceapply_startedapply_completed
direct50
notebook-oscar50
nav20
small-business-grants-open-now20
tipt-tiptree-70-percent-of-cash-after-fortegra-sale10

First touch

Sourceapply_startedapply_completed
direct110
notebook-oscar40

A page stores the source in this browser (first touch once, last touch when a new utm or ref arrives). Starting Funding Chat posts apply_started. A completed application posts apply_completed from those cookies. The public response is counts by source.