3. Feature-by-Feature Scrutiny (The Stated Idea)
Categories:
- Table stakes (TS): customers expect it. It earns nothing, but its absence costs you.
- Differentiator (D): a real reason to choose you over the alternatives.
- Distraction (X): built but unlikely to move revenue or retention.
- Liability (L): creates support, compliance, legal or trust risk.
The idea as stated had three parts: Notion as the content store, a public help center or knowledge base, and a form builder. Below are those parts and the features a buyer would expect from each.
3.1 Classification table
| # | Feature | Category | Reasoning |
|---|---|---|---|
| 1 | Notion as the content store | L | Every feature you sell is a feature Notion can ship. It already shipped Forms and Sites. Notion's API has also shaped your ceiling before: NotionForms launched on the day the API opened and grew because nothing else existed. That window is closed. |
| 2 | Form builder writing to Notion | X / L | Notion Forms is free on all plans. Tally's free plan has unlimited forms and a Notion integration. NoteForms, Fillout and OpnForm cover the power features. There is nothing left to sell. |
| 3 | Public help center from Notion pages | TS | HelpKit, NotionDesk, Notaku, Super, Bullet and Potion all do it. Notion Sites does a basic version free. |
| 4 | Custom domain | TS | Notion Sites sells it for $8–10 a month. Every help-center tool includes it. |
| 5 | Visitor search | TS (for now, a gap in Notion Sites) | HelpKit's main argument against Notion Sites. A gap that exists only until Notion ships search. |
| 6 | Themes and branding | TS | Everyone has them. Competing on design means fighting Scroll and Refined on their best ground. |
| 7 | AI chatbot over the docs | TS → L | HelpKit sells it at +$49. Scroll and Refined include AI search. Without citations, a wrong answer is a support ticket and a trust problem. |
| 8 | Embeddable help widget | TS | HelpKit has one. Help desks (Intercom, Zendesk) have their own. |
| 9 | Analytics (views, searches) | TS | Expected. Only useful when it leads to action. See the doc-owner loop in 04. |
| 10 | Access protection (password, email, SSO) | TS | HelpKit and Scroll both have it. Needed for partner or customer-only docs. |
| 11 | Multi-language | TS / X early | HelpKit has it. Translation workflows are a support burden. Add only when paying customers ask. |
| 12 | SEO and AI-readability (sitemaps, llms.txt) | TS | HelpKit and Scroll both market it. |
3.2 Scoreboard
| Category | Count | Features |
|---|---|---|
| Table stakes | 9 | Help center, custom domain, search, themes, widget, analytics, access protection, multi-language, SEO/llms.txt |
| Differentiator | 0 | None in the idea as stated |
| Distraction | 1 | Forms |
| Liability | 2 | Notion as the only content store, uncited AI answers |
None of what was described is a reason to buy from a newcomer in 2026. Every item is table stakes, already free, or both.
3.3 Building to avoid selling: explicit call-outs
- Forms. A 6-day MVP in 2021. In 2026, a clone of a free native feature.
- Themes and page builders. Fun to build, easy to demo, and exactly what the incumbents already do better.
- A generic AI chatbot. It demos well and sells to nobody who has Intercom Fin, Rovo or Notion agents.
- "Support any source." Notion plus Confluence plus Google Docs feels like a bigger market. It's three integrations and no buyer.
3.4 What a disciplined v1 looks like (the Confluence wedge)
- Keep: publish one Confluence space as a public help center on a custom domain, full-text search, a help-center theme, article feedback, and the doc-owner loop (failed searches, "not helpful" votes and unviewed pages filed weekly as Jira issues or Confluence tasks).
- Add after 10 paying: AI answers with citations (Advanced edition), gated sections, a Zendesk and Intercom import.
- Cut: forms, Notion, Google Docs, page builders, internal wiki features, Data Center.