User research findings as a private HTML report (2026)

The research lead ships the findings deck Friday afternoon — 47 slides, 80MB, embedded video clips, the methodology appendix, the verbatim quote highlights, the recommendations. Stakeholders open it over the weekend.

Sunday night, the VP of product opens the deck on a phone. The first slide renders at 12-point type. The embedded video doesn't play in the mobile slide viewer. The chart on slide 23 — the one showing the segmentation difference — is a 1200-pixel image scaled down to a thumb-sized box.

Monday's review opens with three minutes of "let me find the slide where..."

The content was fine. The format mismatch was the bottleneck.

What UXR deliverables actually need

Four properties that recur across most research workflows:

1. Multi-stakeholder accessibility. PMs, designers, leadership, and engineers all read findings; each reads on different devices, in different windows, looking for different signal.

2. Interactive evidence. A direct user quote is more persuasive than a paraphrase; a 15-second video clip is more persuasive than a verbatim transcript. Static formats lose the interactivity that makes the evidence land.

3. NDA-bound by default. Research participants are usually anonymized but the data is sensitive — competitive intel, pre-launch findings, internal pain points. The container needs access control matching the engagement's confidentiality.

4. Multi-month reference life. Findings get cited during sprint planning, design reviews, roadmap conversations, post-launch retrospectives. The deliverable outlives the readout meeting.

The typical slide-deck format addresses none of these particularly well; it was built for the read-out meeting, not for what happens before and after it.

What each delivery format does

Slide deck (.pptx, Keynote). Designed for the read-out moment. The presenter walks the room through; the slides anchor each section; the meeting is the deliverable. Right shape when the readout is the only consumption moment; misses the phone-pre-read and the months-later re-read use cases.

PDF export of the deck. Designed for archive. Pages don't reflow on mobile; embedded video becomes a flat thumbnail; the methodology appendix is a 30-page scroll that nobody scrolls through. Right shape for the regulatory-archive copy; lossy as the working deliverable.

Notion page. Designed for collaborative working documents. Strong for ongoing edits, comments, multi-author contributions. The Notion subdomain in the URL reads as a working doc rather than as a finished deliverable; per-document permissions exist but are coarser than what an NDA usually requires.

Dedicated research repository (Dovetail, EnjoyHQ, Condens). Designed for ongoing research ops. Tag user quotes across studies, search across the history, build the institutional memory of the research function. Right shape for the team's working environment; less natural as the deliverable shape — the consumers (PMs, executives) shouldn't need a repository login to read a report.

Self-contained HTML at a private link. Designed for delivery. The page renders cleanly on mobile, holds embedded video, supports interactive evidence (hover for transcripts, expand for methodology, click for raw quotes), and lives at an unguessable URL with optional password protection for NDA-bound material.

What changes when findings ship as HTML

Three small properties compound across the lifecycle:

1. Mobile-readable from the first paint

Executives read findings on phones between meetings. An HTML page sized for mobile renders the first chart at a legible size, lets the executive scroll through the executive summary, and bookmarks the URL for re-read.

For internal review cycles, the mobile-first read is the moment that decides whether the findings drive the next conversation or sit in someone's inbox waiting for "when I get to my laptop."

2. Interactive evidence stays alive

UXR's strongest evidence is the verbatim — here's exactly what the user said, in their own words. Static formats compress this to a pull-quote; HTML lets the quote live as text with a hover-to-expand for the surrounding context, or a play-button for the 15-second clip.

The same goes for interactive charts — segmentation differences are more legible when the reader can filter by demographic themselves, instead of seeing a static screenshot of one filter state.

3. One URL outlives the read-out meeting

A PM cites the research in sprint planning three weeks later. The URL still works. The methodology section is one anchor link away. The specific quote about the friction point opens with one tap.

When the findings update (a follow-up study, refined interpretation), the URL stays the same and the content updates. No "v3_REVISED" file proliferation across the team's drives — the topic covered in why HTML beats PDF for deliverables at a more general level.

NDA + expiry — the privacy combination

Research findings are usually covered by an NDA between the research team and the broader org (or between an external research consultancy and the client). The container should match the NDA's actual properties:

• Default-noindex — the findings page is not findable in search. The unguessable URL is the access control; randomly typing URLs doesn't surface the report.
• Password protection — for findings going to a specific stakeholder list, the password layer keeps casual forwards stopped at the prompt.
• Expiry tied to the engagement window — for studies covering a specific quarter or product cycle, the expiry window maps to the relevance of the findings (90 days for quarterly studies, 6 months for product-launch research, permanent for foundational research a team will reference for the lifecycle).

The same controls that work for consulting deliverables and agency previews apply here; the UXR-specific layer is whether the participant data is anonymized enough to be shared at the scope the link is going to.

The build path

For research teams already using a notebook or markdown-based workflow:

• Quarto / Observable / Jupyter export — HTML is the native format; the export pass is mostly making sure embedded videos and interactive plots are inlined (or stay as small external references where appropriate).
• AI-built reports (Claude, ChatGPT for synthesis) — already HTML-native; the cross-tool prompts for portable HTML cover the output shape.
• Slide-first teams — the practical path is to keep the slide tool for the read-out anchor and ship the parallel HTML report for the consumption use case. The slide deck handles the live presentation; the HTML report handles the rest of the lifecycle.

For teams that already use a research repository (Dovetail, EnjoyHQ), the typical pattern is: repository for the team's working environment; HTML report for the deliverable to consumers. The two formats handle different consumption shapes.

What this isn't

The HTML report shape doesn't replace:

• The research repository — for ongoing tag-and-search across studies, the dedicated repository is the right shape. The HTML report is the deliverable output of that repository.
• The read-out meeting — for the synchronous interpretation conversation with the team, the meeting is the right format. The HTML report covers the async re-read after.
• Raw data and recordings storage — sensitive recordings need their own access-controlled storage; the HTML report references and embeds clips, but the recording archive lives separately.

The right framing: the HTML report is the deliverable that fits how findings are actually read after the readout meeting — mobile, async, multi-stakeholder, over months. For everything else (research ops, recording storage, live presentation), the right shape is a different tool.