This is a short report on what I built on the site this month. Some things worked, some things were broken and got fixed, and some things are still to do.
Every quarter I'll write down what shipped, what didn't, and what changed. This first note covers the start of October 2026. A roadmap explorer, collection pages, a game about AI dilemmas and audio summaries arrived. Broken links got fixed. Lots of collections were still empty, and filling them is the next job.
Q4 2026 interim note covering 10 to 11 October. Shipped: Microsoft 365 roadmap explorer, per-collection pages with a build-time link check, AI trolley problems, page media and a podcast feed. Open: empty collections (in review), method write-ups, tooling. Measures are delivery against plan; no traffic or commercial metrics are reported.
Why I write these
The plan is a decision model, and decisions need reviewing. A quarterly note is that review in public. I write down what I said I'd do, what actually happened, and what I learned. This is the first one, so it covers a short stretch at the start of October 2026 and sets a baseline.
All the facts here come from the site's own change log.
What shipped
On 10 October, the Roadmap page became a searchable explorer for Microsoft's public Microsoft 365 roadmap. You can filter by product, status, release phase, platform, cloud and date, and every item links back to Microsoft. If the live feed is unavailable, it falls back to a saved snapshot.
The same day, AI trolley problems went up: five tracks of twelve dilemmas each, for boards, leaders, operations, agent builders and users, with a report at the end.
On 11 October, the home page, sections and existing articles gained images, an infographic and a short spoken summary each, and a podcast with one short episode per page. The voices are synthetic, and each page says so.
What was broken
The bigger lesson of the month was about things that looked finished but weren't. Every collection card on the site looked clickable, but none of them led anywhere. Some section headers had dark text on a green background that was hard to read.
Both are now fixed. Each collection has its own page, and the build now checks every internal link on every page. If any link points nowhere, the build fails. I'd rather the site refuse to publish than publish a dead end.
What I planned but haven't done
Many collections were empty. A collection page that says "nothing here yet" is honest, but it isn't useful. Filling them is the work in progress as I write this, and it's in review rather than live.
Some planned pieces are still waiting, such as the full write-up of my Decision Diagnostic and the tools listed under What I'm building.
A worked comparison
This is a simple planned-versus-actual count of the kind I'll use each quarter. I didn't write a formal plan before this quarter, so the planned column is illustrative. The done column follows the change log.
| Item | Planned | Done | Note |
|---|---|---|---|
| New features | 3 | 3 | Explorer, trolley problems, media |
| Fixes | 2 | 2 | Collection links, contrast |
| Empty collections filled | All | In review | Biggest open item |
| Tools | 2 | 0 | Next quarter |
What I learned
A site can pass a visual check and still be broken. Automated checks catch the dull failures, so I can spend my attention on whether the writing is clear and correct.
What I'm still checking
I haven't decided what to measure each quarter beyond delivery against plan. I don't want traffic numbers to drive the writing, and the site carries no offers until 2027, so there are no sales figures either. Something like "questions answered" might be better, if I can find an honest way to count it.
Sources
- craigstanley.work change log (CHANGELOG.md in the site repository), entries for 10 and 11 October 2026, accessed 11 October 2026.