Read across to compare shape and weight. The right template is the smallest one that still captures everything a future reader will need.
| Template | Category | Sections | Size | Best for | |
|---|---|---|---|---|---|
| ADR-Lite Common community variant | Lightweight | 3 (3 req) | ~100–200 words | Your team is just starting an ADR practice and any friction will kill it. | Open → |
| Agile ADR Hybrid (Atlassian, ThoughtWorks, IndustrialLogic patterns) | Reversible / agile | 6 (6 req) | ~200–400 words | The decision has a known shelf life — a vendor contract, a build-vs-buy that depends on team size, a temporary workaround. | Open → |
| MADR (Markdown Architectural Decision Records) Adam Bien, Olaf Zimmermann, et al. — adr.github.io | Rich (with options) | 8 (7 req) | ~400–800 words | There are 2+ real options and the comparison is the substance of the decision. | Open → |
| Michael Nygard's ADR Michael Nygard, 2011 | Standard | 5 (5 req) | ~250–500 words | Your team has never written an ADR and needs a low-friction starting point. | Open → |
| Tyree-Akerman Jeff Tyree & Art Akerman, IEEE Software 2005 | Rich (with options) | 12 (8 req) | ~1,500–3,000 words | The decision is permanent (one-way door) and the blast radius spans teams. | Open → |
| Y-Statement Olaf Zimmermann, 2018 | Lightweight | 1 (1 req) | ~50 words | You want a one-line decision register that humans actually read. | Open → |
Picking the heaviest template you've heard of. Tyree-Akerman written for a one-line call gets skimmed; ADR-Lite written for a one-way door gets re-opened in twelve months and nobody can reconstruct the alternatives. Match the template to the shape of the call, not to your team's appetite for ceremony.