40+ products, four capability categories, refreshed quarterly. It stopped being a document and became the thing teams consulted before prioritising — and four roadmap items came out of it, one of which now generates revenue.
Cadmus doesn't have one competitor, it has four different kinds.
Every conversation about “what the market does” was really a conversation about whichever product the speaker had seen most recently.
Platforms that run the assessment itself — authoring, sitting, submitting.
The LMS layer that owns the course, the cohort and the gradebook.
Tools that help students produce the work — drafting, referencing, feedback.
Similarity detection, AI detection, proctoring and process assurance.
Worrying about a “competitor” operating in a different category entirely.
Missing a real gap because the product holding it wasn't the one anyone was thinking about.
Prioritisation errors caused by a missing artefact, not by bad judgement.
Marketing pages describe intent; feature lists describe ambition. Neither tells you if a capability is usable.
The most direct route to a real demo. A vendor's pitch shows what they lead with and what they avoid — neither is on the website.
Published demos and user-recorded walkthroughs, which often reveal the awkward parts of a workflow that an official demo skips past.
Support docs and help centres are the honest version of a feature list — they describe limitations because they have to answer real user questions.
Where a product could be accessed directly, I used it as an academic would — the approach that produced the Smart Comments differentiator.
Any team could filter to their slice — group assessment only, integrity only — instead of reading 40 products to answer one question. A reference nobody can navigate gets rebuilt from scratch every time.
The refresh is what makes it an instrument rather than a document. A stale comparison is worse than none — it produces confident wrong answers.
The test of competitive research isn't whether it's thorough — it's whether anything is different because it exists. Here's the chain.
Engineering was pointed here for group assessment prioritisation, filtered to the competitors that mattered.
Checking whether a proctoring provider already had an invigilator-key equivalent before we scoped ours.
Read the Work Playback case study → · Read the Smart Comments case study →
The most valuable thing a comparison gives you isn't a list of gaps to close — it's the confidence to leave some open.
Where a missing tick loses a deal regardless of how good the rest of the product is. These are table stakes: you don't win on them, you only lose without them.
Categories where the sheet showed Cadmus holding a genuine, defensible position — worth investing in precisely because competitors can't easily follow.
A parity gap is a cost of doing business — close it cheaply. A differentiation gap is a bet worth over-investing in. Treating both the same way is how roadmaps become feature-matching exercises.
I never instrumented the artefact — I know it was used, but not how often, or which categories drove decisions. That would have told me where to spend refresh effort. I'd also have written the parity-versus-differentiation call into the sheet rather than carrying it in my head.
Booking the sales calls. Slightly absurd, and the single highest-yield method I used — you learn what a company thinks its own strengths are, which isn't in any documentation. It led directly to the Smart Comments insight, so the return was concrete.