Skip to content
SkyRosterBook a working session

Evaluate

A method that survives challenge, not a demo that impresses a room

Most rostering evaluations fail the same way: a vendor runs a canned scenario tuned to nobody's real rules, the room nods, and six months later the live rollout hits a constraint the demo never tested. Here is a shorter path, and then a plain statement of what we have to show for it and what we do not.

The method

Four steps, in this order

1 · Write your constraints down first

Before you see any vendor's screen, list your own hard and soft constraints from your collective agreement and the regulation that applies to you. A demo tuned to a generic scenario cannot fail against rules it was never shown, which is exactly why it always looks good.

2 · Ask for the mechanism, not the number

When a vendor states a percentage or a saving, the useful question is not whether you believe it. It is what was actually measured, on what sample, versus what was extrapolated or modelled. A vendor that cannot separate the two has not done the separation internally either.

3 · Test with your own data

A month that actually fought your current process, a real staffing shortfall, a real collision of two collective-agreement clauses, tells you more in one working session than a polished walkthrough of someone else's unit ever will.

4 · Ask what the tool refuses to do

A system that always says yes has no rules worth trusting. Ask to see a roster the tool rejected, and why, not only one it accepted. The refusal, and the reason attached to it, is where the actual rule engine shows itself.

What we actually have to show for it

Evidence you can check against a primary source

What exists

  • Five case studies, with their unit sizes and their method stated, published inside a EUROCONTROL rostering guideline that names SkyRoster as its source, not a testimonial we wrote ourselves. Read them against the source →
  • The guideline's own attribution, quoted exactly, with what it does and does not mean. See the exact wording →
  • A rule catalogue compiled from the live product on every build, not a document that can drift out of date from the software it describes. Open the catalogue →
  • A claims ledger listing every figure on this site against its source and its evidence state. Read the ledger →
published

What does not exist, said plainly

  • No named reference customer. No ANSP is identified anywhere on this site as a SkyRoster rostering customer, and no reference call is on offer. The organisations named in the EUROCONTROL guideline's acknowledgements contributed to that guideline; they are not disclosed as customers, and the guideline's own case-study customers stay anonymous, in the source and here.
  • No production solve-time benchmark. No controllers-per-second figure, no "N employees in M seconds" claim, appears anywhere on this site, because no production-scale benchmark exists to publish honestly.
  • No customer logos, no testimonials, no third-party review-site badge.

Where to go next

Depending on what you actually need

Bring the constraints you wrote down.

A working session is 45 minutes with your own rules on screen, tested against the same engine referenced on this page, not a canned scenario.