Every guide, in one list

Most selections start with a product somebody liked. The requirements are then written to describe it.

A list of forty requirements that every serious product satisfies has not narrowed anything. It has just taken three weeks to write.

Every selection contains a requirement that eliminated a good product and mattered to nobody once it was gone.

Selections fail late because somebody with a veto was consulted after the decision rather than before it.

Every business case compares products to each other. The comparison that matters is against changing nothing at all.

The build option is usually dismissed on cost and the buy option on fit. Both dismissals are made too early and on the wrong grounds.

The narrowing is where the decision is really made, and it is usually done by whichever vendor markets hardest.

Marketing pages are optimised to survive comparison. The informative parts are the documentation, the pricing page and the changelog.

A standard demo is a rehearsed path through the product's strongest features. You learn nothing you could not have learned from the website.

Vendor-supplied references are selected to be positive. That does not make them useless — it changes what you should ask.

A weighted scoring grid looks like objectivity and is mostly a mechanism for converting a preference into a number.

Familiarity with the current system and enthusiasm for the newest option pull in opposite directions and are both about comfort rather than fit.

A trial that cannot produce a no is not a trial. It is a familiarisation exercise with a purchase order attached.

Trial time is short. Spending it re-verifying that a mature product can do the obvious wastes the only chance to find what it cannot.

Sample data is clean, consistent and complete. Yours is none of those, and the difference is where implementations fail.

The evaluator is rarely the daily user, and the gap between their two experiences of the same product is enormous.

The most informative twenty minutes of any trial is the one the vendor has no incentive to make easy.

The hardest outcome to act on is a trial that produced no clear winner, which is the most common outcome.

The same product can be cheap or ruinous depending on how the meter runs against how your organisation actually behaves.

Subscriptions are compared monthly and lived with for years. The monthly figure is usually the smaller part of the total.

Most software contracts are signed unread. A small number of clauses determine what happens when things change.

A competitive first-year price is frequently an investment the vendor plans to recover at renewal, when you are least able to leave.

Migration and configuration are your costs, paid in your people's hours, and they do not appear in any proposal.

Price differences between candidates are usually small next to the differences in what they cost you to operate.

Most small organisations implement software as a side task. That is workable, but only if the sequence is right.

A single session before go-live is the standard approach and the one most reliably forgotten by the time it is needed.

The success criteria were written at the start precisely so this question could be answered. Almost nobody goes back and answers it.

Adoption rarely fails loudly. It fails as a slow drift back to the spreadsheet, and by the time it is visible it is established.

The half-decommissioned system is the most common artefact of a software migration and the most expensive to leave alone.

The review is booked, skipped, and forgotten. It is also the only mechanism by which an organisation gets better at buying software.