Unthought.

Vendor and platform selection

Feature lists converge between finalists; what separates them is renewal terms, data portability, and what happens after the invoice is paid.

Choosing a platform means choosing a relationship with a company whose incentives are not yours, and keeping it for longer than anyone expects at the time. The evaluation that matters concerns itself less with the feature list — which will be broadly similar between any two finalists — than with what happens after the first invoice is paid.

Requirements are written down before any demonstration is watched. A demo is engineered to make one tool look inevitable, and a criteria list assembled afterwards will, without anyone intending it, describe that tool. The order of those two activities does more to determine the outcome than the evaluation itself.

Essential and desirable are separated before scoring begins, and the essential list is kept deliberately short. A requirements document in which everything is mandatory cannot discriminate between options, which means the decision will be made on something else and recorded as though it were made on the list.

There are no referral fees, commissions, or reseller margins from any vendor named in a recommendation. This is structural, not a matter of good intentions: an advisor compensated by one of the options cannot produce a recommendation a reader is able to evaluate, however honest the advisor happens to be.

Data portability is tested rather than read about. A real export is pulled during the evaluation and inspected — its format, its completeness, whether historical records come with it — because documentation describes the export somebody intended to build.

The contract terms that decide the relationship are the renewal and the exit, not the price. Auto-renewal windows, notice periods, price escalation clauses, what happens to the data after termination, and how long it remains retrievable. These are negotiable at signing and immovable afterwards.

Vendor stability is assessed as a real risk: how long the company has operated, how it is funded, and whether the product is its main business or an adjacent line that could be discontinued without much affecting its revenue. The second kind gets discontinued.

References are requested from customers of similar size and shape, and the productive question is not whether they are satisfied. It is what they would do differently, what surprised them after signing, and how long implementation actually took against what was projected.

The evaluation is written down and kept, so that when circumstances change the decision can be revisited against what was known at the time rather than re-argued from memory by people who now recall having preferred something else.

Exclusions

What this does not cover.

  • Referral fees, commissions, or reseller margins from any vendor under consideration.
  • Evaluations where requirements were written after the demonstrations.
  • Portability assessed from documentation instead of from a real export.
Elsewhere in this area

Advisory