Everything on this site converges here. Whatever class was chosen and whichever product is in front of you, a short list of questions put in writing before signature will settle more than any demonstration.
The eight questions
The full list of what is captured, per feature rather than per plan, in writing. The ladder in the endpoint capture article is the checklist to hold it against.
Whether window titles are included, asked specifically, because it is the item most often described as application usage and it is the content of the work.
How one person's data is exported, demonstrated rather than asserted, without exposing anybody else in the same file. Regimes that grant rights of access make this a legal question and a procurement one at the same time.
Where the data physically sits, and which sub-processors are involved, because the answer determines which rules apply.
Who at the supplier can read it, including support staff during a ticket, and whether their access is logged.
What happens to it at the end of the contract: what is returned, what is deleted, on what timescale, and how deletion is evidenced.
The retention controls, demonstrated, including whether raw capture and aggregates can be given different lifetimes, which the retention article argues is the decision that matters most.
The price at three times current usage, since the quoted figure describes today and the invoice will describe growth.
The demonstration is designed, and the pilot is self-selecting
A demonstration is run by somebody who runs demonstrations, on data chosen to suit the product, with the alerting configured by the person showing it to you. Nothing dishonest is happening and nothing about it predicts a Tuesday in your organisation.
A pilot is better and carries its own distortion: pilots are run by the people who wanted the tool, on teams that agreed to it, during a period when everyone is paying attention. The result systematically overstates how the full deployment will go, and the correction is to pick the pilot team for being ordinary rather than for being willing.
The one test with no such bias is installing it on a machine that holds nothing, using it for a day, and reading the store afterwards.
Ask in writing, before signature
- Best forEvery purchase in either market
- PricingFree, and delays the decision by a week
- StandoutA supplier who will not put an answer in writing has answered the question
- Watch out forIt requires somebody to actually read the response, which is where this usually fails
Run a scoped pilot on real work
- Best forAny deployment large enough to be hard to reverse
- PricingUsually free or heavily discounted, and it costs the pilot team's attention
- StandoutReveals the operational reality that no demonstration shows, including the default alert volume
- Watch out forPilots are run by enthusiasts on willing teams, so they systematically overstate how it will go
Deploy on a test machine and inspect the store
- Best forWorkforce tools specifically
- PricingFree
- StandoutThe only method that establishes what is actually collected rather than what is documented
- Watch out forIt takes a day, and nobody schedules it before a purchase decision
Get the exit path in the contract
- Best forAnything you will still be running in three years
- PricingNegotiation time
- StandoutExport formats, notice periods and deletion evidence agreed while you still have leverage
- Watch out forLeverage exists only before signature, and nobody feels like negotiating an ending at the start
The two questions behind the eight
What decision will this let us make that we cannot make today. If nobody can answer that in a sentence, the purchase is a proxy for a different conversation, and the article on proxies applies to the purchase as much as to the measurement.
And what would have to be true for us to switch it off. An arrangement with no exit condition has none, and the tool will never raise the question, because its own cost is not among the things it collects.
The last thing to check is the one nobody checks
Whether the people affected have been told, in terms they could understand, before it starts rather than after. In the workforce market that is an obligation in most places. In the infrastructure market it is the difference between a team that maintains the alerting and a team that silences it.
Both failures are the same failure, and both are cheaper to avoid than to repair.
What we cannot verify
Contract terms, capabilities and pricing are specific to each supplier and each negotiation, and nothing here is legal or procurement advice. We have tested no product, name none, publish no scores, and reproduce no rates. The only account of a specific tool worth relying on is the one you produce yourself, from your own test, in writing.
The short version
- Eight questions in writing before signature settle more than any demonstration.
- Ask specifically about window titles; it is the item most often understated.
- Have the export of one person's data demonstrated, not asserted.
- Ask the price at three times current usage, because the quote describes today.
- Name the decision this enables, and the condition under which it stops.
- Tell the people affected before it starts, in both markets, for the same reason.