Buyer's Guide
The evaluation toolkit for technical buyers preparing the case to the CFO and IT.
A single bundle of the documents committee evaluation actually needs: a defensible TCO, a real timeline, the pre-answered security questionnaire, a sample RFP, and an integration plan with rollback.
THE SCORING RUBRIC
Seven questions, and what a straight answer sounds like
Every demo works. These are the questions that separate a platform your team still trusts in month six from one they quietly stop opening. Ask all seven, and score the answers rather than the slides.
- 01
Show me how this specific score was calculated.
A straight answer
A published formula and a per-event deduction table you can recompute on paper. A vendor who cannot will describe the model family instead, which is not the same thing.
- 02
How long until the first useful alert, and what happens before that?
A straight answer
A number in days, plus what the system does during the learning window. Silence while a baseline builds is fine if it is stated. Implied day-one accuracy is a demo, not a deployment.
- 03
What is your false positive rate?
A straight answer
Precision at your fleet size, not an accuracy percentage. On rare events accuracy is close to meaningless. Ask how many alerts land per day and how many of them are real.
- 04
Who holds the key to your audit log?
A straight answer
The signing key belongs outside the database. If whoever can edit a row can also re-sign it, the log proves nothing about tampering.
- 05
What happens when the network drops?
A straight answer
How long the edge buffer survives, in days, and whether readings backfill in order once the link returns. Ask about day ten, not day one.
- 06
Can you write to my PLCs?
A straight answer
Read-only is the safe answer. A platform that can write to your controllers is a safety and liability conversation with your process engineers, not a feature bullet.
- 07
What does leaving look like?
A straight answer
Export format, export scope, and how long access lasts after cancellation. If answering needs a call with the account team, that is the answer.
BUILDING THE CASE
What a TCO model has to include
Most vendor TCO sheets count the licence and stop. A model your CFO will accept counts what the deployment costs your own team, and puts your downtime figure on the other side of the equation.
- Software licence
- Per machine, per year. Ask what the rate does as the fleet grows and whether it steps at thresholds you are about to cross.
- Edge hardware
- One Linux box per site or per network segment. Often hardware you already have. Count it anyway, so the model survives someone checking it.
- IT effort during onboarding
- Hours from your own team for firewall rules, credentials and the CMMS integration. This is the line most vendor models leave out, and the one your IT lead will ask about first.
- Integration cost
- CMMS, ERP or BI connections. One-off, unless the vendor charges per connector, in which case ask for the list before you sign.
- Avoided downtime
- Hours avoided multiplied by the fully loaded cost of an hour down at your plant. Use your own figure. A vendor supplying both sides of a savings calculation is marking their own homework.
Three-year TCO = (licence + hardware + integration + internal hours) − avoided downtime cost
If a vendor will not run that subtraction with your numbers rather than their benchmarks, the model belongs to them and not to you.
WHAT IS IN THE BUNDLE
Five documents in one file
Total cost of ownership framework
Three-year and five-year TCO that adds up software licence, edge-agent hardware, IT effort during onboarding, and integration cost against the cost of unplanned downtime you avoid. A spreadsheet with the formulas exposed, not a marketing PDF.
Implementation timeline reference
The typical pilot-to-production schedule for a single-site deployment and a multi-site rollout, with the gates between each phase named and the people involved on both sides identified.
Pre-answered security and compliance questionnaire
SIG Core and CAIQ Lite answered in advance against the current Haltless control framework, so procurement onboarding does not begin with a blank questionnaire.
Sample RFP template
An adaptable request-for-proposal template covering data architecture, integrations, security, support, pricing model, and exit terms. Drop your own logo and edit the parts that are not relevant to your environment.
Integration and rollback plan checklist
A line-by-line checklist for the change-management ticket: pre-deployment validation, cutover steps, rollback triggers, and the data-export plan that keeps you in control of your own telemetry from day one.
HOW TO GET IT
Tell us where to send it
The bundle is not gated by a sales call. Send us your work email and we will reply with the file plus a single follow-up to answer questions specific to your stack. No marketing drip.
Email us for the toolkitCOMMON QUESTIONS
How do I compare predictive maintenance vendors?
Score the answers to a fixed set of questions rather than the demo. The seven on this page cover explainability, time to value, alert quality, audit integrity, offline behaviour, write access and exit terms. Ask every vendor the same seven and compare the answers side by side.
What should a predictive maintenance TCO model include?
Licence, edge hardware, your own IT hours during onboarding, and integration cost on one side. Avoided downtime, priced with your plant figure rather than the vendor benchmark, on the other. Internal hours are the line most models omit.
Do I have to talk to sales to get the buyer’s guide?
No. The rubric and the TCO framework are on this page in full. The bundle adds the file versions: a TCO spreadsheet with the formulas exposed, the pre-answered SIG Core and CAIQ Lite, a sample RFP and a rollback checklist. Email info@haltless.io and it comes back as a file, with one follow-up and no drip sequence.
What is a pre-answered security questionnaire?
SIG Core and CAIQ Lite are the two questionnaires most industrial procurement teams send to a new vendor. Answering them in advance against the current control framework removes a round trip from onboarding, which is otherwise the slowest part of a technical evaluation.
