How do you choose phones for an NFC deployment?
Test the exact model with the exact tag on the exact surface before ordering either. A specification saying NFC tells you the hardware exists and nothing about range, antenna position or how much the housing costs you, which are the three that decide daily reliability.
What a specification does not say
Where the antenna is, which decides how easy the phone is to aim.
How strong the field is, which decides how much margin operators have on a marginal tag.
How much the housing or the intended case costs you, which can be most of that margin.
Whether the manufacturer's software restricts anything, such as background behaviour or NFC under power saving.
All four are established in ten minutes with a sample and none appear on a datasheet.
The test that matters
Use the actual tag, on the actual surface, in the actual place, with the case you intend to issue.
Have someone who will do the job perform it, at their pace, with gloves if they wear them.
Test the worst location in the deployment rather than a convenient one. A tag on a metal doorframe at knee height is the real requirement.
And test more than one unit, since a single sample can be unrepresentative in either direction.
Standardise the fleet
One model means one instruction, one antenna position, one case, one behaviour.
Two models mean operators who learn one gesture and fail with the other, and support conversations that begin by establishing which phone someone is holding.
Where a mixed fleet is unavoidable, record which models are in it and test each one. Undocumented variety is what makes a deployment feel unreliable.
Buy a small reserve of the same model at the same time, since a replacement bought a year later is frequently a different device with the same name.
Planning for replacement
Phones are replaced on a shorter cycle than tags, so the fleet will change while the tags stay.
That means the deployment has to tolerate a phone you have not tested yet, which argues for margin: larger tags, better placement, visible targets.
Write down what you tested and what the result was. When the next model arrives, that record turns a re evaluation into a comparison.
And keep one phone from the outgoing generation. It is the reference device for deciding whether a new problem is the phone or the tags.
Frequently asked questions
Is a more expensive phone better at NFC?
Not reliably. Antenna position and housing matter more than price, and both are established by testing rather than by specification.
Can I specify NFC in a tender?
State the outcome instead: an operator reading the deployment's worst tag reliably with gloves. That is testable and a bare NFC requirement is not.
How many units should I test?
More than one, since a single sample can mislead. A handful across the intended configuration is enough.
What if the model is discontinued?
Expect it, and buy a reserve at the same time. Design with enough margin that an untested successor is likely to work.