A punch time clock is a century old idea that is still sold new, and for some workplaces it remains the honest answer. It works without a network, costs little, needs no explanation, and nobody has to consent to anything. What it costs is arithmetic: the card is the record, and somebody has to total it.
Where a punch clock is still right
A small site with a stable crew and a bookkeeper who already totals cards by hand. Replacing that with a subscription service can add cost and administration for no gain anybody would notice. The limits are equally plain: a lost card is a lost week, one person can punch another's card without difficulty, and nothing is queryable later without leafing through paper.
The arithmetic is the running cost
Totalling cards for twenty people takes an evening every period, and it is exactly the kind of work where errors are invisible until somebody complains. Price that person's time into the comparison, because it is the real cost of a punch clock and it is the thing a card or fob reader removes for very little money.
The card reader middle path
A proximity card or fob reader produces an electronic record without reading anybody's body, costs little, clears a queue faster than a card and a slot, and needs no consent conversation. For most employers still running a punch clock this is the upgrade that makes sense, and biometrics only become proportionate once you have evidence that cards are being shared.
Questions people ask about punch time clock
Are paper punch cards acceptable as records?
Generally yes, provided hours can be produced when asked and retained for the required period. The practical risk is loss and legibility.
Can somebody punch a colleague's card?
Easily, and that is the known weakness. It is the one problem a card with photo capture or a biometric reader actually solves.
What is cheapest over five years?
Usually a card or fob reader, once somebody's time totalling paper is counted honestly.