No link on this site currently earns us anything. We have not been approved by any affiliate program yet.

Per user, per device, or per server — the unit is not the same everywhere

Not every price in this market is per person. ZeroTier counts devices, Pritunl charges per server you run, and several per-user plans have a device allowance behind them.

Rows priced in a different unit are kept in this dataset and left out of the per-user rankings, because averaging them together would compare things that are not alike.

"$5 per user" and "$5 per device" look identical in a table and describe different bills. A five-person company with laptops and phones is ten devices. A three-person company running one server is one server.

The three units in this dataset

Per user. The common case: Twingate, Tailscale, NordLayer, GoodAccess, uTunnel, Firezone and others price by the person, sometimes with a device allowance included.

Per device or machine. ZeroTier's plans are built around a device count with a base fee, and NetBird prices additional machines beyond an allowance. The person is not the unit; the thing joining the network is.

Per server you run. Pritunl licenses the software you host yourself, priced per server per month. Your user count does not enter the price at all, and your hosting bill does.

Why this site does not blend them

Because the conversion is not a constant. Devices per person range from one to four depending on whether phones, tablets and a home machine are enrolled, and the ratio is a fact about your company that no published page knows.

If we picked a number — say two devices per person — every ZeroTier row would carry an assumption invented here, printed in the same type as figures that carry a URL and a date. That is the trade this site exists to refuse, and it is the same reason we never convert between currencies.

So the rankings are per-user rows only. The other rows appear on the vendor list and on their own pages with the unit named, and they are excluded from the ordering with the reason given.

How to compare across units yourself

You can do the conversion we will not, because you know the ratio.

per-device product: monthly cost = base fee + (devices you will enroll x per-device price) then divide by people, not devices, to compare with a per-user plan per-server product: monthly cost = license per server x servers + what hosting costs you then divide by people

The second one has a term nobody publishes: what running the server costs you in hosting and attention. A self-hosted license at a low monthly figure is not the same as a managed service at a higher one, and the difference is your time.

The device allowance hiding inside per-user plans

Even where the unit is the person, a device limit often sits behind it. Where a vendor publishes the included count, it is in the row. Where it does not, the field is empty rather than assumed.

This matters most for the smallest teams, for the same reason everything else does: a two-person company with four devices between them can cross an allowance that a fifty-person company would never notice per head.

What we do with the rows we cannot rank

Keep them, show them, and say why. Seven vendors here publish no price at all, four publish figures we could not read, and a handful price in a unit that is not the person. Every one of those is a fact about the market rather than a hole in the dataset, and each row is labeled with which it is — the vendors that publish nothing covers the first two groups.

The alternative — quietly dropping every row that does not fit the ranking — would produce a tidier site that misrepresented the market as more uniform than it is.

Who this is not for

If every person in your company uses exactly one device and you will never self-host anything, the unit question does not affect your shortlist and you can read the per-user rankings straight.

It is also not a recommendation against per-device or self-hosted pricing. For a company with many machines and few people, or with the skills to run its own server, those units can work out far cheaper than a per-user plan. They are simply not comparable on a single sorted column.

Questions people ask

Which unit is cheapest? It depends entirely on your ratio of devices to people, and on whether your time is free. That is why the arithmetic above is written out rather than resolved into a ranking.

Could you add a devices-per-user input to the calculator? That would mean printing figures that depend on a number the reader supplies, which is defensible — it is what the team-size input already does. It is not built yet, and if it is added it will be described as an input rather than a published price.

Do any vendors publish both units? Some publish a per-user price with an included device count, which is the closest thing to both. Where they do, both figures are in the row.

Sources


Written by Kaz (pen name). Figures are copied from providers' published pricing pages on the dates shown. We have not used these products. Corrections: contact form.