The only question that predicts failure
Before features, price, integrations or anything else, ask this: how many of my people have to open something they do not currently open? The answer predicts whether the tool is alive in three months better than every other factor together.
If the answer is "everybody", the odds are poor - not because your staff are difficult, but because opening a new app is unpaid extra work performed at the end of a tiring day by somebody who was not consulted about buying it. Some teams do it; desk-based teams with a manager enforcing it usually do. Technicians, kitchen staff and field engineers usually do not, and the list is stale within a month.
Every other question in this guide matters less than this one.
- Everybody must open a new app - expect decay unless somebody enforces it daily
- Only managers must open it - considerably better odds
- Nobody must open anything new - the constraint is removed
- Ask the vendor what their month-three usage looks like, and watch whether they answer
What each pricing model really costs
Software is sold four ways and the monthly figure is the least useful number in the comparison. What matters is what happens when you hire, and what the total is over the period you will actually use it.
| Model | Looks like | What it really costs | Watch for |
|---|---|---|---|
| Per seat, monthly | $5 - $30 a person | Rises with every hire, forever | Shared logins; seasonal teams paying year-round |
| Flat rate, monthly | $29 - $99 to a user cap | Predictable until you cross the cap | The per-user rate above the cap; feature tiering |
| One-time purchase | $1,000 - $5,000 once | Plus hosting and running costs you pay yourself | Whether updates and support are included, and for how long |
| Free tier | Nothing | Limits arrive exactly when it becomes useful | User caps, history limits, features moving to paid |
Do the three-year arithmetic for your actual team size, including the people you expect to hire. For a fifteen-person team, a $20-a-seat tool is $10,800 over three years; the same team on a $1,999 one-time purchase plus running costs is under $3,000. That gap is not a rounding error, and it is invisible if you only compare monthly rates.
Twelve questions to ask every vendor
Ask all twelve, including of us. The pattern of what a vendor avoids answering tells you more than the answers themselves.
The last three are the most revealing. A vendor who cannot name a single weakness or a single competitor who beats them somewhere is either not paying attention or not being straight with you.
About adoption
- How many of my staff have to install or log into something?
- What does a member of staff have to do every day, in seconds?
- What happens when somebody simply does not use it for a week?
About money
- What does this cost me over three years at my team size, including hires?
- What happens to the price when I add five people?
- Which of the things I just saw are on a higher tier?
About the data
- Where is my data stored, in which country, and who can read it?
- How do I get all of it out, in what format, without asking you?
- What happens to it if I stop paying, or if you are acquired?
About the truth
- What does your product do badly, or not at all?
- Who should buy a competitor instead of you?
- Can I speak to somebody using it who is not in your testimonials?
How to run a trial that tells you something
Most trials are run by the most organised person in the business, on the quietest fortnight of the year, on work that was already under control. They prove nothing except that the software functions.
Design the trial to fail if it is going to. That is the only way it is worth the fortnight.
- Pick your least enthusiastic member of staff, not your most organised one
- Run it through a normal-to-busy fortnight, never a quiet one
- Use real work, including the messy jobs nobody wants to write down
- Do not remind anybody to use it after day three - that is the thing you are measuring
- On day fourteen, ask one question: is the list true? Not "did people like it"
If the list is accurate on day fourteen without you having nagged, the tool works for your team. If it is not, no feature on the roadmap will fix it.
Hidden costs to price in
The licence is rarely the whole cost. Four things are routinely left out of the comparison, and two of them are usually larger than the subscription.
- Setup time. Somebody in your business spends days on this. That is real money, whoever the vendor is.
- Your own attention. Any tool needs somebody to look at it. If nobody will, buy nothing.
- Getting out later. Ask about export before you put three years of history in. "Contact support" is not an export path.
- Tier creep. Check which of the features you were shown are on the plan you are quoted, and which arrive at a higher one.
Where we fit, and where we do not
We sell one of these things, so here is the plain version rather than a conclusion dressed as advice.
Bosun is a one-time purchase, self-hosted, with no per-seat fee, and nobody on your staff installs anything because it works through WhatsApp. That combination suits businesses whose people are not at desks and who have already watched an app die on their team.
It is the wrong purchase if you need rotas or a time clock, if you want somebody else to run the server, if you need task dependencies and Gantt charts, or if your team is under five people and a group chat honestly still works. We have written those comparisons out with the competitor prices linked.
- Cheaper than per-seat tools from about the eighth month at fifteen seats - the arithmetic, or run your own numbers
- More expensive than Connecteam, which also does rotas - said plainly here
- Less capable than Trello or Asana as a project tool - and why that may not matter
- Not a customer inbox - that is a different tool
Run the twelve questions on us. If we dodge one, that is information you should act on.