Every hosting company publishes a pricing page. Almost none of them tell you how to pick a row on it. This guide is the sizing conversation we have with new clients, written down: the arithmetic that actually predicts what a low-traffic site needs, and the honest answer to when the least expensive plan is not a compromise but the correct size.
Start with the number nobody asks for
The useful question is not “how powerful is the server” but “how many people visit on your busiest day.” Suppose your site sees 400 visits on a strong day — a healthy figure for a local business, a studio, or a professional practice. If each visitor reads three pages, that is 1,200 page views. Spread across the fourteen waking hours when traffic actually arrives, it averages well under two page views per minute.
Traffic never spreads evenly, so engineers apply a peak factor: assume the busiest minute carries ten times the average. That is still under one page view every two seconds. A well-tuned node serves a cached page in a few tens of milliseconds, which means your busiest minute of the year occupies the machine for well under a second of actual work. This is why raw horsepower is the wrong axis for sites like these — on the traffic question alone, almost any plan fits. What separates the tiers is everything else.
What actually separates the tiers
Every Hostfluency plan runs on the same fleet: NVMe storage, low-density nodes kept deliberately under-filled, 24/7 monitoring with checks every sixty seconds. You do not buy a faster machine by paying more. What changes across Ember, Forge, and Foundry is the shape of the service around the machine — how many sites, how much database, and how much of the operation moves off your desk.
| Plan | Monthly | Database | Fits |
|---|---|---|---|
| Ember | $19 | 1 GB | One site, steady content, backups minded daily |
| Forge | $39 | 10 GB | One site, plus managed email & DNS and priority answers |
| Foundry | $79 | 100 GB | Up to three sites, Continuous Protection included |
When Ember is enough
Ember is the right answer more often than a pricing page would like to admit. It fits when three things are true: you run one site; its database sits comfortably under a gigabyte — a brochure site or portfolio typically lives between 100 and 500 megabytes; and losing at most a day of changes would be an annoyance rather than a loss, because daily off-node backups define your worst case.
If you update the site every week or two and money does not move through it, Ember is not the budget option. It is the correct size, and paying more would buy capability you would never use. We would rather tell you that on day one than discover you resent your invoice on day ninety.
Where Forge earns its keep
Forge is the tier most clients graduate to — not for capacity, but for coverage. The 10 GB database allowance matters if you run a busy blog with years of archives or a site with a heavy media catalog. But the practical difference is the layer of work it removes: email and DNS managed for you, priority on the answer queue, a monthly health report in plain words, and small content changes handled on request. If one misconfigured DNS record could cost you a week of missed email, Forge is cheaper than the mistake it prevents.
When Foundry is the honest answer
Foundry stops being a luxury the moment your site takes money or bookings between backups. Its 100 GB database allowance covers serious archives, and it carries up to three sites under one plan. But its defining feature is Continuous Protection, included: live replication of every change to a second node, and point-in-time restore that can roll a site back to 14:32 this afternoon rather than to last night. On Ember and Forge that protection is available as an add-on; on Foundry it is the default posture.
The two signals it is time to move up
- Your database is approaching its cap. Archives, orders, and media libraries grow quietly — the plan should grow before the cap pinches, not after.
- A lost afternoon has a price. When orders, bookings, or member activity happen between backups, point-in-time restore stops being optional. Move to Foundry, or add Continuous Protection to the plan you have.
- A third signal outranks both: your “site” has quietly become an application — logins, APIs, background jobs, sustained concurrency. That is a different conversation. Start at our applications page, or talk to us about Enterprise.
The pattern behind all of this: for low-traffic sites, plans are not about horsepower, because horsepower was never the constraint. They are about how much of the operation you want to think about — and what your worst Tuesday is allowed to cost.
Filed under sizing & plans · HF-G-001
Still a question? Ask us — same-day answers.