Report · updated 2026-08-02

The State of Self-Hosted Software, 2026

We keep a running ledger of what business SaaS actually costs — every figure transcribed from the vendor's own pricing page, on a date, with the link. This report is what those 83 figures say when you line them up against the one-time price of owning the same tool. No market-size estimates, no analyst forecasts. Just the prices, and the arithmetic they imply.

83 pricing figures 57 with a public monthly price 26 quote-gated or unpriced 82 vendors $39 median monthly

Start with the single most useful number in the whole dataset: the median business-SaaS tool in our ledger costs $39 per month. Half the priced tools cost more, half cost less. The cheapest priced figure we recorded is $1.65 per user per month; the most expensive is $789 per month. That spread — from a couple of dollars to the better part of a thousand — is the first thing the data teaches: "SaaS pricing" is not one market but a dozen, and the honest comparison is always category by category, never a single headline figure.

The second thing the data teaches is subtler, and it is the reason this report exists. A monthly price is not a number you pay once. It is a number you pay every month, for as long as you use the tool, whether or not the tool changes. Over a three-year horizon — a conservative life for a business tool — the median $39/month becomes $1,404. The same job, done by software you buy once and run yourself, is a one-time price in the low tens or low hundreds of dollars. The gap between those two numbers, repeated across 57 priced tools, is the entire subject of this report.

A necessary caveat, stated up front because it governs everything below: this is a price ledger, not a total-cost-of-ownership study. Renting SaaS buys you managed hosting, updates, and support; owning software means you (or a host you choose) run it. Self-hosting has real costs — a server, patching, backups — that a monthly bill folds in. What the ledger measures is narrow and exact: the list price the vendor publishes, versus the one-time price of an owned equivalent. Where that comparison flatters owning, it is because the recurring meter, compounded over years, is genuinely large — not because the owned tool is free to run. We say so again in the method note, and we mean it.

A word on what the dataset is and is not. It is 83 pricing figures across 82 vendors, each a real line read off a real pricing page on a real date — not a survey, not a sample weighted to look representative, and not a projection of where prices are heading. We add vendors when we build or compare against them, so the ledger leans toward the categories a small business actually shops in: invoicing, booking, scheduling, property management, expenses, compliance, document tooling. That is a feature for a reader in those categories and a limit for anyone expecting a census of the whole software market. Read it as a well-kept price book for the tools ordinary businesses rent, not as a market map.

What makes 2026 worth a report rather than a spreadsheet is the shape the numbers now hold. With 57 published prices in hand, the distribution is wide enough to have a meaningful middle and long, instructive tails — and the split between priced and unpriced tools is large enough to be a story in itself. The sections that follow walk that shape from the middle outward: the median first, then the categories, then the extremes, then exactly how every figure was obtained so you can check any of them yourself.

The numbers, by category

Grouping the priced figures by the owned tool they map to gives a category-level view. For each category below, the median column is the middle monthly SaaS price among the vendors we priced in it; the one-time column is what the owned equivalent costs; and the three-year gap is the median monthly, run out to 36 months, minus that one-time price. Categories are sorted by that gap — widest first.

Category	Vendors priced	Median SaaS /mo	Own once	3-yr gap at median
Approva	1	$499	$39	$17,925
Supplia	1	$499	$59	$17,905
Specta	1	$499	$99	$17,865
Tokora	2	$410	FREE	$14,760
Extracta	2	$272	$44	$9,748
Cargora	2	$269	$49	$9,635
Restock	1	$199	$89	$7,075
Consigna	2	$149	$84	$5,280
Lendra	2	$107	$34	$3,800
Commissa	1	$105	$69	$3,711
Gymora	1	$89	$59	$3,145
Confida – Self-Hosted Whistleblowing Portal	1	$79	$119	$2,725
Revup	1	$75	$69	$2,631
Fixora	3	$72	$129	$2,463
Vendra	2	$69	$34	$2,450
Rentara	5	$62	$119	$2,113
Safora	2	$52	$44	$1,810
Nexura	1	$39	$49	$1,355
Votera	3	$29	$59	$985
Leavora	2	$27	$49	$932
Slotly	4	$23	$34	$794
Ledgira	2	$22	$49	$760
Clockora	2	$22	$49	$733
Assetora	1	$20	$39	$681
Invora	3	$19	$34	$650
Waiverly	2	$19	$39	$645
Rostera	2	$19	$59	$616
Secreta	2	$18	$44	$616
Privara	1	$9	$49	$275
Expensa	2	$5	$49	$113

Three patterns come out of that table, and each is worth naming honestly.

The steep end is document-and-data tooling

The widest gaps cluster in a specific place: tools that meter document processing, procurement, product-information management, and OCR/extraction — the categories where the SaaS median sits in the hundreds of dollars a month rather than the tens. These are usage-priced or high-tier platforms, and over three years the recurring figure dwarfs a one-time price by five figures. That is not a trick of the arithmetic; it is what a $200–$500/month meter does to a budget when you hold it next to a tool you buy once. If your work lives in one of these categories, the ledger is telling you plainly that the rent-versus-own decision is not close.

The cheap end still adds up — just slower

At the other extreme sit invoicing, scheduling, booking, and staff tooling, where the SaaS median is roughly $20–$30/month. Here the monthly number is small enough to feel painless, which is exactly the point: it is designed to. But $20/month is $720 over three years, and the per-user tools in this band multiply that by headcount. The cheap end is not cheap; it is patient.

The per-user multiplier is the quiet compounder

The pricing model matters more than the sticker. A flat monthly price is a known quantity; a per-seat price is a bet on your own growth, made in the vendor's favour. Several tools in the ledger are priced per user — scheduling and booking tools in the $5-per-seat range, and heavier operational tools well above it. The headline figure looks small because it is quoted for one seat. Run it out to a real team and the arithmetic changes character.

Take a $5-per-user scheduler, a figure squarely in our ledger's booking-and-staff band. For a solo operator it is $5/month — trivial. For a five-person team it is $25/month, or $900 over three years. For a ten-person team it is $50/month, or $1,800 — the same three-year total as a tool that flatly charges $50/month, reached one hire at a time. The point is not that per-user pricing is unfair; it is that the number you evaluate at signup is the smallest number you will ever pay, and it grows precisely as your business succeeds. An owned tool priced once, for unlimited staff, inverts that: the cost is fixed at the moment you can most afford it, and every subsequent hire is free.

This is why the ledger records a per-user baseline where one applies, rather than quoting the single-seat price and letting it flatter the vendor. A comparison that comes out honestly at three seats is a comparison you can trust at thirty.

The stack, not the line item, is where the money is

No business runs one tool. It runs a stack — invoicing here, scheduling there, a booking page, a place to track staff, something for expenses. Each individual subscription clears the "small enough to approve without thinking" bar, which is how the stack assembles itself unnoticed. But the median tool at $39/month is $1,404 over three years, and a modest five-tool stack at that median is $7,020 over the same period — before a single per-user multiplier or price increase. The decision that matters is rarely any one line item; it is the standing total of the whole stack, renewing quietly every month. Owning even two or three of those tools outright is what visibly bends the curve, because each one you buy once is a meter you switch off permanently.

Nearly a third of the market won't show you a price at all

Of the 83 figures in the ledger, 26 — close to a third — carry no public monthly number. Some are quote-gated enterprise tools that route you to a sales call; a few would not render a price to our automated read and are recorded honestly as "not readable." We do not estimate these, ever: an unpriced row stays unpriced. But the count itself is a finding. When a large share of a market declines to publish its price until it has you on the phone, opacity is part of the product, and the buyer's first cost is the time it takes to find out what anything costs.

Reading the table honestly

At the top of the table sits Approva's category, where the median rented tool runs to about $17,964 over three years against a one-time $39 owned equivalent — a $17,925 gap at the median, before the tails. At the bottom sits Expensa's category, where the rented median is low enough that the three-year gap is modest and the honest answer is "it depends" — a free tier or a cheap flat plan may genuinely be the right call. Both belong in the same report, because the useful conclusion is never "always own" or "always rent"; it is "here is the category, here is the number, decide with it in front of you."

One caution about the medians themselves: several categories rest on only a handful of priced vendors, so a single tool's pricing can swing the middle. Where a category shows just one or two priced rows, treat its median as an anchor, not a verdict — the more vendors behind a figure, the more weight it carries. The full per-vendor detail, so you can see exactly what each median is built from, is one click away in the Observatory.

The three largest gaps

These are the individual vendor rows where three years of the published price, minus the one-time price of the owned equivalent, is largest. They are the tails of the distribution, not the typical case — but they show how far the meter can run when a high monthly price meets a long horizon.

The honest reading of these three is not "SaaS is a rip-off" — it is that a high monthly price compounded over years is a genuinely large commitment, and that for a well-defined job an owned tool removes the meter entirely. The extremes are dramatic; the median is the number to plan around. A $39/month tool you keep for three years is $1,404 — not catastrophic, but not nothing, and entirely recurring. Multiply by the number of tools a small business actually runs, and the stack, not any single line item, is where the money goes.

What to do with this

The point of the ledger is not to declare a winner but to make the trade legible. The decision comes down to four questions, and the data answers the last one for you:

• Does the tool's job change quickly? If the category evolves fast and you want the vendor to keep up for you, renting buys that. If the job is stable — invoicing, scheduling, a rent ledger — the thing you are renting mostly stopped changing years ago.
• How much do you care where the data lives? Owning means your client list, revenue records, and rosters sit in a database you control, not on a vendor's server behind a login they can change. For some businesses that is a compliance point; for others it is simply peace of mind.
• What is your real operational appetite? Self-hosting is closer to running a WordPress site than to running a data centre, but it is not nothing. If you cannot give it any attention at all, a managed rental — or a hosted-but-owned arrangement — is the honest choice.
• What does the three-year number actually say? This is the one you do not have to guess at. Run the tool you are weighing through the calculator; it uses these same cited prices to turn a monthly figure into your own three-year total against a one-time price. If the gap is small, rent with a clear conscience. If it is a category from the top of the table, the number will not be close.

Nobody should own every tool, and nobody should rent every tool. The businesses that come out ahead are the ones that look at the standing monthly total, pick the two or three tools whose job is most stable and whose data matters most, and buy those outright — leaving the fast-moving remainder on a subscription where the managed service earns its meter. That is not an ideology; it is just reading the table. Every figure here is yours to check: the full dataset, with source links and verification dates for all 83 figures, is the Price Observatory, and it is free to cite.