Skip to content
Lead magnet in preparation

TCO calculator +
PDF guide

Excel with live formulas to run your real case, plus an 22-page PDF guide with migration step-by-step. Three-year cost of open source vs Digisini SaaS.

No card required · Free download

Available now
📊

TCO Calculator
+ PDF Guide

Excel with live formulas
22-page migration guide
Free download

TCO calculator + guide to migrate
from open source to SaaS

Excel with live formulas to run your real case, plus an 22-page PDF guide with migration step-by-step. No marketing, just numbers.

Two resources in one download

📊

Excel TCO Calculator (6 sheets)

Live formulas, adjustable assumptions, bar chart "your cost per year".

  • Sheet 0 — Read me first
  • Sheet 1 — Your current situation (inputs)
  • Sheet 2 — Your real 3-year cost (open source)
  • Sheet 3 — Equivalent cost on Digisini
  • Sheet 4 — Your net saving (the punchline)
  • Sheet 5 — Tweak the assumptions (advanced)
  • Sheet 6 — Notes and methodology (sources: Hetzner 2026, ONS, Eurostat)
📄

PDF Migration Guide (22 pages)

For when the calculator says "migrate" and you want to know exactly how.

  • Chapter 1 — Real audit of your installation
  • Chapter 2 — Calculate your real TCO
  • Chapter 3 — When open source STILL is the answer
  • Chapter 4 — Migration step by step (with screenshots)
  • Chapter 5 — Mistakes almost everyone makes the first time
  • Chapter 6 — What to do when something doesn't fit

TCO numbers at a glance
what the calculator shows

These are reference figures. The calculator uses your real inputs.

ScenarioSelf-hosted Xibo (3 yr)Digisini (3 yr)Saving
Cafe with 2 screens~£3,876 + 129 h£570 + 0 h~£3,300 + 129 h
Small chain, 5 screens~£7,230 + 234 h£1,770 + 0 h~£5,500 + 234 h
30-screen network, in-house IT~£26,310£5,970~£20,000

Assumes sysadmin at £30/h, owner at £25/h, Hetzner CX22 VPS. Adjust everything in the calculator.

TCO Calculator + PDF Guide

Your Excel calculator and PDF guide to your inbox in seconds.

Drop your email and the 2 files arrive instantly. No newsletter bombardment. We email when we update market figures (1–2 times a year, max).

We handle your data per GDPR. Both files arrive by email within seconds.

While you wait — try it free.

The calculator will show you the numbers. But if you already know you want cloud, the Free plan is live now. One screen, no card, 5 minutes setup.

Start free

Free forever · No credit card · Upgrade anytime

No expensive hardwareLive in 5 minutesManage it from your phoneMulti-screenHuman supportFree plan foreverNo installersCancel anytime

Three real costs are deliberately absent
because putting a number on them would make the result look more certain than it is

A calculator that includes everything produces a single confident figure, and confidence is the wrong output here. These three are real, they affect the total, and they cannot be estimated honestly without knowing your situation — so they are named rather than guessed.

What is missing, and roughly which way each one pushes:

  • The cost of an outage. A screen dark for a week costs a restaurant something and a waiting room almost nothing. It can be large, it is entirely context-dependent, and any default value would be fiction.
  • The value of the time you get back. If a hosted tool removes two hours a month of maintenance, those hours are worth whatever your next best use of them is. That is a real gain and it is not ours to price.
  • The risk of the person leaving. On a self-hosted setup this is the single largest tail risk, and it is a probability rather than an amount. Multiplying a guess by a guess produces a number that looks like arithmetic.

What the calculator does include is only the checkable part: hardware, subscriptions, hosting, and hours at a rate you supply. That makes the output narrower than a full business case, and it also makes every line of it something you can defend to somebody else.

There is one consequence worth knowing before you read the result: because the three items above are excluded, a close result is not a tie. If the two options land within ten per cent of each other, the decision is being made by the things that are not in the table — and at that point the honest answer is to choose on the risk you would rather carry.

Run it twice with different hourly rates — the lowest you would accept and the highest. If the answer flips between them, you have learned that this is a decision about time and not about money.

Four of the five numbers are on an invoice
the fifth is how many hours this actually takes, and nobody has ever written it down

Hosting, licences, hardware and subscriptions are all documented somewhere. The time spent maintaining, fixing and updating is not, because nobody bills for their own hours. It is also, in almost every case, the largest single line in the result — which means the calculation depends most heavily on the one figure that has to be estimated.

Three ways to get a defensible number instead of a guess:

  • Count the last three incidents instead of a typical month. People remember incidents. Multiply by how often they happen and you have something built from events rather than impressions.
  • Ask the person who actually does it. The estimate from whoever chose the system is consistently lower than from whoever maintains it, and the second is the one you need.
  • Or measure forwards for four weeks. Slower, and it turns the weakest input into the strongest. Worth it when the decision is large.

What matters more than precision here is direction: this input is almost always underestimated, never the reverse. Forgotten evenings, the half hour before opening, the message answered on a Sunday. None of it feels like work at the time and all of it belongs in the total.

So the practical rule is unglamorous and holds up: take your honest estimate and check what happens at double. If the conclusion survives, you can stop refining. If it flips, you have found the one number worth spending a month measuring properly.

Before filling anything in, write down the three most recent times something broke and how long each took. That list is a better input than any average you can recall.