We sell a service that helps businesses get off expensive subscriptions. There's exactly one honest way to back that up: showing our own bill.

So here it is. No anonymous case study, no range, no "up to 70% savings". The lines below are on our own invoices.

Also read the third and fourth section. That's where it says why this isn't free money.

01 / the billWhat did we pay, and what do we pay now?

Five subscriptions out, four lines back in return.

As of 3 June 2026.

The biggest item didn't disappear because we bought it cheaper. It disappeared because the four functions in it (CRM, email, blog and site) are now four separate programs that each cost nothing.

02 / the contextSo does DataDrift run on €24 a month?

No. And that question belongs here, because without an answer the figure above is misleading.

What you see in section 1 is the set that's migrating, not our entire company bill. Company-wide, our fixed costs are roughly €812 a month. After the full migration we'll land at roughly €522.

The difference is in everything we deliberately don't run ourselves: business email, telephony, payments, the language model that does the thinking. Those are in section 4.

A savings figure without the rest of the bill is a sales trick, not a measurement.

So the honest summary is two-fold. On the migrating set we save €327 a month. Company-wide, after all steps, roughly €290 a month: the difference is small because the heaviest lines are exactly the lines we keep.

03 / the real priceWhat did that saving cost?

Weeks. Not hours.

The software is free. Installing, connecting, securing and keeping it running isn't, and that's exactly the item missing from every "save on your stack" post.

A few concrete examples of what actually happened. Outgoing mail from our blog worked half the time, because the mail server randomly switched between two network addresses and only one was registered. A routine job on the web server dropped a manually added piece of configuration and took the blog offline. Neither is in any manual.

So don't calculate with the subscription price, but with this sum:

• What does the subscription cost per year?
• How many hours does it take to get the alternative working, and what's your hour worth?
• How many hours a year does the maintenance cost after that? Expect more than zero.
• What does a day of downtime cost you?

For us that sum comes out positive, for two reasons that don't apply to everyone: this is our trade, and it's our shop window at the same time. We'd do this even if it broke even.

04 / what we don't do ourselvesWhy is there still €522 on the bill?

Because running some things yourself is unwise, even if you can.

The pattern: we run what is software ourselves, and buy in what is a service with an obligation to the outside world. That line is useful, even if you take nothing else from this post.

05 / the honestyWhat's not quite right here yet?

Three things, and you'd rather hear them here than discover them yourself.

The migration has been decided, not completed. Right now two of our own servers are running: one for the blog, one for the community. The third (the server where the CRM and automations will land) has been decided but not yet ordered. So the €24 is a calculated end state, not a statement.

The number of programs is counted, not estimated. Ten of the components in our stack are open source that we run ourselves; the whole totals twenty-two components. We mention those figures because you can check them on our stack page. A round number nobody can verify isn't an argument.

We're not you. This sum works for us because we already have the administrator in-house. If you don't, then "free software" buys you an extra job. Then €297 a month for something that just works isn't a foolish expense, but a sensible one.

This text was produced with AI assistance and checked and approved for publication by a human.