He kept a text file with his own DNS checklist. MX, CNAME, TXT, SPF, DKIM, in the order that had burned him least over the years. Running a small web agency meant he was the de facto IT person for every client, and the DNS step was where onboarding always slowed down. First he had to recover the login for a registrar nobody at the client could remember signing up for. Then he had to find the DNS panel, which looked completely different at every registrar he touched.
He was good at it. That is the part worth sitting with. He was not fumbling. He had a system, and the system worked. It just cost him time, every single client, every single time.
The tax nobody put on the invoice
The time was one thing. The error mode was worse. Enter a record by hand and a single typo does not announce itself. You save the panel, it looks fine, and then you wait for propagation to find out whether you were right. If you were wrong, you fix the character and wait again. A five second mistake could cost an hour of dead time.
The quirks were real, and no amount of experience made them go away. None of this is hard in the way real engineering is hard. It is hard in the way a hundred small, undocumented differences are hard. You cannot reason your way through them. You can only remember them, and remembering is exactly the kind of work that does not scale.
| The quirk | Why it bites | With one-click connect |
|---|---|---|
| Trailing dot on a CNAME target | One registrar needs it, another breaks if you add it | The record is written correctly for that provider, so there is no dot to second-guess |
| TTL field | A dropdown at one registrar, a plain number field at the next | The value is set for you, never typed into the wrong kind of field |
| The menus move | A memorized layout gets reorganized the following year, and the Add Record button hides | There is no panel to navigate, so the layout never matters |
Thirty seconds, and a different feeling
The first time he ran a provider-hosted one-click connect for a client, he was ready for it to be a slightly faster version of the same job. It was not. A consent screen appeared, hosted at the client's own provider. It listed exactly which records would be created. The client clicked once to approve. That was it. No panel to navigate. No field to fill. No trailing dot to second-guess. About thirty seconds, start to finish. If you want to see how a one-click connection works, you can watch a client connect their domain in one click.
- Recover the forgotten registrar login
- Find the DNS panel, different at every registrar
- Type MX, CNAME, TXT, SPF, DKIM by hand
- Remember the trailing-dot and TTL quirks
- Wait for propagation to learn if a typo slipped in
- The client sees a consent screen at their own provider
- It lists exactly which records will be created
- The client clicks once to approve
- The records are written correctly for that provider
- Done in about thirty seconds
He told me the strange part was not the speed. It was the realization that came right behind it. The speed was welcome. The realization was categorical.
All the accumulated knowledge, the checklist, the per-registrar quirks, was solving a problem the new flow does not have.
This is the moment that stings a little for anyone who has earned a manual skill. His expertise was not being sped up. For every registrar that supported the one-click flow, it was being routed around. The consent screen did not need him to know that this registrar wants a trailing dot, because the records were written correctly for that provider without a human transcribing anything. He had spent years becoming a good translator between the SaaS tool and the client's DNS panel. The one-click flow removed the need for a translator.
The calls that stopped coming
Here is how he knew it was durable. The my-website-is-not-loading calls stopped, for the clients connected this way. Not because those clients got smarter or more careful. Because the failure mode that produced those calls, the midnight typo that nobody catches until the domain is dark, simply cannot happen when no human types the record.
Why the calls stopped
The 11pm typo cannot happen when no human types the record. Remove the manual step and you remove an entire class of support ticket, not just some minutes.
He is not out of work. Nobody who is genuinely useful to their clients is put out of work by a DNS step getting easier. The recovered time went somewhere better. It went to the parts of the job that actually require judgment: the choices about structure, the trade-offs a client cannot see, the work you cannot reduce to a checklist. That is the work he wanted to be doing anyway. The DNS step was never the point. It was the tax he paid to get to the point.
Manual setup has not vanished from his life. Some registrars have not adopted the one-click flow yet, and for those he still opens the panel, still remembers the quirks, still waits for propagation. He keeps the checklist. But it is shrinking, one registrar at a time, and he is not sad about it.
What changes when the step disappears
For a SaaS platform, the lesson generalizes past one agency. Every one of your customers has a version of this person: someone who has to get a domain connected, who is either an expert paying the time tax or a novice about to abandon at the DNS step. When the record entry disappears, you do not just save them minutes. You remove an entire class of support ticket, the one that comes in at 11pm and reads my site is down and turns out to be a typo. The expertise stops being the bottleneck, and the people who had it get their afternoons back. You can give every customer that thirty-second connect without asking any of them to open a DNS panel.
Custom domains, on autopilot
Let your customers connect their own domain in one click. We detect their DNS provider, write the records, verify ownership, and issue HTTPS, while they just approve one screen. No checklist, no trailing dots, no midnight typo.
Get started