For eighteen years, every client who hired me had already bought hosting from somebody else.
They’d buy a domain and hosting, usually a builder or a theme on top, sit with it for a few weeks, get nowhere, and go looking for someone to do it for them. That’s when they found me.
The host was making eight dollars a month. I charged six thousand for the build, and a few hundred a month to look after it after that.
I’m writing this for the people at the other end of that.
If you run a hosting company, that customer was yours first. You paid to acquire them and you onboarded them, and by the time they got to me you’d effectively lost them, even though they were probably still paying you.
The maths never worked in your favor
Whatever you spent winning that hosting account, it doesn’t come back in year one. Probably not year two either. That same person handed me six thousand in week one, and the only reason they came looking is that they’d given up on what they bought from you.
I used to think that was just how the industry worked. Hosts sell infrastructure, agencies sell the work, everyone stays in their lane. After a few years on this side of the business, on advisory boards and at CloudFest and in a lot of late bar conversations with people who run hosting companies, I don’t think that anymore.
Nineteen out of twenty never finish it on their own
I was shocked to find out how few small business owners finish a website on their own. The failure rate sits at around 95%, and you’d think with AI, completion rates would skyrocket! But then I realized the tool was never the blocker.
They open the builder, look at a blank homepage, and have no idea what the headline is supposed to say.
Some of that 95% will hire someone like me, but really most of them never finish.
So think about the ones who never finish, and what they look like on your books:
- No live site, so there’s nothing to renew for
- Nothing to upsell
- No reason to remember your name when the invoice lands
So they drift and eleven months later they see a cheaper offer and take it, and it gets logged as price sensitivity. You already paid to acquire that person. The only thing standing between them and a much bigger number on your books is a website nobody ever built.
Your support tickets are full of opportunities
If I were you, here’s what I’d go and look at this week.
Pull three months of support tickets. Read the actual tickets, not the SLA report. Tag every one that isn’t really a support request, but a request for work:
- “Can you just build this for me.”
- “My site’s been down since Thursday.”
- “I think I got hacked.”
- “Nobody can find us on Google.”
Every one of those is a customer who already pays you, and already trusts you enough to ask. They’re asking for something you have to turn away, so they go and find someone like me.
I’ve asked a lot of hosting people what that number looks like for them, and many of them hadn’t counted. The ones who do go and count it come back a bit annoyed.
Nobody fixed this because both options were bad
I want to be fair to the two options that have been on the table for a while.
Build a services team in-house. Developers, a lead, a salesperson, and the ops behind all of them. A second business inside your first one, with different margins and different ways to fail. From my experience talking to hosting companies who have tried this, it is very expensive and slow.
Hand it to a vendor. This is cheaper, but your brand is on work you don’t control. If the site’s late or ugly, the customer doesn’t blame the vendor they’ve never heard of; they blame you.
So most hosts did neither, and that was a defensible call. Services didn’t scale, because every project was bespoke and every project was hours, and hours have never had software margins.
That second one, the brand problem, is the objection I get asked about most. It’s the right one to ask, so I’ll come back to it.
Whose side I’m on now
Everything up to here I’ve written as the guy who took your customer. Before I tell you what I think has changed, you should know what I do these days, because it should change how you read the rest of this.
I run Atarim and I build the software that does agency work, and for hosting companies we run it under your brand. The point of it, for you, is to become the place your customers get everything, not just the place their site sits.
What changed: the work turned out to be repeatable
The work agencies do is far more repeatable than agencies like to admit:
- Back up the site before you touch it
- Run the updates
- Check the forms still submit
- Sweep the broken links
- Check in at 30 days, before the client starts wondering whether anyone’s doing anything
- Send a report, not just an invoice
Those got skipped because there were never enough hours in the week, and the hours that did exist went to whoever shouted loudest.
We’re not inventing anything. I’ve always thought about it as digitizing human experiences. You watch what good people already do well, you teach the machine to do it faster and at a scale no human team could hit, and the humans stay in the creative seat.
Here’s what that looks like from where you’re sitting. Our team does the building and the software does the parts that repeat.
You can sell it two ways, and it works either way.
Put it on your checkout, next to hosting, with a price on it, then the customer can buy it themselves.
Or don’t put a price on it. The customer clicks a button to ask about it, your sales team calls them and quotes them.
Whichever you already do, you don’t have to change anything.
From there it’s the same. Once they’ve paid, someone on the team gets in touch from your address, within the day, and the customer fills in a form. Once the first version is up, they get a link where they click on the thing they want changed instead of writing an email describing it. That link is Atarim, running under your name.
Three rounds of that, then the site goes live.
After that, care runs monthly, and every month they get a report showing what was done. Plenty of customers buy care on its own, without a build.
Now the money, because there are two numbers and they behave differently.
The build is the big one, and I’ll be straight with you, it’s a fraction of what I used to charge for the same work. But it happens once. We split it with you, and your share lands in the first month, on a customer you’d already paid to acquire and mostly written off. Put that number next to what it costs you to win an account. You know your own CAC better than I do.
The care plan is the one that changes your business. We build the model on around four in ten staying on one after launch. That’s monthly revenue landing against an account that was worth a hundred dollars a year to you before, and it’s still landing in year three. Troy Dean has been telling agencies this for years and he’s right about it. Nobody ever said it to hosts.
Which leaves the brand question
Every screen has your name on it. The emails come from your domain, and they never find out who did the work. Nothing goes to your customer without someone on our side checking it first. If that isn’t true, don’t do it. Not with us, not with anybody.
I spent eighteen years being paid well to serve customers somebody else had already acquired, already onboarded, and already lost. The revenue was sitting in your book the whole time.
Book a call with me and I’ll walk you through how this could work for your business.
