The most expensive way to save $25,000

What building your own website with AI really costs. And what nobody puts on the invoice.

I have some version of this conversation almost every week.

A founder tells me their hourly rate. Usually somewhere between $150 and $200. Then, almost in the same breath, they tell me they're going to build the website themselves. AI writes the copy now. AI designs the pages now. So why hand us $25,000 when they could sit down with Claude for a month and get it done for the price of a subscription?

I get it. I really do.

The tools are good enough now that a founder can produce something that looks like a website. Copy, layout, design, the whole thing, straight out of a chat window.

But here's the turn: good enough to build something is not the same as good enough to build the right something.

And the entire cost of this decision is hiding in the gap between those two sentences.

So let's run the numbers. Because I don't think anyone actually does.

What you're comparing against

Our basic package is a full rebuild. Home, pricing, four feature pages, a blog, a demo page, an about page. Custom design, custom AI-assisted code, a CMS wired in underneath. About a month, soup to nuts. $25,000.

Hold that number. Everything below gets measured against it.

What it costs to do it yourself

Say you take it seriously and try to do it right. Here's where the hours go:

Content strategy and information architecture: 15 to 20 hours
Copywriting, plus the back-and-forth with AI: 40 to 60 hours
Design direction, plus the back-and-forth with AI: 40 to 60 hours
QA, CMS setup, wiring it all together: 20 to 30 hours

115 to 170 hours.

Let that land. At 40 hours a week, the high end is more than four full weeks.

You didn't save a month. You spent one.

Now put a number on those hours, and let's be deliberately conservative about it. Not the $150 to $200 you'd bill a client. Call it a flat, hard-to-argue-with $100 an hour. That's still $11,500 to $17,000 of your own time. Before a single dollar on tokens. Before we get anywhere near the costs that never touch an invoice.

And $100 is lowballing you on purpose. Once you account for everything a founder's hour is actually worth, the expenses you don't carry, the perks, the leverage on your time, you're realistically closer to $150 or $175. Run the same hours at that rate and you land somewhere between $17,000 and $30,000. On time alone, doing it yourself already meets or beats what we charge.

And that estimate is generous, because it assumes you already have a system. Some sense of how to prompt, how to iterate, how to know when the copy is actually right and the design is actually done. Most founders don't. We've built proprietary skills at Conversion Factory precisely so our team isn't rediscovering the same prompting patterns from scratch on every project, and even with all that scaffolding, a full site still runs 1,200 to 2,000 prompts. Without it, learning what "good" even looks like as you go, you're realistically burning 1,800 to 4,000. Probably more.

Do the math on the tokens everyone's so worried about. Even at a generous three minutes of your attention per prompt, 3,000 prompts is 150 hours. Gone.

The tokens are basically free. Your calendar is the thing getting robbed.

This was never about the AI spend. It's about how many hours of a founder's life, worth $100 an hour if we're being stingy and a lot more if we're being honest, get quietly eaten by trial and error a trained team would never have to touch.

And all of that is before the part that actually matters.

The stuff that isn't on the invoice

Here's where founders miss the real math entirely.

Every hour on the website is an hour not spent running the company. That hourly number isn't just a price tag on your time, it's a stand-in for everything else that hour could have been. The sales calls that didn't happen. The features that didn't ship. The go-to-market that lands a month late because you were buried in Figma-by-way-of-Claude instead of out talking to customers. This is a second cost running right alongside the first one, not the same cost counted twice. You don't get to spend the same hour on two different scoreboards.

Then there's the pattern I've watched play out over and over. DIY sites without a system get rebuilt anyway. You build it, it's fine for a while, and twelve to eighteen months later you're right back in the market for a real rebuild. So the money and time you spent the first time wasn't a substitute for hiring a team. It was a down payment on hiring one later. Build it yourself now for $15,000 or so of your time, rebuild it properly in eighteen months for $25,000 more, and you've spent $40,000 to land exactly where hiring once would've put you on day one. A year and a half behind.

And then there's conversion. The one that compounds. The one that gets missed the most, because everything above is a one-time cost and this one absolutely is not.

I know what you're thinking. My DIY site will be fine.

Fine is the whole problem.

Say you're pulling 5,000 visitors a month and selling software at $30 a month. The gap between your DIY attempt and a site built by people who do this every single day doesn't have to be dramatic to matter. One point of conversion, 2% to 3%, is 50 more signups a month.

Fifty new subscribers at $30 is $1,500 in new recurring revenue. And here's the part that should keep you up at night: it isn't $1,500 once. It's $1,500 more, every month, and next month's fifty stack right on top of this month's. That's what recurring means.

By the end of year one, that one missing point of conversion is 600 subscribers you never signed and roughly $18,000 in monthly recurring revenue you're not collecting. About $216,000 a year in run-rate. And climbing.

Now put the two numbers next to each other.

Your savings is $11,500 to $17,000. Once.

The revenue you left on the table adds up to around $117,000 in the first year alone, and it never stops adding.

Your entire savings is erased in about four months. Then the gap just keeps widening, every single month, for as long as that site is up. By the time a DIY site gets torn down and rebuilt at eighteen months, you're north of a quarter million dollars in signups that quietly went somewhere else. Scale it to your real traffic, your real price, your real churn, the shape never changes. Most founders run the first number and never even glance at the second.

And underneath all of it sits a cost that shows up on no spreadsheet: decision fatigue. Doing unfamiliar creative work, context-switching in and out of it while trying to run an actual business, is a tax on your focus that never gets itemized anywhere. It's real. It's just not the kind of thing you can slap a dollar figure on, so people quietly pretend it isn't there.

So here's the actual comparison

$25,000. One month. A team that does this every day. A site built to convert from day one.

Or $11,500 to $17,000 of your own time if we lowball it, closer to $30,000 if we don't, a launch that slips, a business that got less of you during the month it needed you most, a rebuild inside a year and a half, and a compounding revenue leak that clears six figures in the first year and only widens from there.

One is a purchase. The other is a bet you keep paying to lose.

The cheapest quote in the room is almost never the cheapest option.

— Zach 😎