by John Parkins
Haunted Town Hall is a haunted attraction out in Lafayette, fifteen seasons in. Over the last four weeks we replaced almost everything they run on.
Here is what went away: an aging WordPress site, a third party ticketing vendor, a rented quiz widget that ran their trivia game, a texting service, and the pile of Zapier automations holding all of it together.
Here is what stayed: Mailchimp. Airtable. Twilio. Stripe.
That second list is the point of this entire post, and it took me a couple of years to understand why it looks the way it does.
Short version: the cost of building something shaped exactly like your business fell off a cliff, and most owners have not noticed yet. That does not mean build everything. It means the line between what you rent and what you own moved, and it is worth knowing where it sits now.
The 60 percent problem
Here is what I find in almost every shop I walk into around Allen County.
Four or five subscriptions. A scheduler. A CRM. An email tool. Something for invoices. Maybe a review tool a salesman talked them into on a Tuesday. Each one does about 60 percent of what the business actually needs. Not one of them was built for a two truck HVAC company in Northwest Ohio, because there is no version of anybody's product roadmap where that company is worth building for.
So the owner bends the business to fit the tool. Uses the notes field for something it was never meant to hold. Keeps a spreadsheet on the side for the part the software will not do. Pays per seat, every month, forever, for the privilege.
Then somebody sells them a sixth tool to make the first five talk to each other.
I wrote that sentence before I went back through my notes on the Haunted Town Hall project and remembered that their Zapier account was moving data between six different systems. It was not a metaphor. It was a line item.
So should you just build your own software?
No. Some of it, maybe, and probably not the parts you would guess.
Custom software just means a tool built to fit one business instead of a hundred thousand of them. That is the whole definition. It does not have to be big.
Look at that list of what survived. Mailchimp, Airtable, Twilio and Stripe all came out the other side of a four week teardown where nearly everything else got replaced, and it was not because I ran out of time.
What Mailchimp actually sells is a reputation with Gmail and Outlook, built up over years. I can write code that sends an email this afternoon. I cannot write code that lands in an inbox, because that part is not code. Twilio is phone infrastructure, carrier agreements and number pools, and the compliance paperwork alone will humble you. Stripe is money movement and the liability that comes with it, which I am delighted to let somebody else hold. Airtable is the interesting one, because I genuinely could build it. I do not, because it is a general purpose container, and because the owners can open it without calling me first.
That is the rule, and it took me a while to say it out loud. Rent the plumbing. Build the thing that is shaped like your business.
No vendor on earth is going to build a ticketing system shaped like a haunted attraction in Lafayette, Ohio. There are not enough haunted attractions in Lafayette, Ohio. But every vendor on earth will happily rent you email delivery, and they should, because they are better at it than you or I will ever be.
Most owners have this backwards. They rent the specific thing, at 60 percent fit, and never think about the plumbing at all.
How do you tell which is which?
Ask what you are actually paying for.
Keep renting when:
- The rules change every year and staying current is the actual product. Payroll and sales tax, mostly.
- You are buying somebody's reputation rather than their code. Email delivery is the clearest example.
- It runs on physical infrastructure. Phone numbers, card processing, shipping labels.
- Your team needs to open it without going through you.
- The vendor is carrying liability you do not want to hold.
Start asking harder questions when:
- You pay per seat for something four people touch once a week.
- You pay per transaction on volume you generated yourself, through your own door, with your own marketing.
- You have bent your process to fit the software instead of the other way around.
- You are paying a sixth tool to connect the first five.
That second one deserves a minute, because it is the one that hides. On most ticketing platforms the fee is a percentage plus a flat charge on every ticket, which sounds harmless until you run it on a cheap ticket. Eventbrite lists 3.7 percent plus $1.79 per ticket in service fees, with payment processing stacked on top.
If you are a touring act filling a room in a city you have never set foot in, that fee is buying you an audience and it is worth every penny. If you are a seasonal attraction that sells out because everybody in the county already knows about you, you are paying a finder's fee on customers you found yourself.
What I actually build with
Claude Code, mostly. I have written before about the AI tools I use every day, and this is the one that moved the math.
For years I ran Zapier, then Make, then n8n. Those are what you use when you can describe exactly what you want but cannot build it, so you drag boxes around a screen and connect them with lines. They are good tools and plenty of smart people run their businesses on them. I do not use them anymore. Between Claude Code and the newer standards for getting one piece of software to talk to another, I can build the connection instead of renting a middleman to hold two things together.
Worth knowing what the middlemen run, because it creeps up on people. Zapier's entry level paid plan is $29.99 a month for 750 tasks, and a task is one step, not one automation. A five step workflow that fires 150 times in a month eats the entire allowance.
The part I am not going to pretend about
I can do this because I spent 25 years in print and design before any of it, and because I have been buried in these tools since before most people knew they existed. If this post talks you into believing you will be shipping software by Friday, I have done you a disservice.
The honest framing is that the distance got short. The gap between "I need a thing that does exactly this" and "somebody built it" used to be six figures and most of a year. Now it is close enough that a small business can reach it.
Whether you cross that distance on your own or with somebody sitting next to you is a separate question, and the answer is different for everybody. That is where I have been spending my time lately. Less building custom software for hire, more sitting down with owners who want to know what is possible in their own shop and what is not worth the trouble. It is case by case. Some people need one afternoon and a nudge. Some need considerably more than that.
Where to start
Pull up your bank statement and find every software charge on it. Add them up. Then write down, next to each one, what you are actually buying. Reputation. Infrastructure. Compliance. Convenience. Or a shape somebody else designed for somebody else.
Most people find one line item that makes them a little angry. Start there.
If you want a second set of eyes on that list, that is a conversation we have all the time. BeyondVivid sits on West Elm here in Lima, and this kind of work lives with the Digital Squad. Fifteen minutes, no pitch, no deck: grab a spot on the calendar and bring the list.
