How do you test a local event without risking a public flop?

The scariest part of a local event is the empty room everyone sees. That fear kept me from running more events.

"If you put out there that you're doing this event and then you don't get any traction or you just change your mind, now you've got to go tell everybody you're not doing it and you've got to refund them if you charge money. I don't want to take that risk. So for me, we just don't do it."

An events operator gave me the fix, and it removes that risk completely.

What is the apply-to-join trick?

List the event on Luma as "apply to join" instead of open RSVP. Enough applications? Run it. Not enough, or three people? Delete it. Nobody was ever approved, nobody has it on their calendar, and there's no refund cleanup.

It also makes the event feel selective. I've applied to Luma events in Austin and caught myself wondering if I was good enough to get in. That feeling helps you.

I'd seen "request to join" as a user plenty of times and never connected it to testing demand. The same week I learned it, I had just canceled a free event that got almost no reaction after running in the newsletter and on social a couple of times.

Why not just poll your readers first?

Polls overstate interest. I ran one in my Round Rock newsletter about meeting other locals: about 200 people said they were interested, and the comments were energetic. Then I launched a dinner club there and got very little interest. A friend running a local dinner club elsewhere had to cancel and refund everyone.

The poll still taught me who wants events: moms, dads and business owners, plus the group I missed, seniors. Seniors were the ones writing in asking for events made for them.

Polls can work. Marissa from Boise told me on my podcast she'd decided to launch her dinner club if 100 readers said yes, and 650 did. She was upfront that it was an experiment. Apply-to-join just gives you a real signal instead of a hypothetical one.

What kinds of events should you test?

Lots of them. The advice I got was to try them all over time and see which ones stick. With apply-to-join, you can launch event ideas all day:

• A casual AI meetup
• A 30-minute talk on Meta ads for local businesses, then networking
• A business-owner dinner
• A meetup for parents, or for seniors

Even the ones you kill aren't wasted. You collect the email of everyone who applied, so you know who to invite next time.

For in-person events, smaller is often better. In a smaller community, word spreads faster and people are more likely to already be connected. In a big metro, people do their own thing.

Start small. My second AI meetup ran for 13 people with zero venue coordination. That bare-minimum version worked.

In local media there are a thousand good ideas. The problem is capacity. That's why I'm picky about events: I only want formats that are easy to run.

How do you de-risk the cost?

Get someone else to carry it. Two ways I think about it:

• Find a venue that wants your audience. I ran my AI training at a really nice senior living community for free. They knew my audience was mostly 35 to 55, people whose parents might be looking for a place to live. Win-win. Some businesses with space would even pay you to bring people in.
• Cap it and get a sponsor. An idea I've floated but haven't run: a free event at an indoor playground that normally charges about $20 a kid, capped at 50 kids in one corner section. A real estate agent already told us they'd pay for events. Worst case, nobody shows up and the venue lost nothing.

For a free business event, sell a sponsor table instead of tickets. A business bank or a vendor selling AI phone answering to small businesses would pay something like $500 for a table and a quick intro before the presentation. Ticket sales are possible once you've built trust, but they're hard. If you can avoid them, do.

What did a partner's first dinner teach us?

A partner of mine ran his first business-owner dinner and shared the honest breakdown:

• 18 people came. About 14 came from his personal network, one from the reader list, none from ads.
• A sponsor paid $750 for introductions, a materials table, a custom web page and a newsletter ad.
• The venue hurt it. No private room meant no cocktail hour, bad acoustics and no clean moment for the sponsor.

My take: don't blame the ads yet. Three weeks of lead time is the minimum before you judge whether paid ads work for events. One week is too tight for Meta to optimize.

What makes the format work:

• Get a private room. It makes the cocktail hour, the acoustics and the sponsor moment possible.
• Use name tags with a conversation prompt, not just a name.
• Borrow a structure. Nick Gray's The Two-Hour Cocktail Party is a solid playbook for exactly this.
• Sell a community, not a night. A one-off dinner is a ticket sale. The business is the ongoing membership. Membership communities like Hampton are the model to grow toward.

Free or paid?

For your very first event, free is probably the right call, because you don't know how it'll go. The trade-off: free brings out some people you wish weren't there. A higher-end local newsletter operator told me those people usually self-select out once you charge.

After that, test paid. One operator I know went upmarket with a $60 wine-tasting event instead of a free one, and it beat her regular dinner club on both revenue and how happy attendees were. Going upmarket can be the better event.

How do you target business owners for a dinner?

For a business-owner dinner, raising the Meta targeting from $250,000 businesses to $1 million-plus businesses is what brought in qualified people. A young operator in Georgia ran that version and got three qualified leads quickly, including a Shark Tank alum. Use manual business-owner interest targeting, age 35 and up, with Marketplace placement off.

Once a format works, the money side gets bigger: sponsor tables, sponsor-paid free events and dinner clubs. That's in how local newsletters make money from events, and why free events still pay off is in how free events build a local network.