Before you add another piece of software, another AI model, or another agent to your business — watch this.

The problem: New business owners get told they need this tool and that tool and this agent. Next thing they know, software is 60% of monthly revenue. Scott has seen it firsthand. That’s not sustainable, especially when you should be protecting runway.

Where Scott learned the principle — and it wasn’t from technology.

About 20 years ago, Scott’s dad — a photographer in his second career — invited him to a weekend workshop in Cincinnati with David Ziser, an internationally in-demand wedding photographer. Ziser’s images are iconic. Twenty years later they still hold up.

Scott watched him work: setting the scene, posing the subject, finding the light. Then the secret — Ziser would put a flash behind the subject, firing back toward the camera, so it lit up the veil, the train, and the white of the dress. All light. Nothing to do with the gear.

Then Scott saw the gear. Cases of lenses. Multiple $5,000 camera bodies. Tens of thousands of dollars of equipment. Scott had a cheap camera and started to discount the whole thing — I can’t replicate this.

Ziser stopped him:

“You shouldn’t go out and buy all this equipment. You don’t need it. I would not invest in any equipment, any lenses, any camera bodies, until you know with confidence that you are missing shots that you could get.”

And then the part Scott never forgot:

“Everything I’ve ever bought has not come from this desire to buy this tool. It’s come from frustration from not being able to do what I wanted to do.”

When Ziser hit a wall, he bought the cheapest thing that solved the problem. His equipment grew with his capability — not ahead of it.

Applied to software:

Earn it. The tool has to be justified by the business, not by what someone else in your industry uses.

Keep it in budget. Don’t let software outpace what’s reasonable, especially early.

Don’t grow into it. Grow, then earn it. No business ever grew because of the software.

People on task are often cheaper. Transactional contractors cost less than a stack of subscriptions, and they don’t become a fixed expense.

On the agentic-company videos: If your intent is to build an agentic business and you have the capital, go do it. But if you’re early and you’re not there yet, the tech becomes a boat anchor on your expenses at exactly the wrong moment.

The mindset shift: The question isn’t “should I add this tool?” It’s “what constraint am I hitting consistently that this tool would solve?”

That’s thinking first. Not tool-first.

Got a business question? Ask Scott here: scotttodd.net/ask

Mall developer Alfred Taubman figured out that you can build the best mall in the world, but if people won’t walk into the stores, it flops. He called the problem threshold resistance — if getting in the door takes too much work, people just don’t go.

The same thing happens with your call to action.

The diagnosis: Most CTAs ask for too much. We all want the sale today, because sales pay the bills. But if the ask is bigger than what the buyer is ready to commit to, they walk.

Match the ask to the commitment level. A book might be an impulse buy. A property isn’t — even when the monthly payment is low, the commitment is high. “Buy now” on a high-commitment purchase puts the brakes on.

Make the next step a micro step. What’s the smallest thing they won’t say no to? Request information. Download a property report. Tell me about your situation.

The attorney example: “Book a 30-minute consultation” sounds inviting, but with legal or medical the result might be something you don’t want to hear. That makes it scary. “Tell me a little about your case” or “download this” is a smaller door.

Honor the micro-commitment. Once they take the small step, don’t immediately jump to the ask. They’re not ready yet.

How to test it with AI:

Take two CTAs you’re considering — A and B. Give your AI model the full context: whether the buyer is cold or already in your world, what the offer is, the pricing, everything. Let it ask you questions. Then have it simulate a thousand prospective buyers and compare response rates between the two.

The warning: The numbers will be wrong. The model is wildly optimistic — it’ll tell you cold traffic converts at 10-15%. It doesn’t. Scott has had to walk it back every single time, and it agrees every single time.

Use it for comparison, not prediction. You’re not looking for a forecast. You’re checking whether you’ve lowered the resistance enough, or whether you’ve picked the highest-resistance option without realizing it.

Got a business question? Ask Scott here: scotttodd.net/ask