Why We Fired Our Happiest Customers

Historical product and unit-economics clarification — October 11, 2026. The original October 2025 wording below does not establish that the app closed then. The author now dates the fintech sunset decision around December 2025 and the official pivot around June 2026, with two years for customer transition and the app still online; see the dated clarification. The cohort and financial figures are a personal report. The customer-comparison table records assumptions from that business, not measured traits of every contractor or knowledge worker. No willingness-to-pay experiment is reported. The later product/consulting descriptions are historical, not current offers; see About.

We had 254 active contractors.

They loved us. Real usage. Good retention. Genuine appreciation.

We fired them anyway.

The Decision That Didn't Make Sense

The original account placed the financing sunset in October 2025. My later clarification dates the decision around December 2025 and the official pivot around June 2026, with a two-year transition and the app still online. The October reference needs my records to distinguish an earlier discussion from the eventual decision.

Six years of building. 13 lender integrations. Tens of thousands of loan applications processed. 2,600 total signups across 20+ states.

And 254 contractors who genuinely loved using the product.

They weren't churning. They weren't complaining. They were happy.

We walked away from them.

Here's why.

The Difference Between "Liked" and "Need"

This is the lesson that took me six years to learn:

Happy customers and valuable customers are not the same thing.

Our contractors liked SimpleDirect. They appreciated the beautiful mobile app. They enjoyed having 13 lenders in one place. They told us the platform was "so much better than what everyone else offers."

But here's what they wouldn't do:

• Pay 10x more for it
• Refer other contractors
• Expand their usage beyond the basics

They liked us. They didn't need us.

And "liked" isn't a business. "Need" is.

The Math That Killed Us

Let me show you the unit economics that made this decision obvious:

Annual revenue per contractor: $1,500

Annual cost to serve:

• Customer support: $600-1,200
• Platform maintenance: $200
• Sales and marketing: $300
• General overhead: $200

Total cost: $1,300-1,900

Net margin: $200 to −$400 per year using the listed $1,300–1,900 costs and $1,500 revenue, before any omitted costs

At $1,500 per year, we couldn't afford to provide the level of service contractors actually needed.

Home financing is complex. Homeowners have questions. Contractors need hand-holding. That's the nature of the business.

We tried to solve it with better technology—a more intuitive app, faster pre-qualifications, automated workflows.

But the fundamental issue wasn't technology. It was that contractors expected white-glove service at self-service prices.

And honestly? They were right to expect it. The product category demands support.

We just couldn't afford to provide it at our price point.

The Customer Success Trap

A few years ago, we hired a customer success person. He was personable, contractors loved him, and he genuinely wanted to help.

But he had one fatal habit: he gave out his phone number to everyone.

Suddenly, every contractor expected phone support for every question. Every homeowner expected a call to "activate" their financing. What started as excellent customer service became an operational nightmare.

When we realized this, we made the hard decision to let him go. But the damage was done—the expectation was set.

Contractors were accustomed to calling us for everything. We couldn't scale that model without hiring a call center.

The lesson: You can't build a scalable SaaS business on a services business model. At least not at $1,500/year.

Our customers loved the high-touch experience. But that experience was killing us.

The Framework: How to Know If Your Customers Are "Liked" or "Need"

Here's what I wish I'd asked earlier:

1. Would they pay 10x more?

If your product disappeared tomorrow and came back at 10x the price, would they still buy?

Our contractors? No. They'd go back to calling lenders directly. It's annoying, but it works.

That's "liked." Not "need."

2. Do they actively refer others?

Not "would you recommend us?" (everyone says yes). Do they actually send you new customers without being asked?

Our contractors didn't. They were happy, but not evangelical.

3. Are they expanding usage?

Do they keep finding new ways to use your product? Do they ask for more features because they've hit limits?

Our contractors used the basics and stopped there. No expansion. No growth within accounts.

4. What happens when you raise prices?

We did not test those higher prices. My expectation that customers would leave is a hypothesis, not observed price sensitivity or an established market ceiling.

If you can't raise prices without losing everyone, you don't have pricing power. You have a charity.

Why Happy Customers Can Kill You

Here's the trap:

Happy customers feel like success. They say nice things. They don't churn. Your metrics look good.

But if they're only happy because you're undercharging and over-serving, you're building on sand.

Every month, you're losing money on them. Or breaking even. Or making margins so thin that one bad quarter kills you.

Meanwhile, you're not investing in finding the customers who would actually pay what you're worth.

Happy customers can be the most expensive customers you have.

Not because they're demanding. Because they're keeping you stuck in a market that doesn't work.

What Happened After We Fired Them

We pivoted SimpleDirect to a completely different market.

Old: Embedded financing for home improvement contractors; our selected cohort required support that was costly relative to our pricing.

New: AI-powered tools aimed at knowledge workers; lower support cost and higher willingness to pay were hypotheses to test, not established traits of this group.

Table context — October 11, 2026: The table below records my earlier generalizations, not measured customer cohorts. Its binary technology/support labels and willingness-to-pay amounts do not establish every contractor’s capabilities or the economics of every knowledge-work customer.

The difference:

Same team. Same company. Completely different economics.

We launched SimpleDirect Chat in beta. Customers are self-onboarding. Support tickets are minimal. We're actually making money on each customer.

I was concerned that satisfied customers could still be uneconomic for us. The stated annual cost range includes both a $200 surplus and a $400 deficit, so it does not establish a loss on every customer.

At publication I reported better economics for the new customers; an itemized cohort record would be needed to establish those margins.

That's the difference between "liked" and "need."

The Bigger Pattern: Why Some Markets Are Traps

I want to share something I've been thinking about since we made this decision:

Our contractor-financing model had difficult unit economics. This account does not establish that every product serving this market has the same economics.

Here's why:

• Low tech adoption — Contractors don't naturally embrace new software
• High support needs — They prefer phone calls over documentation
• Price sensitivity — They can't afford enterprise pricing
• Relationship-driven — Trust matters more than features
• Slow decision-making — Long sales cycles, low conversion

This article does not supply dated fee schedules or comparable margins to establish that these are the only viable embedded-finance models.

The middle market—where we were—is a death trap.

You're too expensive for DIY self-service. But too cheap to afford white-glove support.

We got out before burning more time and money on a model that doesn't scale.

The Permission Nobody Gives You

Here's what I wish someone had told me earlier:

You're allowed to fire customers who aren't working.

Not because they're bad people. Not because they don't appreciate you. Just because the math doesn't work.

254 happy contractors sounds like success.

254 happy contractors at negative margin is a slow death.

The hardest part wasn't the decision. It was giving myself permission to make it.

We'd invested six years. We had relationships. People depended on us.

But staying would have meant:

• Continuing to lose money on every customer
• Never having resources to invest in something better
• Eventually failing anyway, just slower and more painfully

Walking away wasn't giving up. It was growing up.

The Framework For Your Business

If you're wondering whether your customers are "liked" or "need," ask yourself:

1. Unit economics: Are you making money on each customer, or losing it?

2. Price elasticity: Could you raise prices 2x without losing everyone?

3. Referral behavior: Do they actively send you new customers?

4. Expansion revenue: Are they growing their usage over time?

5. Support ratio: How much does it cost to keep them happy?

If the answers are bad across the board, you don't have a customer problem.

You have a market problem.

And the solution isn't better customer success.

It's different customers.

The Real Lesson

We didn't fail with SimpleDirect Financing. We learned.

We learned that product quality doesn't matter if unit economics don't work.

We learned that beautiful design can't overcome fundamental business model flaws.

We learned that happy customers and profitable customers aren't the same thing.

And we learned that sometimes the smartest move is walking away from people who love you—so you can find people who need you.

That's not failure. That's clarity.

Have you ever had to fire customers who loved you? What made you finally do it?

Reply in the comments. This is one of the hardest decisions founders face, and I'm curious how others have handled it.