The 'Fancy Dinner' Effect: Why Customers Feel Ripped Off Even When They're Not

A sales leader at a high-growth startup sits across from a vendor at a steakhouse. The wine list is extensive, the appetizers arrive unbidden. Instead of feeling valued, she feels suspicious. Her mind runs a quick calculation: this dinner cost maybe $500. Her company pays this vendor $50,000 a year. The vendor can afford that dinner only because of her contract. She quietly resolves to review the invoice when she gets back to the office.

This scene isn't hypothetical. It's a composite of thousands of interactions our data picks up every quarter—and it's exactly what Jason Lemkin describes in his piece on SaaStr. Lemkin recounts taking a top customer to a fancy dinner after Dreamforce, only to have the customer joke, “This is a pretty fancy dinner. That probably means I’m overpaying.” The jab stung because, as Lemkin says, it wasn't a joke. The customer was reading signals. And the signal the dinner sent contradicted the real deal they were getting.

Lemkin's story is a founder's memoir, not a data set. But his intuition lines up with what we see across thousands of businesses. The perception of overpayment isn't random noise. It's a measurable phenomenon with clear triggers—and it shows up in churn rates, expansion velocity, and support ticket spikes. Let's put numbers on it.

The signals customers actually notice

Lemkin lists four signals that trigger the “I’m overpaying” feeling: mediocre support, unfair upcharges, a customer success team that only upsells, and AI agents that are too hard to deploy. Our data tracks all four. Here's how they show up in the wild:

Mediocre support – We're tracking 1,847 problems across 89 industries related to slow, unhelpful, or canned support responses. Average severity: 3.9 out of 5. That's not a minor annoyance. When a customer pays serious money and gets an auto-reply that ignores their actual question, they don't just get frustrated—they start questioning the entire contract. Lemkin is right: support is the most visible proof of whether a vendor respects the customer's time and money.

Unfair upcharges – There are 612 complaints in our system about unexpected fees, “SSO taxes,” API access charges, and other nickle-and-diming. Average severity: 4.1 out of 5. These hit hardest because they feel like punitive taxes on using the product the way it was meant to be used. Nobody wants to pay extra to log in securely. Every one of those fees tells the customer the vendor's pricing model is designed to extract value, not deliver it.

AI agent deployment pain – We're seeing 594 problems about AI agent training difficulty, up 37% in the last 90 days. This is the newest and fastest-growing trigger. With usage-based pricing for AI agents, every failed attempt, every retraining session, every human cleanup task becomes a visible line item on the bill. Customers don't just feel the dollar cost—they feel the time cost. And as Lemkin notes, with AI the value audit gets faster and more granular. A single frustrating week can undo months of goodwill.

Unpredictable billing – Our data adds a fifth signal Lemkin didn't emphasize: meter anxiety. We've logged 890 mentions of unpredictable billing in the last quarter alone. Usage-based pricing promises fairness, but it also creates fear of runaway costs. Customers hesitate to scale AI agents not just because they're hard to train, but because they can't predict next month's invoice. That fear colors every interaction, making even a reasonable price tag feel risky.

The SMB multiplier effect

Lemkin's story features a large enterprise customer. But our data shows the overpayment perception problem is even more acute for small and medium businesses. SMBs report 3.2x more value-audit problems per employee than enterprise customers. That makes sense: in a 20-person company, the founder personally approves every software expense. They feel every dollar. A $50/month tool that underdelivers creates more resentment than a $50,000 contract that mostly works does at a Fortune 500 company.

For founders selling to SMBs, this means the fancy dinner is doubly dangerous. You can't afford to send the wrong signal because the founder is already one bad support interaction away from churning. Every touchpoint—from the onboarding email to the help center response time—is a value signal. And SMB founders talk to each other. A single bad review or a tweet about hidden fees can poison your pipeline for months.

What the data suggests you do differently

Lemkin's post-dinner changes were smart: better support for top customers, fewer add-on charges, success teams that actually help. But how do you operationalize that at scale, especially with a lean team?

1. Kill the nickel-and-dime charges now. The 612 complaints about upcharges have an average severity of 4.1. That's the highest severity score among all categories we track. Customers hate paying extra for things they believe should be included. If your pricing page has a separate line for SSO or API access, you're actively breeding resentment. Bundle those costs into your base price. Yes, it lowers perceived ACV, but it raises perceived fairness—and fairness drives retention.

2. Make support a product feature, not a cost center. With 1,847 support problems at 3.9 severity, you can't afford to treat support as an afterthought. At minimum, answer fast (under 2 hours for paid tiers). Answer like a human who read the question. And give your support team the autonomy to solve problems without escalation. The fastest way to make a customer feel overcharged is to make them repeat themselves three times.

3. Attack AI deployment pain with onboarding, not docs. The 37% quarter-over-quarter increase in AI training complaints means customers are trying to adopt these tools and failing. The vendor response is often to point to documentation or offer paid onboarding. Don't. Build a guided, in-product setup that gets the agent working on a real task in the first session. Better yet, offer a white-glove deployment for new AI customers—not as an upsell, but as a standard part of the launch. The cost of that service is far less than the churn you'll prevent.

4. Give customers clear, predictable usage dashboards. The 890 mentions of unpredictable billing tell us customers want transparency. If you use usage-based pricing, show them real-time cost estimates, budget alerts, and the ability to cap usage. Make the bill feel like a tool, not a surprise. That builds trust and actually encourages more usage, not less.

The bigger picture: engineering value signals

Lemkin's essay is a reminder that customers are constantly auditing value—and they read every signal you send, intentional or not. A fancy dinner, a slow support response, a mysterious $14.99 line item—these aren't isolated events. They're data points in a running calculation the customer updates every single week.

Our data confirms the audit is real, and it's intensifying. The rise of AI agents and usage-based pricing means customers can now measure value with terrifying precision. They know exactly what each task costs. They know when the agent fails and a human has to take over. They know when they've spent three hours training a model that still doesn't work. Every one of those data points goes into the “am I overpaying?” column.

For founders, the answer isn't to skip the fancy dinners. It's to make sure the fancy dinner is the least important signal you send. Because in 2026, the customer isn't just reading your gestures. They're reading your product logs.

---

Data cited from PainSignal's tracking of software vendor problems across 89 industries. Statistics reflect the most recent 90-day period unless otherwise noted.