Type "customer support best practices" into Google and you'll find the same advice on 50 different blogs. Some of it is good. Some of it is actively harmful. And because it's repeated so often, people assume it's true.

Here's the advice that sounds smart, gets shared constantly, and makes your support worse.

"Always Exceed Customer Expectations"

This is the most popular support advice on the internet and the most damaging to sustainable support operations.

The problem: if you exceed expectations on every interaction, you've raised the baseline. What was "exceeding" last month is now "meeting" this month. You're on a treadmill. And you're teaching customers that the exceptional response is the standard one.

Ritz-Carlton gives every employee a $2,000 authorization to solve guest problems. That's great for a luxury hotel charging $800/night. It's absurd for a $29/month SaaS product. But the advice "exceed expectations" doesn't come with a qualifier about unit economics.

Better advice: consistently meet expectations with occasional moments of delight. A fast, accurate, friendly response every time. An unexpected gesture once in a while (proactive outreach, a small credit, a personal note). The consistency is what builds trust. The occasional delight is what builds stories. Trying to delight on every interaction builds burnout.

"The Customer Is Always Right"

This phrase dates to at least 1905, when the Boston Globe attributed it to department store magnate Marshall Field. Harry Gordon Selfridge and John Wanamaker popularized it further. It was a marketing slogan, not a management philosophy. It was designed to make shoppers feel welcome. It was never intended to mean that every customer demand is legitimate.

The customer is often wrong. They're wrong about how the feature works. They're wrong about what the policy says. They're wrong about what they were charged. And sometimes they're wrong in how they treat your agents.

"The customer is always right" taken literally means: refund every request regardless of merit, never push back on unreasonable demands, and tolerate abuse because the customer's feelings matter more than your agent's.

Better advice: the customer's experience is always real. Their frustration, confusion, and disappointment are genuine even when their facts are wrong. Validate the experience ("I can see why this is confusing") while correcting the facts ("The charge on March 1 was for your annual renewal, which was scheduled when you signed up").

"Reduce Ticket Volume at All Costs"

This advice drives companies to hide the contact button, add friction to the support flow, build chatbots that prevent human contact, and celebrate "deflection" even when it means customers gave up.

Ticket volume isn't inherently bad. A growing company with growing users should have growing support volume. The question isn't "how do we get fewer tickets?" It's "how do we handle tickets more efficiently?" and "which tickets represent problems we should fix?"

Some tickets are signals that your product is working: customers asking how to do advanced things, requesting integrations, reporting edge cases in complex use cases. Deflecting these tickets means losing product feedback and customer engagement.

Better advice: reduce unnecessary tickets (ones caused by bugs, confusing UX, or poor documentation) while welcoming necessary tickets (product feedback, complex use cases, genuine help requests). The distinction matters. Reducing all volume blurs it.

"First Response Time Is the Most Important Metric"

We covered this in detail in our "Stop Measuring FRT" post, but it bears repeating here because this advice appears in nearly every support tool's marketing.

First response time measures when you first said something, not when you first said something useful. Auto-replies game this metric trivially. And optimizing for FRT leads to agents sending quick, low-quality initial responses followed by longer, better follow-ups.

The metric customers care about: time to resolution. How long until their problem is actually solved? A 2-minute FRT with a 48-hour resolution is worse than a 2-hour FRT with a 2-hour resolution.

Better advice: measure time to resolution, messages per resolution, and customer effort. These metrics capture the complete experience. FRT is one input, not the primary output.

"You Need to Be On Every Channel"

Email, chat, phone, SMS, Twitter, Instagram, Facebook, TikTok, WhatsApp, Discord. The "omnichannel" advice says you need to be everywhere your customers are.

In practice, being on 8 channels with a 5-person team means being mediocre on all of them. You can't staff 8 channels adequately. Context gets lost between channels. Customers contact you on four channels for the same issue, creating duplicate work.

Better advice: be excellent on 2 channels (one async, one real-time). Add channels only when you can staff them well and connect them to a unified system so context follows the customer.

"Automate Everything You Can"

The counterpart to "reduce ticket volume." This advice leads to over-automation: chatbots handling emotional complaints, AI making promises it can't keep, and customers unable to reach a human for genuinely complex issues.

Automation should handle the repetitive, high-volume, low-complexity queries. Password resets. Order status. Business hours. FAQ answers. These are genuinely better automated because the customer gets a faster answer and the agent's time is freed.

But automation shouldn't handle complaints, emotional situations, edge cases, or high-value accounts. The marginal cost saving of automating a complaint ($5 saved on agent time) is wiped out by the churn risk ($500 in lost revenue) if the automation botches it.

Better advice: automate the boring stuff. Humanize the hard stuff. Supp's classification draws this line automatically: simple intents get auto-responses, complex intents get routed to humans with context. The technology exists to automate selectively. The advice to "automate everything" ignores the selection.

The Pattern

All of these bad advice pieces share something: they're oversimplifications that sound good in a tweet but fail in practice.

Good support advice is conditional. "Exceed expectations when the cost is low and the impact is high." "The customer's experience is valid, their demands aren't always." "Reduce tickets caused by product problems, not tickets caused by customer engagement." "FRT matters for chat, less for email; resolution quality matters everywhere."

Conditional advice is harder to fit on a slide. It doesn't go viral. But it works.