The other day I read a customer experience report from my company that said:
“Customer satisfaction score: 87%.”
My reaction:
It looked great. Really. But I kept thinking: how many of those 87% would actually renew? How many are just polite? How many gave us a 4/5 because they couldn’t be bothered to write a comment?
Because between a customer saying “I’m satisfied” and a customer who is actually happy — there is a galaxy of difference. And the problem is: most dashboards don’t see the difference. Neither does your NPS score.
The silent churn problem nobody talks about
Here’s the stat that should make every B2B team uncomfortable.
25 out of 26 dissatisfied customers never complain. They leave. Quietly. Without a ticket, without a bad review, without a negative NPS response. They just stop renewing. (Source: thinkJar research via Armatis)
According to a 2024 Gartner Customer Experience Survey, 43% of customers who churn do so without ever voicing their concerns. And 56% of B2B companies only discover dissatisfaction after the churn event — when it’s already too late. (Source: C-Tribe Society citing Gartner)
This is the fundamental problem with NPS as your primary measure of happy customers. NPS captures the people who responded. It systematically misses the people who have already mentally left.
User research expert Jared M. Spool puts the average B2B NPS survey response rate at 4–7%. Even generous estimates put it at 12–13%. (Source: Itamar Gilad) And crucially, as Rob Markey — one of the official co-leads of the NPS framework — has confirmed, the most likely responders are Promoters. The least likely? Detractors. Which means your NPS score is structurally biased toward making you feel better than you should.
Your KPIs are green. Your NPS is stable. And a portion of your customer base has already decided to move to a competitor.
Why scores aren’t enough to spot a happy customer
Let me give you an example from real life.
In a company I worked for, a customer once gave us 10/10 on a CSAT survey. People were proud. Until I read the comment:
“Thank you for replying after 6 days of silence. Finally.”
TTechnically a perfect score. Emotionally, a thinly veiled indictment.
That’s the problem with treating NPS, CSAT, and CES as complete pictures of a happy customer. They’re indicators of behavior, not emotion. They tell you what someone did (clicked a number), not how they feel about staying.
- NPS tells you likelihood to recommend — not why, not how they felt, not whether they’ll renew.
- CSAT gives a number after a specific interaction — you can get a 5/5 with “Issue was fixed.” Cool. But did they feel heard, or did they feel like they wasted 20 minutes?
- CES tells you how easy the experience was. Ease is not joy. You can make something easy and forgettable.
Scores are useful. But a genuinely happy customer doesn’t show up primarily in a survey. They show up in their behaviour.
The free proxies for a happy customer
Here’s what I’ve learned: the most reliable signals of a happy customer are almost all behavioural, observable without a survey tool, and free to track if you know where to look.
Proxy 1: They come back without being asked
The most honest signal a customer can give you is returning voluntarily. Not because you sent a campaign. Not because you triggered an onboarding reminder. Because they thought of you on their own.
In practice, this looks like: repeat purchases, repeat bookings, accounts that expand their usage month over month, people who sign up for your newsletter without being prompted, clients who reply to your check-in emails with something more than “Thanks.”
Track return behaviour in your CRM — even the free HubSpot tier shows you last contact and activity history. If a client hasn’t reached out in 45+ days and you haven’t either, that’s a yellow flag. A happy customer finds reasons to stay in touch.
Proxy 2: They refer people to you
A referral is the single most expensive thing a happy customer can do for you. It costs them social capital. They only do it when they’re genuinely confident you won’t let their contact down.
You don’t need a formal referral programme to track this. Just ask: in the last three months, how many new leads mentioned an existing customer’s name? How many intro emails have come through with “a friend suggested I reach out”? That number — however small — is your real NPS.
According to Apollo research, only 29% of B2B customers are engaged, with 60% being indifferent and 11% actively disengaged. (Source: Apollo) The 29% who are engaged are disproportionately the ones who refer. If you can identify who those people are, you’ve found your happy customers — and your best distribution channel.
Proxy 3: They tell you what’s broken
This one is counterintuitive. A customer who emails you to say something isn’t working, who flags a bug, who says “this part of the process is confusing” — that customer is not unhappy. That customer is invested.
The customer who worries you is the one who has stopped caring enough to bother. Feedback — including complaints — is a form of engagement. It means they still believe you might fix it. The silent customer has already given up.
So: track the ratio of customers who give you direct feedback (emails, calls, comments) versus those you hear from only at renewal time. The first group is more likely to renew. The second group is your actual churn risk.
Proxy 4: Their usage doesn’t drift
For any product or service with measurable engagement — a SaaS tool, a retainer, a subscription — usage drift is one of the earliest and most reliable warning signs.
The VP who championed your product stops logging in. The team that was running five projects a month is now running two. The client who used to send a brief every two weeks has been quiet for a month.
You don’t need a paid customer health platform to see this. A simple Google Sheet updated weekly with “last meaningful interaction” per account, flagged anything over 30 days, is enough for a team managing under 30 clients. It’s low-tech. It works.
As one SaaStr piece on silent churn put it bluntly: “We walked away from a $60,000 annual commitment, and the vendor still has no idea. Zero visibility into our declining satisfaction. No early warning system.” (Source: SaaStr) Usage drift is that early warning system — if you’re watching.
Proxy 5: They expand
A happy customer doesn’t just stay — they grow. They add seats, they upgrade a tier, they bring in a second team, they scope an additional project.
According to EverAfter’s customer health research, expansion signals appear when smaller accounts hit 80% of plan limits with strong engagement. (Source: EverAfter) You don’t need a tool to spot this — you just need to be looking. A client who asks “can we also use this for X?” is telling you more about their happiness than any survey score.
Track expansion conversations in your CRM. If they’re not happening at all, that’s information too.
The free stack to track happy customers without NPS
You don’t need a customer success platform. Here’s what you actually need:
A simple account health sheet — one Google Sheet, one row per client, columns for: last meaningful interaction date, usage trend (up / flat / declining), expansion conversations (yes/no), referrals given (yes/no), direct feedback received (yes/no). Update it weekly. Review it in your team meeting. Flag anything that looks wrong.
Your email inbox, read differently — stop looking at response rates as a marketing metric and start reading them as a relationship signal. Who replies quickly? Who opens but doesn’t respond? Who hasn’t opened the last three emails? That pattern is a proxy for engagement.
One honest call per quarter with your most important clients — not a sales call. Not a renewal conversation. A genuine “how are things going?” call where you ask open questions and actually listen. “What’s one thing we could do better?” and “What surprised you recently?” return more useful signal than any 10-question CSAT survey.
Text analysis on the words your customers use — you don’t need Medallia for this. Take 10 emails or chat messages from clients each week and look at the language. Words like “finally,” “still waiting,” “I had to,” “honestly” — these are emotional signals. Words like “great,” “worked perfectly,” “exactly what I needed” — these tell you what’s creating genuinely happy customers. Start labelling. Patterns emerge fast.
The one question that replaces NPS
If you do want to send a survey — and sometimes you should — here is the one question that returns more honest signal than “how likely are you to recommend us on a scale of 0 to 10?”
“What would have to change for you to feel even better about working with us?”
It assumes they like you (so it doesn’t feel like a complaint form). It’s forward-looking (so it’s easier to answer). And it separates the genuinely happy customers — who will struggle to answer because things are genuinely good — from the politely satisfied ones who will immediately name three things.
The answers to that question, read carefully, are more useful than three months of NPS data.
The honest truth about happy customers
A happy customer in B2B is not someone who gave you a 9 on a survey. It’s someone whose usage is growing, who refers people to you, who pushes back when something’s wrong because they’re invested, and who you don’t have to chase to renew.
Those signals are visible without a single paid tool. They just require you to look in the right places — and to stop confusing the absence of complaints with the presence of happiness.
85% of customers who left a provider say they would have stayed if their problem had been addressed before they left. (Source: Armatis citing Netigate 2025) The proxies above exist precisely to surface those problems before the customer goes silent.
You don’t need a platform for that. You need attention.
I write about B2B customer experience and zero-budget growth — at sebastien.no. Find me on LinkedIn.
Comments