From Stars to Signals: How AI Will Use Social Proof to Recommend Products
ChatGPT and other AI tools are moving into shopping. Here’s how reviews, ratings, and social proof will shape which products large language models recommend.
TABLE OF CONTENTS
The Problem With NPS: Why Your Score Isn’t Telling You the Truth
This in-depth guide breaks down Net Promoter Score (NPS) in plain English - what it is, why companies use it, and where it falls short. You’ll learn the pros and cons of NPS, how to avoid treating it like a vanity metric, and smarter ways to combine it with CSAT, CES, and review analysis.
Net Promoter Score (NPS) is everywhere. From scrappy startups to Fortune 500s, you’ll find leaders obsessing over one number: “What’s our NPS this quarter?” The idea is simple: ask customers how likely they are to recommend you on a 0–10 scale, subtract the detractors from the promoters, and you’ve got your magic loyalty metric.
It sounds neat. It looks tidy on a slide. And yes, it’s useful. But here’s the thing: on its own, NPS won’t tell you the truth about your customers. It’s more like a headline than the full story. If you treat it as the whole book, you’re missing the plot.
Let’s break down what NPS gets right, where it falls short, and how you can use it without falling into the “vanity score” trap.
What NPS Actually Is (In Plain English)
NPS was born in 2003 when Bain & Company consultant Fred Reichheld published “The One Number You Need to Grow.” The premise: you can predict loyalty by asking one simple question — “How likely are you to recommend us to a friend or colleague?”
- Promoters (9–10): Enthusiastic fans who’ll spread the word.
- Passives (7–8): Lukewarm. Satisfied, but not shouting your name.
- Detractors (0–6): Unhappy customers who may churn — or worse, talk others out of trying you.
Your NPS is calculated by subtracting the percentage of detractors from promoters. Passives don’t count. That’s it. Simple. Benchmarkable. Boardroom friendly.
Why Businesses Love It
NPS is sticky for a reason:
- It’s simple. One question, one number. Even the busiest execs get it.
- It benchmarks. You can compare scores across teams, products, or industries.
- It trends. Watch it over time, and you get a pulse on whether things are moving up or down.
For many companies, NPS becomes a north star. It’s easy to track, report, and celebrate.
But simplicity cuts both ways. Which brings us to…
The Cracks in the System
NPS has flaws - not fatal, but important.
1. It Over-Simplifies Human Emotion
People don’t live neatly between 0 and 10. You can be thrilled with a product but angry at the billing. You can love the service but hate the price. A single number can’t capture that nuance.
2. It’s Biased by Timing
Ask right after someone unboxes your product? You’ll likely get a 10. Ask after they’ve been on hold with support for 40 minutes? Expect a 2. Timing skews results more than most people admit.
3. Industry Skew
A “great” NPS in airlines might be 30. In SaaS, 50+. A dental practice? Maybe 70. Context matters. Comparing across industries is apples to oranges.
4. False Positives (and Negatives)
High score ≠ loyalty. Customers might say 9 but still churn because your competitor is cheaper. On the flip side, some detractors will stick around because switching is painful.
5. It Gets Gaming Fast
Tie bonuses to NPS and suddenly you’ll see frontline staff “nudging” customers: “Would you mind giving us a 10?” That defeats the point.
Using NPS the Right Way
NPS isn’t useless. It’s just incomplete. Here’s how to make it work:
- Always pair it with an open-text follow-up. “Why did you give that score?” is the gold. The number gets you the trend; the comment tells you the story.
- Look at trends, not snapshots. Don’t panic over one bad week or celebrate one good one. The long view matters.
- Segment your results. Break down by customer type, product, or lifecycle stage. A 20-year loyalist leaving a “6” is more concerning than a one-time free trial user.
- Don’t chase vanity. A higher number feels good, but the real question is: are we acting on what customers tell us?
Beyond NPS: Smarter Feedback Loops
If you only ask NPS, you’ll miss the real opportunities. Pair it with other tools:
- CSAT (Customer Satisfaction): Great for single interactions like support tickets.
- CES (Customer Effort Score): How easy was it to get what they needed?
- Churn or exit surveys: Why did they leave? Priceless for pattern-spotting.
- Review analysis: Public sentiment often highlights issues private surveys miss.
Here’s the truth: the real value isn’t in the number - it’s in the themes. If your NPS is 45 but everyone’s comment mentions “confusing billing,” you’ve got work to do. That’s where feedback becomes growth fuel.
Example: The SaaS Trap
A SaaS company hits an NPS of 60. Leadership cheers. But dig into the comments and you find:
- “Love the product, but support is slow.”
- “Billing is confusing — not sure what I’m paying for.”
Ignore those, and the score means nothing. A year later, churn spikes. The NPS was positive, but it didn’t predict the problem. Acting on the comments would have.
FAQ: NPS Without the Fluff
What’s a good NPS?
It depends. Above 0 means more promoters than detractors. Above 50 is strong in most industries. But don’t obsess over benchmarks — context matters.
Should we tie bonuses to NPS?
Careful. It creates pressure to game scores instead of fixing root issues. Reward actions that improve customer trust, not just higher numbers.
Can small businesses use NPS?
Yes — but only if you act on it. Asking customers for feedback and then ignoring it is worse than not asking at all.
Is NPS enough on its own?
No. It’s a starting point, not a finish line. Use it as a signal, then dig into the “why.”
At Troof, we see NPS as one piece of a bigger puzzle. On its own, it’s a score.
Combined with reviews, surveys, and customer comments, it becomes a map of where trust is breaking and where growth is hiding.
Our platform pulls in all that feedback, analyzes the sentiment, and shows you what to fix first. Because your reputation isn’t built on a number — it’s built on whether customers feel heard, respected, and valued.
Final Word
NPS isn’t bad. It’s just incomplete. The danger is treating it like gospel when it’s really just the headline. If you want the full story, you need to read the comments, spot the themes, and actually do something with what you learn.
Because customers don’t remember your score. They remember how you made them feel.
👉 Want to go beyond the score? Troof helps you surface the real signals hidden in feedback — so you can close trust gaps before they cost you growth.
Related Posts
How Much Do Star Ratings Really Impact Conversion Rates? More Than You Think.
An in-depth look at how star ratings influence conversions, backed by data and psychology. Explains why small rating shifts lead to big sales impacts and how businesses can manage reviews to stay competitive.
Behind the Queues: What 27,000 Manchester Airport Reviews Reveal About Trust
We analyzed 27,000+ Manchester Airport reviews to reveal how trust is built, broken, and repaired. The findings show that queues and pricing spark anger, while simple wins in service create lasting loyalty.