Why customer interviews are good but not enough (and what to add)

A copywriter on Reddit described their research process: “I typically spend two weeks on customer research. But that is just waiting for survey results and/or to run 8-10 customer interviews followed by a day or two of analysis.”

Two weeks. 8-10 interviews. Then a day or two of analysis.

That is 80+ hours of work for 8-10 data points.

Meanwhile Joanna Wiebe at Copyhackers read 500 Amazon reviews in a few hours. She pulled one sentence from a reviewer. Tested it as a headline. CTA clicks went up 400%.

The interview is not the problem. The problem is stopping THERE.

Interviews are the gold standard. And the smallest sample you will ever get.

Interviews give you depth. Tone of voice. Facial expressions. Follow-up questions. The ability to dig into WHY someone said what they said.

But they also give you 8 data points. Maybe 10. From people who AGREED to be interviewed. Which means they are already more engaged than your average customer. More articulate. More positive. More willing to help.

Research on qualitative sample sizes confirms this. Data saturation (the point where new interviews stop revealing new themes) happens after just 9 interviews. Most copywriters do 6-8.

That means your entire messaging strategy is based on what 8 people told you in a structured conversation where they knew they were being recorded.

The 3 blind spots of interview-only research

Blind spot 1: Interviewer bias. You ask questions. The questions shape the answers. If you ask “What do you like about the product?” you get positive answers. If you ask “What frustrated you?” you get complaints. The framing decides what you find.

Online reviews have no interviewer. No framing. No leading questions. People say what they actually think. Unprompted. Unfiltered.

Blind spot 2: Social desirability bias. People give “correct” answers in interviews. They clean up their language. They avoid sounding negative. They want to be helpful.

On Reddit people say: “The grunt work is still on me.” On Trustpilot they say: “This is predatory and bordering on fraud.” That RAW language is what makes copy feel real. You will never hear it in an interview.

Blind spot 3: Small denominator. If 2 out of 8 interviewees mention pricing anxiety you have a 25% signal. Is that real? Or did you just happen to interview 2 price-sensitive people? With 8 data points you cannot tell.

With 65 sources you can. If 6 out of 65 sources mention pricing (9.2%) across Reddit forums. Review sites. And blog posts. That is a pattern. Not a coincidence. The Frequency x Intensity framework needs volume to work. Interviews alone do not give you that volume.

The 5 customer research methods that give you the full picture

Interviews are source #1. Here are the other 4. Together they give you hundreds of data points instead of 8.

Method 1: Customer interviews (depth)

Still valuable. Still worth doing when possible. The ideal: 6-8 interviews with real customers. Record them. Transcribe them. Pull verbatim quotes.

What interviews give you that nothing else does: Follow-up questions. “You said X. What did you mean by that?” The ability to chase a thread. To hear hesitation in someone's voice. To catch what they ALMOST said.

Limitation: 6-8 data points. 2+ weeks to schedule and complete. Interviewer bias. Social desirability bias.

A copywriter on Reddit put it well:

“Does the company have any reviews of their own. Either collected on their website or on a platform like Trustpilot. G2. Etc? If not. I recommend conducting customer interviews or surveys with their customer base. Another option would be to look at the reviews of their competitors. As you can find a lot of good ideas in there. Too.”

Notice the order. Reviews FIRST. Interviews as a fallback.

Method 2: Review mining (volume + raw emotion)

G2. Trustpilot. Capterra. Amazon. App Store. Any platform where real users leave unfiltered feedback.

What reviews give you: Hundreds of data points in an hour. Language people use when nobody is watching. The 2-3 star reviews are especially powerful — they come from people who TRIED the product and have specific objections and hesitations.

Limitation: No follow-up questions. You cannot ask WHY. You take what they give you.

The math: 30 minutes on G2 and Trustpilot = 30-50 quotes. 30 minutes on Reddit = 40-80 quotes. That is 70-130 data points in one hour. Compare that to 8 interviews over two weeks.

Method 3: Forum and community mining (unfiltered conversation)

Reddit. Quora. Slack communities. Facebook groups. Industry forums. Anywhere your audience talks to EACH OTHER instead of talking to you.

What forums give you: The way people describe problems to peers. Not to salespeople. Not to interviewers. To people like them. This is where you find the exact words your audience uses when nobody is selling to them.

A copywriter on Reddit described this source perfectly:

“Fellow writers! If you cannot directly ask customers these questions for whatever reason. Just do some DIY market research. Strive to answer these questions yourself. Hang out in the forums where they chill. Consume the same stuff they are reading. Watching online. And basically try to get inside their heads.”

Limitation: Noisy. You have to sift through a lot of irrelevant conversation to find the gold. But the gold is PURE because it was never shaped by your questions.

Method 4: Competitor analysis (positioning gaps)

Open your top 3-5 competitors' landing pages. Their pricing pages. Their FAQ sections. Their testimonials.

What competitor analysis gives you: What claims have already been made. What proof has been shown. What objections have been addressed. And more importantly — what has been MISSED. The gaps are your opportunities.

FAQ sections are objection lists in disguise. Every question a competitor puts in their FAQ is an objection they heard enough times to address publicly. That is free research.

Limitation: You see what competitors CHOSE to show. Not the full picture. Their testimonials are cherry-picked. Their FAQs are curated. Use this alongside raw data — not instead of it.

Method 5: Survey data and industry research (statistical proof)

Published surveys. Industry reports. Academic studies. Market research. These give you the NUMBERS that back up what the qualitative data tells you.

The VoC market is projected to reach $32.93 billion by 2030 (Grand View Research). That is a proof element for anyone selling VoC tools. The Copyhackers salary survey found that 56% of freelance copywriters work 31-50 hours per week. That is a proof element for anyone writing about copywriter productivity.

Limitation: No emotional language. No verbatim quotes. Just numbers. Use these for your proof elements section. Not for headlines.

The math: interviews vs. multi-source research

Interview-only approach: 8 interviews × 45 minutes each = 6 hours of conversation. Plus 10+ hours scheduling. Plus 4+ hours transcribing. Plus 2+ hours analyzing. Total: 22+ hours. Data points: 8.

Multi-source approach: 1 hour on review sites = 50 quotes. 1 hour on Reddit = 60 quotes. 30 minutes on competitor FAQs = 15 objections. 30 minutes on industry reports = 10 proof points. Total: 3 hours. Data points: 135.

22 hours for 8 data points. Or 3 hours for 135 data points.

Add interviews ON TOP of the multi-source research? Now you have 143 data points. With the depth of interviews AND the volume of public data. That is the complete picture.

How to combine all 5 methods

Here is the order that works. From fastest to slowest:

Day 1 (3 hours): Review mining + Reddit mining + competitor FAQs. This gives you 100+ quotes. You already know the top pains. Top desires. Top objections. Here is the full source list.

Day 2 (1 hour): Organize everything into categories. Pain points. Desires. Objections. Emotional language. Count frequency.

Day 3-7 (if possible): Schedule 4-6 interviews. But NOW you know what to ask. Because the public data already showed you the patterns. The interviews CONFIRM and DEEPEN. They do not discover.

Day 8 (2 hours): Add interview quotes to your research brief. Compare with public data. Where do they align? Where do they differ? The alignment is your lead message. The differences are worth investigating.

The one thing to remember

Interviews are not bad. They are incomplete.

8 data points from people who agreed to talk to you is not the same as 135 data points from people who said what they actually think in public. To strangers. Without being asked.

Use interviews. But never use ONLY interviews. The best customer research methods combine depth with volume. Structured with unstructured. Private with public. When you combine all five methods you end up with hundreds of real quotes — and the copy almost writes itself.

That is how you get the full picture.