Every copywriter knows the principle: write about benefits, not features. Speak to the transformation, not the product. But knowing the principle and actually executing it are different problems. Most ad copy still reads like it was written by someone who knows the product extremely well and is trying to share that knowledge - rather than someone who understands exactly what the buyer is feeling before they've heard of you at all.

The fix is not a better creative brief. It is primary research. Specifically, it is reading what your customers and your competitors' customers have already written about the category, the problem, and the purchase decision - in their own language, under their own motivation, before any marketer got near them.

This is voice of customer (VoC) research. It is one of the highest-leverage things you can do before writing a single word of ad copy. And most brands skip it entirely.

The Problem with Brainstorms

A typical copy brainstorm starts from the wrong place. The team knows the product inside and out. They know its differentiators, its ingredients, its backstory. So they write copy that reflects that knowledge. They lead with the mechanism. They use the category's professional vocabulary. They describe features in the brand's polished voice.

That copy fails not because the information is wrong, but because it is written in a language the buyer is not speaking when they encounter it. A potential customer scrolling through their feed is not thinking about your product's formulation or your founder's origin story. They are thinking about a specific problem they have - and they are using very specific words to think about it. If your ad does not use those words, it does not register as relevant, and the scroll continues.

"The goal is not to write good copy. It is to find the words your customer is already using in their head and put them in your ad."

The best copy is often not invented - it is excavated. Here is where to dig.

Where to Mine Customer Language

These sources are ranked roughly by signal quality. Work through them in order until you have 50 to 100 usable quotes.

Source 01

Post-Purchase Surveys

If you ask your customers one question after they buy, make it this: "What was the one thing that almost stopped you from buying?" The answers to that question are your objection library. They are also, when flipped, your most compelling proof points. A customer who almost did not buy because they were not sure it would work for their specific situation is telling you exactly what your next ad needs to address.

A second valuable question: "How would you describe what you bought to a friend?" The answer is your word-for-word copy. Not a cleaned-up version of it. That exact phrasing.

Source 02

Amazon and Competitor Reviews

Go to Amazon and find your product or the closest competitor with meaningful review volume. Read the 3-star reviews first. Not the 5-stars (which perform positivity) and not the 1-stars (which are usually logistical complaints). The 3-star reviews are written by people who are honest and specific. They got part of what they wanted. They will tell you precisely what the product did and did not deliver in language that has not been polished by anyone.

Then go to the 5-star reviews and mine for the before-and-after language. You are looking for phrases like "I used to..." or "I was so tired of..." or "after trying everything else." Those openers are your ad hooks.

Real example structure: "I've tried every X on the market and they all [specific failure]. This one actually [specific outcome]." That sentence is worth more than an entire brainstorm session.

Source 03

Reddit Threads

Find the subreddits where your buyers congregate before they are buyers. Look for threads where people describe their problem and ask for recommendations. The framing they use to describe their situation - before they have a solution - is the most valuable copy data you can find. It is problem language from a genuinely problem-aware state.

Pay particular attention to how people describe the stakes. "I've been dealing with this for three years" or "I have an important event in six weeks" - that urgency language is exactly what your ad should reflect back to people who feel the same way.

Source 04

Customer Support Transcripts

Every objection that reaches your support team is an objection your ads and landing page did not resolve. Read your last 30 pre-purchase support conversations or live chat transcripts. The questions people ask before buying ("Does this work if...?", "How long before I see results?", "What's the difference between X and Y?") are the questions your next ad should answer before they are asked. Proactively addressing these in your creative reduces friction downstream and qualifies the click before it happens.

What You're Actually Looking For

When you are reading through raw VoC data, you are hunting for three types of language specifically:

  • Before language - how they describe their life or problem before the solution. "I was constantly..." or "Every time I tried to..." This becomes your hook. The first three seconds of any ad need to create immediate recognition in the person who has this problem.
  • Failure language - how they describe what they tried that did not work. "I had already tried X, Y, and Z..." This positions your product without you having to make the claim directly. When your ad speaks to the failure of prior attempts, people who share that experience self-identify immediately.
  • Outcome language - the specific, concrete results they describe, not the abstract benefits. Not "improved sleep" but "I stopped hitting snooze for the first time in years." Not "better focus" but "I actually finished the thing I'd been putting off for three months." Specificity is credibility.

Flag every phrase that could work as an ad headline or opening line. You are building a copy bank, not a summary. The raw phrases are the output.

From Raw Quotes to Ad Copy

Once you have 50 to 100 tagged quotes, group them by theme. You will start to see 4 to 6 angles that appear repeatedly across sources - the same underlying fear or desire expressed in slightly different words. Each of those clusters is a copy angle.

Your job at this stage is compression, not rewriting. Take the clearest version of each angle and trim it to hook length (under 10 words for a headline, under 20 for an opening line). Keep the customer's vocabulary intact. Do not translate it into marketing language. If they wrote "I was absolutely exhausted by 2pm every single day," your headline should sound like that - not "Combat afternoon fatigue with our formula."

The Distinction That Matters

There is a meaningful difference between copy that describes the transformation your product delivers and copy that reflects the language your buyer uses to describe their own situation. The first is written from the inside out. The second is written from the outside in. This is the same distinction covered in the drill vs. the hole framework - but VoC research is the mechanism that actually gets you there, rather than the principle alone.

You should exit this process with 4 to 6 distinct copy angles, each grounded in real customer language. Those are your creative hypotheses. Now you need to test them.

Then You Test It

VoC research does not tell you which angle will win. It tells you which angles are worth testing. You still need to run the ads to find out which resonates strongest with your current audience, at your current spend level, against your current creative format.

The testing structure matters here. Running five angles in a single ad set with $20/day is not a test - it is noise. You need enough budget and run time for each angle to generate statistical signal. For most Meta campaigns, that means at minimum $50 to $75 per angle before drawing conclusions. If you are running video, the hook variants should be isolated as the primary variable so you can attribute performance differences to the copy rather than the format.

Before you commit significant budget to paid testing, consider validating your strongest angles with email performance data first. Open rates and click-through rates on email sends are a free creative test - they tell you which hooks and angles your audience actually engages with before you spend on ads.

Over time, your VoC mining and your creative testing compound together. Each winning angle tells you something new about your audience's psychology. Each losing angle tells you something too - often something about awareness level or objection hierarchy that is just as valuable. A proper testing framework turns this data into institutional knowledge rather than one-off wins.

The brands that do this consistently do not run out of copy ideas. They run out of budget before they run out of angles. That is a much better problem to have.

Start with 30 minutes on Amazon reviews for your category and your three closest competitors. Highlight every phrase that contains the words "used to," "finally," "tired of," or "I was." That is your first copy bank. Write three headlines directly from that raw material and compare them against whatever you are currently running. The gap will be instructive.

Want copy that actually converts?

We run VoC research as part of every engagement - before we write a word of ad copy. If your creative is not resonating, let's figure out why.

Talk to us →

Frequently Asked Questions

What is voice of customer (VoC) and how does it apply to ad copy?

Voice of customer refers to the actual language your customers use to describe their problems, desires, and experiences - captured in their own words rather than filtered through marketing messaging. In ad copy, VoC research means mining sources like product reviews, Reddit threads, post-purchase surveys, and customer support conversations to find the exact phrases buyers use when they are in a problem-aware state. When your ad reflects that language back to them, it creates an immediate sense of recognition - the reader feels understood rather than sold to. That recognition is what separates a 0.8% CTR from a 2.5% CTR on the same audience.

Where do you find the best voice of customer data for ad copy?

The most useful VoC sources for ad copy are, in rough order of signal quality: your own post-purchase surveys (especially open-ended questions asking what almost stopped them from buying), one-to-one customer interviews (15 minutes on the phone with 5 recent buyers beats 50 survey responses), Amazon reviews on your product or direct competitors (especially 3-star reviews, which are specific and honest), Reddit threads in relevant subreddits where people describe their problem before finding a solution, and customer support tickets and live chat transcripts. Social media comments and testimonials are useful but lower-quality because people often perform positivity there rather than expressing their actual buying psychology.

How do you turn raw VoC data into actual ad copy?

The process is extraction first, editing second. Collect 50 to 100 raw quotes from your sources and tag each one by theme - specific fears, specific desires, specific before-and-after descriptions, specific language about alternatives they tried and rejected. Look for phrases that repeat across multiple sources in similar form: those are your highest-signal copy candidates. Your job is not to rewrite these phrases but to use them nearly verbatim in your ad headline, hook, or opening line. Once you have 3 to 5 candidate angles from the data, run them as creative tests to see which resonates strongest with your audience.

Can I use competitor reviews to find ad copy for my product?

Yes, and competitor reviews are often more revealing than your own. When someone leaves a 3-star review on a competitor product, they are telling you exactly what the market wants that the competitor did not deliver. That gap is your positioning. If 40 reviews of a competing product mention that the taste is tolerable but not great, and your product actually tastes good, that is your ad angle. You are speaking to a documented market complaint and positioning your product as the solution. Reddit threads where people discuss alternatives they tried before finding the right solution are equally valuable - that consideration language is precisely what your ad needs to address.