Research methods

Voice of Customer Research from Social Media: Collect Your Customers' Own Words

How to run voice of customer research with social media: find where customers describe problems, collect their exact words from comments, code them and use them.

Voice of customer research collects the exact words customers use to describe their problems, the outcomes they want and their objections, and social media is one of the richest free sources of it: Reddit threads, YouTube review comments and TikTok comment sections are full of unprompted, detailed descriptions. Surveys tell you what customers say when asked. Social posts show what they say to each other. This guide covers where to look, how to collect and code customers' words, and how to use them in copy, product and onboarding without misrepresenting anyone.

What you are collecting

Voice of customer categories
What it sounds likeWhere it is used
Problem"My heels always slip in wide shoes"Headlines, ad hooks, landing page opening lines
Desired outcome"I just want to run 10k without blisters"Benefit copy, product positioning
Objection"Wide-fit versions never come in good colours"FAQ, comparison pages, product roadmap
Alternatives tried"I tried insoles, then sized up"Competitive positioning, onboarding
Trigger"After my third black toenail I gave up"Ad timing, targeting, content topics
Exact words"toe box", "heel lockdown"Keywords, product naming, search terms

Where customers describe problems in detail

  • Reddit: long threads where people explain their situation, what they tried and why it did not work. Use Search posts with Top for Past year, then Post comments on the best threads.
  • YouTube comments: under review and tutorial videos, viewers add their own experience and questions. Use Search videos for product type review, then Video comments.
  • TikTok comments: shorter, but numerous and candid. Use Search videos, then Video comments, and Comment keywords to see the common words in a video's comments.
  • X replies: replies to brands and creators often contain complaints. Use Post replies on relevant posts.
  • Lemon8 and Instagram: for lifestyle and beauty categories, posts and hashtags show what people value visually.

Collect systematically

  1. Write five seed searches. Use the problem, not your product: wide feet running shoes, heel slip running, "toe box" blisters.
  2. Run them on two or three networks. Reddit and YouTube for depth, TikTok for breadth.
  3. Pick the richest sources. The 10 to 20 threads and videos with the most relevant discussion.
  4. Pull the comments. Run Post comments or Video comments on each, 1 credit per search, with Top sorting.
  5. Export. Select the results and export them as CSV. Exports are free.

For a typical category this is 30 to 60 credits and an afternoon's reading. Every result is saved to your Results, so you can come back to a thread without paying again.

Code the words

Read the export and tag each useful comment with one of the categories above. Keep the exact words in a separate column; paraphrasing loses the value. A simple rule of thumb: a phrase is worth using when you have seen it, or something very close, from several different people in different places. Our guides to comment analysis and building an audience question bank show coding methods in more detail.

Use it

  • Copy: open with the problem in customers' words. "Heels that slip" beats "superior fit technology".
  • Product: rank objections by how often they appear and how costly they are to fix.
  • Onboarding and support: answer the questions customers ask each other before they ask you.
  • Content: each recurring question is a video, a post or an FAQ entry.
  • Search: the exact words become the keywords you target on social and web search. See social media keyword research.

A worked example: running shoes for wide feet

A running shoe brand wants to rewrite its wide-fit product page. The example is illustrative.

  1. Reddit Search posts for "wide feet" running shoes and title:"wide toe box" with Top and Past year find a dozen long threads in running communities.
  2. YouTube Search videos for wide feet running shoes review and Video comments on the three most viewed reviews add viewer experiences.
  3. TikTok Search videos for wide feet running and Comment keywords on the top video show "toe box", "heel slip" and "size up" as common words.
  4. Coding 200 comments finds three dominant problems (heel slip, toe pressure, limited colours) and one recurring alternative (sizing up, which causes heel slip).

The new page leads with "Room for your toes without your heel slipping", answers "should I size up?" directly, and adds two colours. About 40 credits of research.

Do it fairly

  • Quote without names. Use the words, not the usernames, and avoid quotes so distinctive they identify a person.
  • Do not invent or edit testimonials. Customer language from public posts is research input, not an endorsement. Presenting it as a review from your customer would mislead.
  • Respect communities. Read, do not spam. If you join a conversation, disclose who you are.
  • Keep the sample in mind. People who post are not all customers. Combine social VoC with interviews and support tickets before big decisions; see market research using social media.

Frequently asked questions

What is voice of customer research?

Research that captures how customers describe their problems, needs, objections and experiences in their own words, so you can use that language and insight in marketing, product and service decisions.

Can social media be used for voice of customer research?

Yes. Reddit threads and comment sections on YouTube and TikTok contain detailed, unprompted descriptions of problems and products. Combine them with interviews and support data for big decisions.

How many comments do I need?

Enough that the main themes stop changing as you read more. For a single product category that is often a few hundred comments from 10 to 20 good sources.

Can I use customers' social posts in my marketing?

Use the language and insights, not the posts themselves as testimonials. Quoting or featuring someone's post in marketing needs their permission.

How much does it cost in Superlurk?

Most searches, including comment searches, cost 1 credit. A typical voice of customer study uses 30 to 60 credits; exports are free.

How this guide was written. Written by the Superlurk team on 7 November 2026 from the app's current searches. The worked example is illustrative.

The Superlurk TeamSocial research

We build Superlurk for public-data research across 25 supported social networks, with the price shown before every search. We write practical guides to search methods, source verification and AI-assisted research.

Start lurking for free.

Every network, no card needed. See the price first, and pay nothing if a search fails.