Market research should help a business decide what to improve, explain or investigate. User-generated content (UGC) can contribute by showing how people describe products and experiences in reviews, public posts, forums and videos.
These materials can reveal questions you had not thought to ask. They are not automatically verified, unbiased or representative of all customers. People choose whether to post, platforms shape what becomes visible, and commercial relationships or fabricated content can affect what you find.
Start with a research decision
Choose a question narrow enough to act on. For example: why do new users struggle with setup, what information is missing from a size guide, or which use cases appear repeatedly in product reviews? A collection of mentions becomes more useful when it serves a specific decision.
Define the sources, date range and selection rules before reading for conclusions. Include relevant critical feedback as well as praise. Do not discard an inconvenient complaint merely because it conflicts with the result you hoped to find.
Know whose experiences are missing
A busy forum may overrepresent enthusiasts; a review page may attract people with particularly strong experiences. Public posts can also omit customers who prefer private support or do not use that platform. Volume alone does not correct those gaps.
Describe findings as themes in the material reviewed, rather than claiming a percentage of all customers believe something. Avoid inferring age, income or sensitive characteristics from a profile when those details were not reliably provided. Compare the themes with support records, interviews or a deliberately designed survey before making a major product decision.
Collect only what the research needs
Access depends on the platform, account permissions and available tools. Public visibility does not grant unlimited permission to collect, retain or republish personal information. Respect relevant access terms and privacy requirements, and avoid collecting private-group or direct-message material without an appropriate basis.
For a small review, a table may be sufficient: source link, date, product version, topic, observation and confidence. Restrict access to identifiable material and set a retention period. Use de-identified summaries where the research question does not require a person’s identity.
Turn observations into testable themes
Regular monitoring social media can surface recurring product and service questions. Create a simple coding scheme, such as setup, reliability, price and support, then apply it consistently. Separate what a post actually says from your interpretation. Record both a complaint and a successful workaround when they appear together.
Spend time verifying the authenticity of the content: check the original post, its date, the relevant product and whether a quoted experience can be corroborated. Duplicate posts, incentives, spam and synthetic reviews can distort the picture. Automated sentiment labels can also miss sarcasm and context, so review examples manually.
Act without overstating the evidence
Use a repeated theme to propose a small test: rewrite an installation step, add a comparison table or ask users to evaluate a prototype. Document which observations led to the change and check whether the relevant support issue declines afterward.
Over time, this process can help identify trends & patterns in customer behavior, but changes in posting activity are not necessarily changes in the whole market. Keep the source and time period attached to each finding.
If you later want to use a customer’s words or video in marketing, treat that as a separate permission and accuracy check. In the US, the FTC’s consumer review rule addresses practices including fake reviews and incentives conditioned on a positive or negative sentiment. Research should not become a reason to manufacture endorsements or suppress genuine criticism.
Meet the Author
Cutis Lee is a researcher and consultant for multiple industries including information technology, blockchain & business development.
Leave a Reply