User generated content analysis: real product use

Short answer

To analyse product use experience in user-generated content (UGC — reviews, posts, videos and blog write-ups created by customers themselves), code each post on six fields: where and in what situation the product is used, what job the person wants done, at which moment they hit difficulty or delight, what the outcome is, what workaround they use, and what unexpected purpose they use the product for. After 40–60 posts, patterns appear: how the product really works in everyday life and how that differs from what you imagined. It is raw material for product, support and marketing alike.

Why UGC tells you something different

A star rating says "good" or "bad". Detailed user content says "how": which room the person installed it in, what they did not understand on day one, how the children treated it, what changed a month later. Surveys do not ask about these details, because nobody thinks to. UGC is the cheapest source showing real conditions of use — unlike a lab test or a focus group, here people assume nobody is watching.

A six-field coding frame

1. 1ContextWhere, with whom, in what conditions: a small flat, an office, a family with children, a country house.
2. 2JobWhat the person wants to achieve: saving time, tidiness, health, a gift.
3. 3MomentWhich moment of the experience is described: unboxing, installation, first use, a month later, repair.
4. 4OutcomeSuccess, partial success, failure — in the person's own words.
5. 5WorkaroundHow the person compensates for a shortcoming: an extra tool, their own rule, another product.
6. 6Unexpected useWhat unintended purpose the product is used for.

The "first week" lens

The most valuable UGC often describes the first week. In this period people learn the product, make mistakes, look for instructions — or fail to find them. Sentences such as "the first time I connected it nothing worked, then I realised..." point straight at support and instruction gaps. In coding, keep posts whose "moment" is "first week" as a separate group — they explain most of the calls customer service receives.

Who takes what

• Product — unexpected uses and workarounds: a new version, a new accessory, a design change.
• Support — first-week difficulties: instructions, short videos, ready replies.
• Marketing — real contexts of use: does the situation shown in ads match the customer's real situation?
• Sales — "not a fit for" signals: in which context the product does not work, and this should be said upfront.

Sampling: not just the best

When choosing posts to analyse, do not settle for the most popular and most positive ones. The most shared UGC is usually the most striking, not the most typical. Include posts of every outcome — success, partial success, failure — and from different sources: forums, blogs, review sites. Forty to sixty posts are usually enough to see the first patterns; if more stop revealing new ones, you can stop.

Illustrative example

This is an illustrative example. A company selling robot vacuum cleaners codes 50 user posts from forums and blogs. Three patterns emerge. First, context: most posts come from older flats with carpets, while the ads show a new flat with parquet floors. Second, the first week: many people write that the robot gets caught on cables and they tidy the cables into a box themselves — a workaround. Third, unexpected use: pet owners run it twice a day to pick up fur. The company adds a "tidy cables on day one" step to the instructions and prepares separate content for pet owners.

Rights and privacy

• Get the author's explicit permission before using UGC in your own ads, website or catalogue.
• Photos and videos may show other people, especially children — their permission is a separate matter.
• Do not collect the author's personal profile for analysis; coding concerns the post's content, not the person.
• Do not copy sensitive details such as health or family life into the analysis table.
• Keep quotes anonymous in internal reports.

The limit of text-free content

A large share of user content is photos and video, and many of them never write the brand's name. Monitoring that works by word search does not find such content — it shows up only when your name appears in the title, description or comments. So present the results of UGC analysis as "users who write in text", not "all users".

Limitations

Users who create UGC are a small and more active part of all customers: enthusiasts and the very unhappy write more. Coding is subjective — two people may code the same post differently, so the fields need short definitions. UGC should not be the sole basis for a product decision; read it together with sales, returns and support data.

Common mistakes

• Choosing only the most popular and most positive posts.
• Looking at the star count and not reading the post itself.
• Not keeping first-week posts as a separate group.
• Using UGC in ads without permission.
• Forgetting that text-free content is invisible.

Linking it to idea cards

Unexpected uses and workarounds are the strongest product-idea signals, because the person has already spent time or money. Move them onto the idea card from product ideas from social listening. When you want to use praise in advertising, see the permission rule in viral praise.

UGC with Vexvon Monitoring

Vexvon Monitoring searches the text on the sites you add — blogs, forums, review pages — for brand and product names, and gives each post a one-sentence summary, a tone and topic tags. That speeds up choosing posts for coding; a person fills in the six fields. There is no claim of recognising a brand inside images or video — monitoring works with text. More on the Vexvon Monitoring page.

Next step

This month, choose 40 user posts — from every outcome type and at least three sources — and code them on the six fields. Pass the most frequent first-week difficulty to support and the most interesting unexpected use to product. Other articles are in the campaign, content and lead signals section; to look at it together, get in touch.