Unlocking User Insights: Leveraging Online Reviews with NLP for Product Research

During structured interviews, users share their views, attitudes, and opinions about a product. Beyond interviews, the brief comments typed by loyal users online also convey valuable insights. When faced with large volumes of textual user comments, technology employs NLP algorithms for processing, which are widely used in opinion extraction, sentiment analysis, knowledge graphs, intelligent Q&A, and machine translation.

Researchers now have the chance to stand on the shoulders of these technologies, bypassing complex NLP algorithm barriers to see how user comments can aid user research in unfamiliar domains.

User researchers typically use surveys, interviews, and focus groups to understand user perceptions and needs, but many user voices already reside unnoticed in comment sections before any formal study begins.

Comments cover everything—from product pain points like "poor sound quality, high latency in games" to brand trust and affordable pricing, to expectations for new features such as "water‑resistant for swimming and running". By analyzing these comments before recruiting participants, researchers can preliminarily map product pain points and user‑focused attributes, gaining early immersion in real user scenarios and emotions, thus achieving greater efficiency.

Scenario 1: Category Pain‑Point Review – JD.com Comments

TIPS: Summarize purchase‑user comments to extract product pain points and strengths, informing qualitative interview guides and quantitative option design, filling category blind spots, and enriching real‑world usage contexts.

Example: For a fresh‑food category, the business side wants to understand emerging lines’ development status and product issues but lacks sufficient user samples. By analyzing over 10,000 purchased‑user comments and selecting 500 negative reviews, key pain points such as standardization, quality control, cold‑chain logistics, delivery timeliness, and market pricing were identified.

For non‑technical teammates, using Python libraries like jieba or cloud NLP services (JD Cloud, Baidu Cloud, PaddlePaddle) may be daunting. A newly launched tool, EasyIdea (EasyIdea.jd.com), offers a simple interface: input product category, brand, or SKU to obtain scene‑pain analysis and sentiment results.

Example: Fresh‑food category user pain‑point analysis

Scenario 2: Omni‑Channel Business Review – Dazhong Dianping

TIPS: Analyze offline business user consumption comments to extract experience highlights, supplementing research gaps.

Example: For an offline omni‑channel business, the team needs insights on surrounding commercial districts and competitor layouts. Limited on‑site visits and dispersed interview samples restrict feedback. By examining Dazhong Dianping reviews of target districts and city‑wide innovative formats, strengths and weaknesses are mapped, enriching the research report with missing perspectives.

Example: User evaluation of a Beijing offline store

Dazhong Dianping comments contain rich tags, photos, and videos, offering valuable supplementary information for offline venue scouting and trend analysis, though anti‑scraping measures may require Python skills.

Scenario 3: Product Platform Review – App Store

TIPS: Scrutinize user comments on online apps to uncover experience highlights and pain points, identifying opportunities for product feature extensions.

Example: For a new app, researchers examine user experience and pain points beyond interaction flow, including perception and mental models. By sorting App Store or Android store reviews by date, they track attitudes toward the latest version and new features, and compile strengths and weaknesses to align with competitor trends.

Example: User evaluation of an e‑commerce platform app

While user opinions are valuable, researchers should avoid blindly following noisy, emotional, or duplicated comments, as extracting actionable insights can be challenging and may lead to misdirected research. In the long run, encouraging high‑quality original reviews and applying big‑data algorithms to mine them will further enhance user experience strategies.