What a Xiaohongshu Research Report Should Contain

A useful Xiaohongshu research report connects a business decision to evidence that another reader can inspect. It should explain the question, exact sources and collection method, distinguish observation from interpretation, state what the sample cannot prove, and recommend the next validation step.

That is a higher bar than a slide deck full of screenshots. A screenshot can show that one post existed. It does not establish how the post was selected, whether it was promotional, how its language was interpreted, or whether the same view is common among Chinese customers.

For an international brand evaluating China, the report should make those boundaries visible enough for a product, strategy or research team to decide what to do next. This guide provides a buyer's checklist and an explicitly illustrative walkthrough. It is not a client case study and contains no newly collected Xiaohongshu findings.

Start with the decision, not the platform

The opening page should name the decision the research informs. “Understand Xiaohongshu” is too broad. Better questions include:

• Which customer problem should anchor our first positioning test?
• Which alternatives should we include in interviews?
• Which product claims require local proof before launch?
• Which objections should a concept test investigate?

The report should also define the intended customer, category, geography and decision date. A team choosing interview topics needs a different deliverable from one comparing product concepts or planning a pricing study.

This framing prevents a common reporting failure: collecting interesting material first and deciding what it means later. It also makes exclusions intelligible. A bounded Xiaohongshu study may help discover language, questions, comparisons and claimed experiences. It cannot by itself calculate market size, forecast revenue, prove purchase behavior or represent every Chinese consumer.

If the team is still defining its research workflow, begin with our practical guide to Xiaohongshu market research. For the wider launch decision, see what to research before entering China.

Put the answer and its confidence on the first page

A decision-maker should not have to reverse-engineer the conclusion from an appendix. The executive summary should contain:

1. The decision and research question.
2. The most decision-relevant findings.
3. The strength and limits of the supporting evidence.
4. The recommended action.
5. The evidence still needed before committing money or making a public claim.

Use calibrated language. “Maintenance questions appeared in the selected material” is an observation about a defined sample. “Maintenance may be a useful topic for customer interviews” is an interpretation and next step. “Chinese consumers prioritize maintenance” is a population claim that a selected social-media sample usually cannot support.

A good summary can recommend “do not decide yet.” Identifying a missing comparison, inaccessible source, uncertain translation or unverified product claim is useful when it prevents false confidence.

Make the method reproducible enough to audit

The method section should let a reviewer understand why particular material appears in the report. At minimum, record:

• the exact Chinese and English query wording;
• collection dates and relevant access conditions;
• the inclusion, exclusion and deduplication rules;
• the number and type of items inspected and retained;
• whether posts, comments, brand pages or other sources were included;
• how promotional material, reposts and suspected automation were handled;
• who translated and coded the material, including any software assistance;
• how disagreements, missing material and quality checks were handled; and
• the known limits of the design and collection process.

These are not arbitrary details. The American Association for Public Opinion Research says released research should disclose enough about its methods for independent review. Its standards specifically call for content-analysis inclusion rules, dates, sample sizes, coding procedures, quality controls and limitations. AAPOR's scope is public-opinion and survey research rather than a Xiaohongshu-specific protocol, so use it here as a transparency benchmark, not a claim that a social-content sample is a survey. Read the AAPOR disclosure standards.

If counts appear, the report must define the denominator and unit. “12 of 50 distinct retained posts” describes the reviewed material when the selection and deduplication rules are available. It does not become “24% of Chinese customers” or market share.

Give every important claim a source trail

A report needs more than a bibliography. Each decision-relevant claim should connect to an evidence register that preserves enough context for review without exposing unnecessary personal data.

The 2025 ICC/ESOMAR International Code emphasizes fit-for-purpose research, transparency, data minimization, privacy and accurate publication of findings. It also requires researchers to distinguish research activity from non-research activity. Those principles support a report design in which provenance, interpretation and responsibility remain visible. Read the ICC/ESOMAR code.

Do not hide weak evidence behind polished charts. If the source only supports “one creator claimed X,” the chart and headline should not imply verified performance or broad customer agreement.

Separate finding, interpretation, decision and test

Every important section should make four steps explicit:

1. Finding: what the inspected evidence shows.
2. Interpretation: a bounded explanation of why it may matter.
3. Decision: what the team could change or prioritize.
4. Next test: what stronger evidence would confirm or challenge the interpretation.

This structure makes disagreement productive. A product manager can accept the recorded observation while challenging the interpretation. A researcher can then design an interview, survey, concept test or product test around that uncertainty.

It also keeps Xiaohongshu evidence in the right role. Selected social material can be valuable for discovery: vocabulary, use situations, comparison criteria, objections and questions. Other methods are better suited to estimating prevalence, testing causality, measuring willingness to pay or forecasting sales.

An illustrative Xiaohongshu research example

The following walkthrough is entirely illustrative. It does not describe a Xiaohongshu search, a Whatson.red customer, observed posts or measured behavior.

Imagine an international reusable-drinkware brand deciding whether its first China concept test should emphasize temperature retention, portability or easy cleaning.

Question: Which usage situations, alternatives and proof questions should the concept test include?

Planned sources: A bounded set of clearly documented Chinese queries could explore insulated cups, commuting, leaking, cleaning, odors, cup holders, gift buying, domestic and international alternatives, and “not worth buying” language. Brand specifications and care instructions would be checked separately from social claims.

Illustrative evidence record: Suppose a retained post asked whether a lid gasket holds odor after repeated use. The observation would be the question itself and its context—not proof that the problem occurred. A separate official care page might verify how the gasket should be cleaned, but it would not prove customer satisfaction.

Bounded interpretation: Odor and cleaning may be useful evaluation criteria for this concept. The post's selection cannot show how prevalent the concern is, and the author may not be a verified buyer.

Decision: Include an easy-cleaning message and a removable-gasket demonstration as alternatives in the concept test rather than declaring them the winning position.

Next test: Ask intended customers to compare the three concepts, inspect the cleaning steps, explain tradeoffs in their own words and identify evidence they would need to believe each claim.

That chain—from question to source plan, observation, interpretation, decision and test—is the value of the report. The invented scenario is not evidence that cleaning matters in this category. It demonstrates how real evidence should be handled when collected.

Show contradictions, absences and commercial context

A credible report does not only collect support for the team's preferred story. It should include disconfirming queries and explain what would weaken each interpretation. Search plans might pair recommendation language with regret, failure, replacement, alternative and “not worth it” terms.

The report should also identify commercial context. Brand content, affiliate material, free products and creator partnerships can still reveal useful language or comparison criteria, but they should not be presented as independent customer testimony. Engagement does not establish sales, market share or product satisfaction.

For a worksheet that applies these distinctions to customer alternatives and positioning, see our guide to Xiaohongshu competitor analysis.

Absence needs care too. Failing to find a topic in a bounded collection does not prove that customers do not care about it. Query wording, ranking, access, deletion, moderation and the collection window all shape visibility.

Protect people while preserving auditability

Public availability does not make every reuse harmless. Retain only the personal data needed for the research question, limit access to source records, avoid unnecessary searchable quotations, and do not infer sensitive traits from profiles. The Association of Internet Researchers treats internet-research ethics as context-dependent and provides questions for considering data subjects, consent, vulnerability, reuse and potential harm. Read the AoIR ethics guidance.

An evidence register can be auditable without becoming a public archive of user identities. The appropriate balance depends on the collection method, purpose, platform terms and applicable law. A report should name unresolved ethical or legal questions rather than implying that a public URL settles them.

Use this buyer's checklist

Before accepting a Xiaohongshu research report, check that it contains:

• one defined business decision and a bounded audience;
• an executive answer with confidence and limits;
• exact queries, dates, source types and selection rules;
• sample size, deduplication, translation, coding and quality-control details;
• traceable evidence for every material claim;
• a visible separation between observation and interpretation;
• commercial-context and verification labels;
• contradictions, missing evidence and inaccessible material;
• privacy, ethics and platform-access considerations; and
• recommendations framed as decisions or tests, not unsupported forecasts.

The report should leave the reader able to say both “this is what the selected evidence supports” and “this is what we still do not know.” That distinction is more useful than a confident conclusion whose source trail disappears on inspection.

If your team needs to compare a proposed report scope or turn a China launch question into a source-linked English brief, tell Whatson.red Intelligence the decision you need to make. We can define the question, evidence trail and validation boundaries before research begins.