Observed
Describe the change actually present in the material.
Reducing false signals
Reduce noise, duplicates and false alarms
Check identity, independence and coverage before interpreting a change as a reputation signal.
What you get
First the signal. Then the conclusion.
More collected content may reflect republications, broader searches or newly connected sources. Check what changed in collection before interpreting what changed for your brand.
Also check changed queries and identical names before calling a spike a crisis or a success. The shape of the curve does not replace examining the material: growth can indicate better coverage and a decline an interruption. Both outcomes remain distinct from a change in how your brand is perceived.
| Apparent signal | Explanation to check | Evidence check |
|---|---|---|
| Many new mentions | Reprints of the same story or duplicated content. | Trace content origin and independence before counting new experiences. |
| More criticism than usual | A new source or a change in collection criteria. | Compare the common subset and make discontinuities visible. |
| A new topic associated with your brand | A namesake, indirect reference or different offering. | Check the subject, product and context before assigning priority. |
What you get
Confirm the entity
A matching name alone does not identify a company or product: aliases, domains, market and context help verify the match. Doubts remain visible. An ambiguous connection is excluded from brand conclusions until review clarifies attribution.
- Detected name
- Aliases and context
- Confirmed attribution
What you get
Distinguish repetitions from independent sources
An article may be republished across websites or accessible through different URLs: collected content counts should be distinguished from independent source counts. Provenance matters. Brand statements and press observations play different roles, which should remain identifiable when interpreting a topic’s spread.
- Original source
- Copies and republications
- Independent accounts
What you get
Check the collection scope
Compare the same periods, languages, channels and inclusion criteria. Show missing sources, collection errors and exclusions. A smaller or different sample can shift a proportion without a real sentiment change; absence within the sample does not prove general absence.
- Available sources
- Inclusion criteria
- Comparable periods
Checks before comparison
Remove the noise. Keep the evidence.
Deduplication prevents repeated counting of the same evidence. It does not mean deleting different accounts because they express similar opinions or come from the same domain.
Each exclusion needs a consistent criterion and a clear effect on the denominator. For AI answers, a technical error is not brand absence; two planned repetitions are not duplicates merely because their text matches. Collection identity and content identity play different roles.
| Apparent signal | Possible explanation | Required check |
|---|---|---|
| Many news items on the same day | Reprints of an original item. | Distinguish origin, reprints and independent experiences. |
| New brand mentions | A namesake or attribution to the wrong product. | Check subject, aliases, context and market. |
| Declining AI presence | A change in panel, model or provider availability. | Compare versions, eligible answers and errors separately. |
A proportionate conclusion
Say what the data supports.
Reliable interpretation distinguishes observation, explanation and the next check. It can be useful even when it concludes that the sample is insufficient for a decision.
Report the observed scope, exclusions and differences from the previous period. Avoid turning correlation into causation or a few experiences into the voice of the whole market. TrustInsider connects indicators to evidence precisely to make these checks possible, keeping cases requiring review open.
Hypothesised
State the proposed explanation and alternatives still possible.
To check
Identify the check that would strengthen the decision.
Related paths
Choose your next step.
Check your brand’s presence, compare results or choose content improvements: continue with the feature you need.
Guides to reputation, trust and AI
Prepare an AI check, verify a citation and compare results: guides with examples to improve your brand’s content.
ReadFrom question to action
From the question to checking the outcome: context, sources, hypotheses and action plans in the TrustInsider analytical workflow.
ReadA transparent method, not a black box
TrustInsider’s criteria for interpreting reputation and trust: provenance, samples, reviews and the limits of conclusions.
ReadFind the right answer
Find guides, FAQs and methodology to clarify indicators, coverage and access, with dedicated routes for data requests.
Read