NativeData & Provenance
Know what
grounds the twin.
Trace the evidence behind the answer, where it came from, and how it stays current.
At a glance
Start at the result.
Each result links to the twin population behind the result, the sources used to build that population, and down to the population’s individual respondent.
Step 01See what grounds the twin.
Native digital twins are grounded in real customer and market evidence.
Step 03Keep it current.
Native refreshes twins as new data becomes available.
Step 04See the gaps, not guesses.
The result shows the limitation instead of hiding the gap.
Step 06
Start at the result.
Each result links to the twin population behind the result, the sources used to build that population, and down to the population’s individual respondent.
- A twin result
- The twin population behind it
- The sources used to build that population
- Each individual respondent
Each result links back through every step Read the evidence trail.
Provenance is the evidence trail behind a result.
The trail shows - Sources
- Which sources ground the twin
- Origin
- Where those sources came from
- Freshness
- How current the source data is
- Calibration
- How the twin was calibrated and validated
Native records source origin, approvals, and changes through its HumanOS governance layer.
See what grounds the twin.
Native digital twins are grounded in real customer and market evidence.
Native foundation dataNative’s own data agreements, panels and research, plus public signals
- Data-provider agreements
- Native panels
- Multi-year research programs
- Reviews
- Forums
- Retailers
- Social discourse
Your selected researchBrought forward for the decision at hand
- Respondent-level surveys
- Segmentations
- Trackers
- Interview transcripts
- Open-ended responses
Your selected first-party dataWhere permitted
- CRM
- Loyalty
- Behavioral records
- Transactional records
What data can ground a Native digital twin?See how a twin is assembled
Keep it current.
Native refreshes twins as new data becomes available.
The cadence is set with each client: daily, weekly, monthly, quarterly, semi-annual, annual, or when a new study arrives.
Different sources refresh at different rates - Native foundation dataRefreshes daily
- Your sourcesDaily, weekly, monthly, quarterly, semi-annual, annual, or when a new study arrives
- Import datesVisible to users, who can sort audiences by date
Update, then validate.
When a new study, tracker wave, or first-party dataset arrives, Native’s team reviews the data, adds it to the relevant twins, and validates before the updated twins go live.
- 1A new study, tracker wave, or first-party dataset arrives
- 2Native’s team reviews the data
- 3Native adds it to the relevant twins
- 4Native validates before the updated twins go live
What happens when the underlying evidence changes?How validation works
See the gaps, not guesses.
The result shows the limitation instead of hiding the gap.
Twins without enough context to answer a question do not respond, which lowers the response rate rather than filling the gap with a guess. The faithfulness score shows how much inference an answer required.
Answered from its source dataNot enough context, so no answer
The open circles lower the response rate. No guess fills them. Adding more twins improves coverage, but adding twins does not create context missing from the source data.
What happens when there is not enough evidence to support a strong answer?
Add your own research.
Your organization adds its own research and first-party data to make twins native to your customers, competitors, category, and market.
Common starting points are completed segmentations, brand trackers, usage and attitude studies, and concept tests.
Before ingestion, Native - Reviews each dataset
- Removes personal identifiers
- Confirms data rights
Can we add our own research and first-party data?Is our data used to train shared or global AI models?
Frequently asked questions
Native twins are grounded in real customer and market evidence, and every result traces back to the sources behind it.
Native digital twins are grounded in real customer and market evidence. Sources include Native’s foundation data, which draws on Native’s own data-provider agreements, Native’s own panels, multi-year research programs with established research providers, and public signals such as reviews, forums, retailers, and social discourse. Your selected intelligence adds respondent-level surveys, segmentations, trackers, interview transcripts, open-ended responses, and permitted first-party data such as CRM, loyalty, behavioral, and transactional records.
Twin AnatomyProvenance is the evidence trail behind a result. The trail shows which sources ground the twin, where those sources came from, how current they are, and how the twin was calibrated and validated. Native records source origin, approvals, and changes through its HumanOS governance layer.
Yes. Each result links to the twin population behind the result, the sources used to build that population, and down to the population’s individual respondent.
Native refreshes twins as new data becomes available. The cadence is set with each client: daily, weekly, monthly, quarterly, semi-annual, annual, or when a new study arrives. Different sources refresh at different rates; our Native foundation-data refreshes daily. Users can also see when the underlying data was imported and sort audiences by date.
When a new study, tracker wave, or first-party dataset arrives, Native’s team reviews the data, adds it to the relevant twins, and validates before the updated twins go live.
The result shows the limitation instead of hiding the gap. Twins without enough context to answer a question do not respond, which lowers the response rate rather than filling the gap with a guess. The faithfulness score shows how much inference an answer required. Adding more twins improves coverage, but adding twins does not create context missing from the source data.
Yes. Your organization adds its own research and first-party data to make twins native to your customers, competitors, category, and market. Common starting points are completed segmentations, brand trackers, usage and attitude studies, and concept tests. Before ingestion, Native reviews each dataset, removes personal identifiers, and confirms data rights.
PlatformIs our data used to train shared or global AI models?Every twin is assembled from something you can name.
What a twin is built from, how the layers contribute, and what the evidence trail behind a result looks like. See how a twin is assembled
Back to the Trust Center