ESOMAR: Xplorence’s 37 answers for sample buyers
Self-description of processes and limitations. Updated: 9 October 2026.
1. What experience does the company have?
Xplorence develops a research platform and works with its own panel and field interviewers in Uzbekistan. The panel is growing; current size and suitability are assessed before ordering.
2. Where does the sample come from?
Our opt-in panel, open questionnaire links and face-to-face interviews (CAPI). Additional advertising recruitment is assessed for each project. External sample exchanges are not currently used.
3. Which recruitment channels are available?
Web registration, Telegram, participant referrals, partner links and advertising campaigns. The source is recorded when attribution is supplied at registration or response.
4. Is the sample mixed?
A project may combine panel, open-link and field recruitment. Its source mix is project-specific; no single mix is promised for every study.
5. How do people join the panel?
Voluntarily, with consent. Sign-in supports phone or email verification codes and Telegram. Consent to further invitations can be withdrawn.
6. Is the participant source recorded?
Supplied campaign tags, referrer, landing page and partner code are stored for source assessment. Without attribution, the source may be unknown.
7. How are duplicates limited?
Verified contact hashes, Telegram IDs and device or repeated-answer signals are used. These checks reduce repeated participation but do not guarantee an absolute one-person-one-profile rule.
8. Do minors participate?
Surveys are for adults aged 18+. Age is confirmed on the consent screen; this is the participant’s declaration, not a government-ID check.
9. What are the panel size and coverage?
Panel availability is checked by region and quota for each project. The ability to arrange fieldwork across Uzbekistan’s 14 regions does not mean the active online panel covers every region. Unconfirmed recruitment needs an assessment.
10. How are participants rewarded?
Points (1 point = 1 soum) for valid completes and free prize draws are described in the panel rules. Cash payouts and mobile top-ups are currently disabled; a points balance does not mean withdrawal is available.
11. Which targeting attributes are used?
Region and district, gender, age, settlement type, education, income, employment, languages and topic attributes. Availability of matching profiles is checked before ordering.
12. How do quotas work?
Configured quotas are enforced on the server; filled cells close. Quotas define the particular sample structure and do not by themselves establish representativeness.
13. How is the topic disclosed in invitations?
Invitations describe participation and available terms. Wording and reward disclosures need project-specific review: an invitation does not eliminate participant self-selection.
14. Is invitation frequency limited?
Panel recruitment includes a weekly limit and a pause between invitations. These rules apply to panel invitations, not every open link on the internet.
15. How is participation measured?
Invitation statuses and link-funnel stages are recorded. Participation rates must be read with their denominator: invited, opened and started are different bases.
16. Is there overlap with other panels?
We do not buy external panels. Participants may nevertheless belong to other panels; that overlap is not independently verified.
17. How is collection managed?
Pilot launches, split questionnaire forms, repeated waves and quota controls are available. The tools used depend on project settings.
18. How are suspicious responses detected?
Speed, repetitive patterns, attention, duplication and open-text quality are checked. Flags require interpretation and are not proof of deception on their own.
19. What is a valid complete?
A response that passes the applicable screening, quota, duplicate, behavioural and quality-review checks. Rules depend on the collection channel.
20. How is identity checked?
Verified contacts or Telegram are used. Consented recording and geotags support face-to-face quality checks; this is not passport identification of the respondent.
21. How is fraud handled?
Quality history, profile restrictions and source assessment help identify risks. A quality reviewer decides disputed interviews.
22. How is fieldwork checked?
Face-to-face collection uses audio, geotags, automated signals and sampled manual review or callbacks. Recording applicability, review plan and achieved coverage are project-specific.
23. Are responses linked across studies?
Panel participation is linked by an internal ID for quality history and deduplication. Profile erasure removes that link and direct identifiers; free text is also removed.
24. Are AI-generated responses detected?
Text heuristics and duplicates can trigger review flags. Reliable attribution of every text is not guaranteed.
25. Where is AI used?
In questionnaire preparation, open-answer processing, moderation and written findings. Echo separately generates model answers and must not be presented as human research. External processing flows are described on the data-protection page.
26. How is data cleaning disclosed?
Statuses and exclusion reasons are stored; quality reporting shows collected and valid responses, screening, quotas and flags. The specific report scope is agreed with the client.
27. How is personal data stored?
The live database is in Tashkent. Contact hashes support deduplication; a separately encrypted phone may be stored for callbacks, and email and Telegram IDs for invitations. Contacts therefore do not all exist only as hashes. Retention follows the shared matrix.
28. How is consent obtained?
Before the questionnaire, the organizer, processing and right to decline are disclosed; age 18+ is confirmed. Audio and geotags have separate consent. External open-text processing is disclosed before consent.
29. Which safeguards are in place?
HTTPS/HSTS, role-based access, logging of phone reveals and encrypted off-server backups. This does not claim formal security certification.
30. How can data be erased?
A panel member can erase their profile in the account area. Linked contacts, identifiers, media, transcripts and geotags are cleared; closed answers remain without a direct profile link. Mandatory records and backups follow separate retention-matrix rules.
31. What does an external processor receive?
Language-model provider Anthropic Claude may receive individual open answers, dialogues, questionnaires, briefs and materials, and aggregates and context for findings. Automatic de-identification across all flows is not ensured: personal data within free text may leave Uzbekistan. No per-project server block currently exists.
32. How are industry standards used?
ICC/ESOMAR, ISO 20252 and AAPOR are methodological references. This page is Xplorence’s self-description, not evidence of ESOMAR membership or ISO certification.
33. What is the jurisdiction?
Republic of Uzbekistan. The operator and current Xplorence legal details are available on the company page.
34. Which sample metrics are available?
Valid response counts, quota completion, effective base and weighting measures where the method and data permit. A margin of error does not remove recruitment bias; a human-survey margin does not apply to synthetic Echo answers.
35. What share of responses is excluded?
The share depends on the project and exclusion reasons. No universal percentage is promised; rules and calculation base accompany the result.
36. How is panel activity assessed?
From participation history, valid responses, invitations and quality indicators. A registered profile does not imply immediate availability for a survey.
37. Where can I obtain project-specific figures?
Request an assessment of the audience, quotas, available panel, additional recruitment and timing before ordering. Contact details are on the company page; a feasibility estimate is not achieved fieldwork.
Data protection and retention · Panel rules · Company details