predictedby.me
MethodsHow it worksCompareFAQsInvestors
Book a demo

Trust centre

Data & Privacy

Privacy notice, retailer data processing and Attention Studies.

On this page

  • Privacy notice
  • Retailer DPA Outline
  • DPIA – Attention Studies

Privacy notice

Document date: 3 November 2026

Roles assumed: the retailer running a study is the controller of respondent data; PREDICTEDBY.ME LTD (predictedby.me) is its processor, and the controller for retailer user accounts and our website.

1. Who we are

PREDICTEDBY.ME LTD, registered office 66 Paul Street, London, England, EC2A 4NA, company number 17558239, ICO registration [NUMBER]. Contact: hello@predictedby.me

When you answer a survey for a retailer, that retailer decides why your answers are collected and is responsible for them. We run the survey platform on its behalf.

2. What we collect

2. What we collect
DataWhenWhy
Account details (name, email, password hash)You sign upRun your account and sign you in
Survey answers, swipes, ratings, free textYou take a surveyGive the retailer the results
Voice or video interview recordings and transcriptsOnly in voice or video surveys you choose to joinGive the retailer your spoken answers
Attention study numbers (below)Only in attention studies you agree toShow which parts of a picture or advert people look at and how they react
Device information (browser, screen size)Any surveyMake the survey work; check calibration is still valid
Reward and credit recordsWhere a reward appliesPay or record rewards

3. Attention studies – what happens on your phone

If you agree, your phone camera is used while you look at a picture or advert. The camera image is analysed on your phone only. No video or photo is recorded, saved or sent to us.

Only these numbers leave your phone:

• how many moments your gaze was in each of nine areas of the screen (an approximate estimate, not exact eye tracking);

• a timeline of facial movement signals, such as smile strength or brow movement;

• a calibration accuracy percentage;

• your answer to any follow-up question, and basic device information.

We do not use this to identify you, and we do not draw conclusions about your health, personality or emotions as facts. Results are estimates and are shown to the retailer as combined patterns.

To avoid repeating set-up every time, your phone keeps a small calibration file in its own browser storage (valid for 180 days, same device only). We keep a simple record that you completed calibration and your accuracy score. The calibration file is never uploaded.

4. Our lawful basis

• Attention studies: your explicit consent (UK GDPR Art 6(1)(a) and, as a precaution, Art 9(2)(a)). You can say no and still use the app. You can withdraw at any time by closing the study or contacting us.

• Other surveys and accounts: contract (Art 6(1)(b)) and the retailer's legitimate interest in market research (Art 6(1)(f)).

• Voice and video interviews: consent.

5. Who receives your data

The retailer who asked you the questions. Our service providers, who only act on our instructions: [If required you can ask to see the sub-processor list in the records-of-processing document]. We do not sell your data.

The attention page loads software from two third-party servers (jsdelivr and Google Cloud Storage), so those servers see your IP address and browser details when the page loads.

6. International transfers

Some providers are outside the UK. Where so, we rely on the UK adequacy regulations, the UK International Data Transfer Agreement or the UK Addendum to EU standard clauses. [Providers outside of the UK are listed our sub-processors document]

7. How long we keep it

• Attention study numbers and calibration records: 180 days after the study closes, then deleted automatically.

• Survey answers: are kept for up to 12 months or shorter as defined by the retailer..

• Account data: while the account is open plus [6] years for accounting where needed.

8. Your rights

You can ask to see, correct, delete, restrict or move your data, to object to processing, and to withdraw consent. Contact hello@predictedby.me ; we reply within one month. If the retailer is the controller we will pass your request on. You can complain to the Information Commissioner's Office (ico.org.uk, 0303 123 1113).

9. Age

Attention studies and surveys are for people aged 18 or over. We do not knowingly collect data from children.

10. Consent wording for the phone screen

We would like to use your camera to estimate which parts of the picture you look at and how your face reacts. Your video is analysed on your phone and is never recorded or sent. Only simple numbers are sent. You must be 18 or over. You can stop at any time.

Retailer DPA Outline

Part A – Data Processing Agreement outline

Clauses

Clauses
ClauseWhat it should say
1. Parties and rolesRetailer is controller; PREDICTEDBY.ME LTD (company number 17558239, 66 Paul Street, London EC2A 4NA) is processor for survey and attention data. We are controller for the retailer's user accounts.
2. Subject, duration, natureRunning surveys and attention studies for the term of the service; types of data and subjects as in the schedule
3. InstructionsWe process only on documented instructions (the retailer's use of the platform and settings); tell the retailer if an instruction breaches the law
4. ConfidentialityStaff and contractors are bound by confidentiality
5. SecurityMeasures in Schedule 2; reviewed as the service changes
6. Sub-processorsGeneral authorisation for the listed sub-processors; [30] days' notice of changes; retailer may object; we flow down equivalent terms and stay liable
7. International transfersOnly with a valid mechanism (adequacy, IDTA or Addendum)
8. Data subject requestsWe help the retailer respond; we forward requests we receive; response within [5] working days
9. BreachWe notify affected customers without undue delay and within 48 hours of becoming aware of a personal data breach affecting their data, with the facts needed for the retailer's ICO duty
10. DPIA and consultationWe provide information for the retailer's DPIA, including for attention studies
11. Retention and deletionStudy data deleted by retention setting; on termination, delete or return within 30 days; backups expire in 365 days
12. AuditProvide information to show compliance; audits once a year, with notice, in business hours
13. Retailer dutiesLawful basis and notice to its customers; adult-only respondents; use attention studies only with the consent screen intact
14. Liability and lawEngland and Wales

Schedule 1 – Processing details (draft)

Draft attachment to the Retailer DPA Outline, assembled from the Privacy notice and Attention Studies DPIA. It requires review against the service’s actual operation and the retailer’s instructions before it is agreed.

This schedule covers personal data processed on the retailer’s behalf. Retailer user account data, for which predictedby.me acts as controller, is outside this processor schedule.

Schedule 1 – Processing details (draft)
ItemDraft processing details
Parties and rolesThe retailer commissioning the study is controller. PREDICTEDBY.ME LTD (company number 17558239, 66 Paul Street, London, England, EC2A 4NA) is processor. [CONFIRM: retailer’s legal name, address and contact details in the agreement].
Subject matter and purposeOperating surveys, voice/video interviews and attention studies on the retailer’s behalf, collecting participant responses and producing research results for the retailer.
DurationFor the term of the service, subject to the retention periods and termination arrangements below.
Nature of processingCollection, recording where applicable, organisation, storage, analysis, presentation of research results to the retailer, and deletion or return, on the retailer’s documented instructions through its use of the platform and settings.
Categories of peopleAdult survey and study participants, including customers of the retailer. [CONFIRM: any other participant categories, including externally recruited panel members].
Survey and interview dataSurvey answers, swipes, ratings and free text; voice or video interview recordings and transcripts where participants choose to join. [CONFIRM: participant identifiers and any demographic, profiling or other personal-data fields collected for each study].
Attention study dataGaze counts across nine approximate screen regions, facial movement signal timelines, calibration accuracy percentage and completion record, optional follow-up answers, and basic device information. Camera images are analysed on-device; video, photos, face images, landmark data and the local calibration model are not sent to or stored by predictedby.me for attention studies.
Device and reward dataBrowser and screen-size information; reward and credit records where rewards apply. [CONFIRM: which reward records are processed on the retailer’s behalf, the fields involved, and any separately controlled payment or accounting records].
Sensitive data and safeguardsAttention signals are not used to identify people or presented as factual conclusions about health, personality or emotions. The supplied documents propose explicit consent, including an Article 9 precaution. [CONFIRM through legal review: special-category treatment and whether interview or free-text responses may contain special-category data; agree permitted collection and safeguards].
Attention data retentionAttention study numbers and calibration records: 180 days after the study closes, then automatic deletion. The calibration file remains only in the participant’s browser, valid for 180 days on the same device; it is never uploaded.
Survey, interview and reward retentionSurvey answers: up to 12 months, or a shorter period defined by the retailer. [CONFIRM: retention periods and starting points for interview recordings, transcripts, device information and reward records; do not assume the survey-answer period applies to every category].
Termination and backupsOn termination, delete or return study data within 30 days as set out in the DPA outline; backups expire in 365 days. [CONFIRM: return process and any legally required retention exceptions].
Instructions, recipients and transfersProcess only on the retailer’s documented instructions. Provide research results to the retailer and use authorised sub-processors subject to the DPA. International transfers require a valid mechanism as described in the Privacy notice and DPA. [CONFIRM: final sub-processor register, processing locations and applicable transfer arrangements].

Schedule 2 – Security measures (draft)

Draft attachment supplied by predictedby.me. These statements describe the supplied security measures and require review against actual platform operation and provider configuration before agreement. Cryptographic standards, hosting locations and compliance arrangements below have not been independently verified.

The supplied reference to penetration testing and Cyber Essentials is a plan for 2026, not evidence of completed testing or certification. EU region pinning and regional routing capabilities do not establish that those configurations are enabled.

Schedule 2 – Access controls

• Each customer's data is separated at the database level. Every request is scoped to the signed-in retailer, and database row-level security is switched on for our tables. Customers can't query each other's data.

• The database service key is held only in the server-side worker and is never sent to a browser or app.

• Sign-in uses expiring sessions. Passwords are stored only as salted PBKDF2 hashes, and session tokens are stored as SHA-256 hashes, not in readable form.

• Super admin access is a separate login and session type, used only by named PREDICTEDBY.ME staff. Staff access is limited to what their role needs. MFA on admin and infrastructure accounts

• Requests are rate-limited, and the platform sets standard browser protections, including HSTS, a content security policy and frame blocking.

• Spend and service health are monitored, so unusual usage is flagged.

Schedule 2 – Encryption

• Data in transit uses TLS (HTTPS) across the app, the API and every sub-processor connection.

• Data at rest in the database and file storage is encrypted by our hosting providers (Supabase and Cloudflare).

Schedule 2 – Encryption reference table

Schedule 2 – Encryption reference table
ScopeData PhaseVendor / Sub-processorCryptographic Standard / Protocol
User Data TrafficIn TransitCloudflare EdgeTLS 1.2 / TLS 1.3 & Post-Quantum X25519MLKEM768
Edge Compute OperationsIn TransitCloudflare Workers to SupabaseTLS 1.3 (TCP/TLS Connection Pools)
Primary Database StorageAt RestSupabase (AWS Underlying Infrastructure)AES-256 (Keys isolated via FIPS 140-2 HSMs)
System & Database BackupsAt RestSupabase Database BackupsAES-256
Edge Storage / CacheAt RestCloudflare D1 / R2 (if applicable)AES-256-GCM

Schedule 2 – Legal & Cross-Border Frameworks

• Supabase compliance artifacts: Both a signed Data Processing Addendum (DPA) and Standard Contractual Clauses (SCCs) are legally integrated into their terms to govern international data flows. Data can be pinned strictly to EU-based AWS regions (e.g., Frankfurt or Stockholm) to preserve regional compliance.

• Cloudflare compliance artifacts: Supported by the standard Cloudflare Security Exhibit and regional routing rules (Data Localization Suite) to limit where payloads are evaluated.

Schedule 2 – Minimisation and deletion

• Respondents are identified by an ID, not by name. Names and contact details are not sent to AI providers.

• Retention periods are configurable. For example, attention-study numbers are deleted automatically after a set period (default 180 days), and old question-log rows are deleted where a retention is set.

• On a customer's request, or when a contract ends, we delete or return customer data within 30 days. Individuals can ask us to erase their data at hello@predictedby.me. There is no self-serve delete button today, so this is a manual process

• Backups are kept for 365 days and then expire.

Schedule 2 – Breaches

• We keep an incident log and investigate any suspected breach straight away.

• We tell affected customers without undue delay, and within 48 hours of becoming aware of a personal data breach affecting their data. This lets them meet their own 72-hour duty to notify the ICO.

• If we are the controller, we notify the ICO within 72 hours where the breach is likely to risk people's rights, and tell the individuals affected where the risk is high.

• Our notice covers what happened, what data is affected, the likely impact, and what we are doing about it.

Schedule 2 – Sub-processors and governance

• Sub-processors are recorded in our sub-processor register, are bound by written agreements, and customers get at least 30 days' notice of changes.

• We review security measures annually. Penetration testing is planned for later this year 2026 as Cyber Essentials certification.

DPIA – Attention Studies

Document date: 3 November 2026

1. The processing

What: a respondent views a picture or advert on their phone. On-device software estimates which of nine screen regions they look at and facial movement signals. Only numbers are sent to our server.

Who: retailers (controller) commission studies; PREDICTEDBY.ME LTD processes. Respondents are adult customers of the retailer.

Data sent: gaze counts per region, expression signal timeline, calibration percentage, optional answer, device info. Not sent or stored: video, photos, face images, the landmark data, the calibration model (kept in the phone's browser for 180 days).

Scale: 100 respondents per study, 50 studies a year.

2. Why it is not biometric identification

The numbers cannot identify a person and are not used to recognise anyone. Special category treatment under Art 9 applies only to biometric data processed to uniquely identify. As a precaution we still obtain explicit consent.

3. Necessity and proportionality

3. Necessity and proportionality
QuestionAnswer
PurposeShow retailers which parts of a design get attention
Less intrusive alternativeClick or tap heatmaps and written questions; they are weaker, so the retailer may choose them
Data minimisationOn-device analysis; aggregates only; no images
ConsentSeparate, explicit, optional; refusal does not block the app
Retention180 days then automatic purge
TransparencyPlain consent screen and privacy notice
RightsDelete and withdraw by contacting us; records are keyed to the customer account

4. Risks and measures

4. Risks and measures
RiskLikelihoodSeverityMeasureResidual
Respondent does not understand camera is usedLowMediumConsent screen, 3-2-1 countdown, camera lightLow
Video or images leakVery lowHighNever recorded or uploaded; processed in memory on the phoneLow
Gaze/expression numbers treated as health or emotion factsMediumMediumWording "approximate regions" and "movement signals"; no emotion claims in reportsLow
Inaccuracy or bias across faces, skin tones, glasses, lightingMediumMediumCalibration quality check (default 20% limit); failed users excluded and not charged; accuracy testing so far on two faces onlyMedium – test wider before wide use
IP address exposed to third-party file hosts (jsdelivr, Google Storage)CertainLowDisclose in notice; or self-host filesLow if self-hosted
Data kept too longMediumLowDaily automated purge by retention settingLow
Retailer reuses data for another purposeLowMediumDPA purpose limits; tenant isolationLow
Under-18s take partLowHighAge statement on consent; retailers' customers adult-only; age gate requiring participants to be 18 or over before entering an Attention Study is in place (confirmed by predictedby.me).Medium
Pressure to consent (reward offered)LowMediumCamera participation is optional. Rewards are not offered if the camera is not switched on. Attention-study points are paid only to people whose eye set-up passes the quality check. [REVIEW: explain reward conditions before consent and assess whether the incentive affects freely given consent; the reward policy is confirmed, but consent safeguards require review]Pending review – assess incentive and freely given consent
Unauthorised access to resultsLowMediumAccount access control, row-level security per retailer, tokens in URL fragmentLow

Questions? hello@predictedby.me

predictedby.me

Consumer research platform: swipe testing, surveys, AI video interviews and ad testing on one panel, with predictive scoring.

© 2026 predictedby.me. All rights reserved.

PREDICTEDBY.ME LTD · Company number 17558239
Registered office: 66 Paul Street, London, England, United Kingdom, EC2A 4NA

Platform

  • Home
  • How it works
  • All research methods
  • Best market research platforms

Methods

  • Swipe testing
  • Surveys
  • AI video interviews
  • Video surveys
  • Ad testing
  • Feed test in TikTok
  • Concept testing
  • Packaging testing
  • Brand tracking
  • Attention Studies
  • Sales outcomes
  • Ask Ted

PREDICTEDBY.ME LTD · 66 Paul Street, London EC2A 4NA

TermsCookie PolicyData & Privacy