Performance Radar, operated by Project Moneyball (Avner Baruch), scores sales and customer-call transcripts. We built the product around data minimization: we process transcripts to score them, but we do not store the full transcripts.
Who is responsible for your data
For data about your organization's own calls, you (our customer) are the data controller and we act as your data processor — we process that data only to provide the service and on your instructions. For your own account information, we are the controller.
What we collect
Account & identity
Your email address, received from our identity provider (Cloudflare Access) and used as your account identifier.
Content you submit
Call transcripts you upload or send for scoring. Transcripts may contain personal data about your representatives and the other call participants.
Configuration you create
Scorecard templates and settings, stored under your account.
Usage & security data
Counts of how often the service is used (stored against a hashed email) and technical logs needed to operate and secure the service. Logs never contain transcript content or model output.
How we handle transcripts (the important part)
- A transcript is processed in memory for a single scoring run.
- It is sent to our scoring sub-processor, Anthropic, to analyze the call.
- We do not store the full transcript at rest, and we do not log its content.
- We retain only the derived analysis — scores, confidence, short verbatim excerpts used as supporting evidence, summaries, and recommendations. This is cached in Cloudflare KV, keyed to a one-way cryptographic hash of the transcript, for up to 30 days, so re-running or adjusting weights is instant and doesn't reprocess the call. You can request a fresh run that ignores and overwrites this cache.
Sub-processors
| Sub-processor | Purpose | Data involved |
|---|---|---|
| Cloudflare | Hosting, database, key-value storage, identity/access | Account email, templates, derived-analysis cache, usage counters |
| Anthropic | AI scoring of the call | Transcript content (processed for analysis) |
| Transactional email provider (added when the email feature launches) | Delivering report emails | Recipient emails, report content |
We require each sub-processor to protect data under terms no less protective than this notice. Transcript data sent to Anthropic via the API is not used to train their models.
Retention
- Full transcripts: not stored.
- Derived-analysis cache: up to 30 days, then automatically deleted.
- Scorecard templates: kept until you delete them or close your account.
- Usage counters: short-lived (rolling daily / per-minute windows).
- Logs: limited operational retention; contain no call content.
International transfers
Our sub-processors may process data in the United States and other countries. Where required, we rely on appropriate safeguards such as Standard Contractual Clauses.
Security
We protect data with encryption in transit (TLS) and at rest, identity-based access controls, prompt-injection input filtering, output validation, and rate limiting. See our Security Brief.
Your rights
Depending on where you live (e.g. GDPR, CCPA/CPRA), you may have rights to access, correct, delete, or port your personal data and to object to certain processing. For call content where our customer is the controller, please direct requests to them; we assist as their processor. For your own account, contact us below.
Cookies
We use a session cookie from our identity provider (Cloudflare Access) to keep you signed in. We do not use advertising cookies.
Children
The service is for business use and is not directed to anyone under 16.
Changes
We'll post changes here and update the date above; material changes will be communicated to customers.