Privacy Policy
Last updated: November 14, 2026
Who we are
Triumph Interactive, Inc. ("Triumph," "we," "us," or "our") provides Triumph Analytics, a web analytics service for businesses. This policy covers our website, customer accounts, and the analytics our customers ask us to collect on their sites.
For visitor analytics, the customer decides what to collect and why. We process that data to provide the service. We also handle account, sign-in, billing, and support information to run Triumph Analytics. We do not sell personal information.
Each customer is responsible for its own websites, notices, and legal duties. Visitors of a customer's site should contact that site's operator. Our tools are not legal advice and do not guarantee that a customer's setup is lawful.
What we collect
Account information. Name, email, company and website details, a password hash, optional encrypted authenticator secrets and recovery-code hashes, sign-in records, and billing records. Sign-in uses an emailed magic link or a password, with optional authenticator codes. We do not offer Google sign-in. Card-number and CVV entry uses processor-hosted fields, so those values do not pass through our servers. ACH account and routing numbers are entered in our page's bank form and sent by the browser to Stax for tokenization, not entered in processor-hosted fields. We keep payment tokens and safe display details, not card numbers or CVV.
Visitor identifiers. A first-party visitor cookie, a session cookie, and an email hash only if the site operator sends one through the identify API. We use the IP address for approximate location and rate limiting. The analytics event does not store the raw IP address. It does store location fields derived from that address. Unless Global Privacy Control (GPC) is on, or the site has IP anonymization on, we also store a server fingerprint derived from the IP address, browser, and language.
What a raw IP is not. We do not put the raw IP address in the analytics event. A raw IP can still appear in short-lived rate-limit keys and in operational access logs (load balancer, proxy, and similar logs), for the retention periods below. GPC does not erase those operational copies.
Internet activity. Pages, referrers, on-site search terms, clicks, form events, downloads, scroll depth, and outbound clicks, as configured by the site.
Location and network. When a request comes through our Cloudflare edge with authenticated location headers, we store country, region, city, postal code, coordinates, timezone, and the network's autonomous system number (ASN). We do not use a local IP-location database and we do not collect device GPS. Direct requests without those headers, including Measurement Protocol calls that override the IP, are stored with unknown location and ASN. Under GPC we drop coordinates and postal code and keep coarser location and ASN.
Device. Browser, operating system, screen size, device type, and language. Unless GPC is on, the tracker also collects canvas rendering, WebGL vendor, renderer and capabilities, an audio-processing value, detected fonts, timezone, CPU core count, device memory, touch points, display and browser capabilities, media preferences, math-rendering values, storage availability, and network type and connection estimates. These signals are sent on the first pageview of a site session and used to compute a device fingerprint for visitor matching. This device fingerprint is separate from the IP-derived fallback described above.
Where data comes from
Visitor browsers, customer API and Measurement Protocol calls, customer imports, authenticated Cloudflare location headers, and account or billing forms. Customers can also send custom fields and ecommerce data. Collection is not limited to the automatic tracker.
Why we use it
We use visitor data to produce the reports, profiles, and attribution the customer asked for. A customer can also forward conversions to Meta, Google Ads, or TikTok. GPC-marked events are left out of that forwarding. We use account and billing data to run accounts, take payment, keep the service secure, and send service messages. We do not use one customer's visitor data to advertise to another customer's visitors, and we do not combine visitor data across customers.
Do Not Sell or Share My Personal Information
We do not sell personal information. We do not share it for cross-context behavioral advertising. Service providers listed below receive information only to run the service. A customer can turn on conversion forwarding to Meta, Google Ads, or TikTok. That forwarding is the customer's instruction, and GPC-marked events are excluded. To opt out of sale or sharing on a site that uses Triumph, turn on GPC and contact the website operator. GPC reduces what we keep and stops conversion forwarding. It does not switch off all analytics. There is no separate opt-out cookie banner. For a Triumph account, use our contact form.
Your data, export, and account deletion
Customer analytics data belongs to the customer. While the account and the data are still available, you can export through CSV, Excel, JSON, and the API, subject to the plan and a 100,000-row limit on each raw JSON request. Large CSV and Excel exports can run in Amazon S3. Those export files expire after 30 days by default (you can set 1 to 365 days). Download links last 24 hours by default and can end sooner. Seven days is the longest a download link can be signed for, not the default. Visitor erasure revokes overlapping async export objects and deletes those accessible objects before completion. Copies already downloaded by the customer are not recalled. Noncurrent versions and backups keep the limits stated here.
After a trial expires we may suspend the account. Analytics cleanup becomes eligible 30 days after that suspension, or 90 days after any other suspension or cancellation. The job checks eligibility again, deletes events, visitor identities, and profiles in batches, refreshes the affected reports, and marks itself complete only after that purge succeeds. The eligibility date is not a promise that deletion finishes that day. Account, billing, and audit records can remain. A finished analytics purge cannot be undone by turning the account back on.
California privacy rights
If you are a California resident, you can ask us to:
- Know the categories and the specific pieces of personal information we have about you, within the limits of the available records.
- Delete personal information, as described below.
- Correct inaccurate personal information we hold about your account.
- Opt out of sale or sharing by using GPC and contacting the website operator, as described above.
- Not be discriminated against for using these rights.
How visitor deletion works. Contact the website operator. An account administrator must turn on account-wide API erasure for the site that submits the request. The secret-key API requires deletionScope=tenant and rejects a website-only delete. We process each visitor id you send across every website in that customer account. Send each known visitor id separately. A 202 response means the work is queued, not that erasure is finished. COMPLETED means a bounded check of the events, identity records, profiles, and affected aggregates that existed at request time, plus overlapping async-export revocation and an ingestion-watermark check. It is not a statement that every copy of a person is gone. Ingest retries are limited to 90 minutes, and we recheck once in a short follow-up window. Collection after the request can create new records. This API does not delete backups, noncurrent export versions, operational logs, copies already downloaded by the customer, or billing and audit records.
Account requests. Customers can use our contact form. We will respond within 45 days. We may need information to verify you. You may use an authorized agent. We may ask the agent for written permission and may ask you to verify your identity directly.
GDPR and other privacy laws
The Service is offered from the United States and is governed by United States law. Customers anywhere may use it. Each customer is solely and fully responsible for compliance with the laws of that customer's own jurisdiction, including the GDPR and UK GDPR, and for notices, consent, and lawful basis. We do not appoint an EU or UK representative, we do not offer Standard Contractual Clauses, and we do not provide a cookie-consent banner. For visitors of a customer's website, that customer decides the purpose of the analytics. Contact that website. We will help the customer with access, deletion, and export requests we can actually perform. Account holders can ask, through the contact form, for access, correction, deletion, or a copy of the account information we hold. Visitor data may be processed in the United States and, for delivery and outage queueing, on Cloudflare's global network.
Service providers
We use the following providers. The Data Processing Agreement has the same list.
- Amazon Web Services (us-east-1, Virginia). Application and database hosting, networking, cache, S3 exports and backups, logs, and monitoring. Amazon SES sends transactional email and scheduled reports over SMTP.
- Cloudflare, Inc. Content delivery and TLS for customer tracking hostnames (Cloudflare for SaaS), script delivery, IP-derived location headers (country, region, city, timezone, and ASN), the outage event queue, and R2 replay receipts. Processing uses a US and global edge network, not US-only storage.
- Stax. Card and ACH processing and the token vault for customer billing. Live hosted card-number and CVV fields use BlockChyp at
https://api.blockchyp.com. Those fields receive what you type directly, without passing card numbers or CVV through our servers. They do not receive visitor analytics. - Customer-site tools. Meta, Google Ads, and TikTok can receive conversions the customer configures. For non-GPC visitors, automatic video tracking can load YouTube and Vimeo SDKs and enable their player APIs when those embeds are present on a customer's page; it does not require a separate video-tracking opt-in. Those providers receive the resulting browser requests. They are not required when the customer's page has no such players.
Cloudflare terminates TLS and forwards tracker and event requests to our origin. We do not use Cloudflare challenges or Cloudflare cookies on those tracker paths. The browser's GPC signal is forwarded and honored as described above. On our account signup, sign-in, password-reset and billing access forms, we use Cloudflare Turnstile to prevent automated abuse. Cloudflare processes browser and network signals for this security verification. Turnstile is not loaded into customer tracking scripts.
Cloudflare Turnstile provides abuse protection on public signup and protected account or billing forms where enabled. It can process IP addresses, browser and device signals, challenge results, and verification tokens on Cloudflare's global network. These form checks are separate from visitor tracking and do not make storage or processing exclusively US-based. Stax may also load its own fraud-prevention services, including Google reCAPTCHA within a processor-controlled frame; our choice of the plain Stax SDK does not control those provider settings.
If the origin is down and the outage buffer is on, Cloudflare may hold queued event payloads, including URLs, referrers, identifiers, and IP addresses, for up to 14 days in the main queue and up to another 14 days in the dead-letter queue (28 days total under the configured retention). A queued 202 is a positive ACK and the tracker stops retrying. Eligible events are replayed into our database, and queue messages are deleted after confirmed processing. After that ACK, the only permitted terminal refusals are visitor erasure completed before replay, an event that is more than 72 hours old when it reaches our origin, a site that does not exist, and an event that live intake would itself have refused as invalid. The erasure outcome is equivalent to storing and then erasing the event. The event-age outcome is the documented outage-window SLA exclusion. Other replay failures remain retryable. A queued 202 is a positive acceptance ACK, but the event is saved only after origin persistence. The queue's longer retention does not extend the live 72-hour window. GPC-flagged events are queued like other events, and their opt-out is kept and honored on replay. Cloudflare R2 stores replay receipts (tenant, website, event id, time, and status) for up to 29 days so we can reconcile the queue. Those receipts are not the analytics record.
We will update this page, and email the account address, at least 30 days before a new provider starts processing customer personal information. An emergency security replacement may start sooner. We will disclose it as soon as we reasonably can. You may object by cancelling.
Security
We use TLS in transit, encrypted storage at our cloud provider, access controls, hashed passwords (BCrypt), and monitoring. No method of transmission or storage is perfectly secure.
How long we keep data
Live tracker, server API, and Measurement Protocol events must be no older than 72 hours when they are saved at our origin, not merely when they are received or queued. Older events are rejected with event-too-old. They are not stored as analytics and they are not billed. For an event already ACKed by the edge, that age rejection is the documented outage-window SLA exclusion and is one of the four permitted post-ACK refusals named above. Times up to five minutes ahead of receipt are accepted as sent, subject to other checks. Times more than five minutes and up to 24 hours ahead are stored at the receive time. Times more than 24 hours ahead are rejected with event-in-future and are not stored or billed. Authenticated Google Analytics import is the historical-data path. Imports are not billed and are not subject to the 72-hour window. Plan limits still apply.
Usage counts deduplicate retries that keep the same event id and time. A retry with no stable id can be counted more than once. Per-event billing fingerprints are removed by whole UTC day, during hourly maintenance, after that day's last timestamp has expired. That cleanup can take longer than 72 hours. It does not extend the 72-hour acceptance window. Older billing disputes use monthly and daily totals and verification digests, not a permanent copy of every event. Imports and supported retroactive settings can still change historical reports.
Analytics retention is 14 months on Starter, 26 months on Growth, and 38 months on Pro, Business, and Enterprise. A paid extension can add 12 months on Starter, Growth, Pro, and Business. Enterprise longer or unlimited retention is a contract term. A website can choose a shorter period. Removing a paid extension lowers the ceiling immediately. Up to 12 months of older history then becomes eligible for permanent deletion on the next normal retention run. Export first. Adding the extension again does not restore deleted history. Scheduled retention deletes eligible events only when its safety checks allow it. Otherwise it keeps them and reports the condition. These periods cover events, not every identity, profile, account, billing, export, log, or backup copy.
Backups cover both databases in our PostgreSQL cluster. We target weekly full backups, daily differentials, and hourly incremental backups. Retention keeps four full backups and fourteen full or differential backups; these are backup counts, not fixed day limits. Four weekly full backups normally span about 21-28 days from the oldest backup's start. That is not a hard erasure deadline. Backup chains, a 28-day storage lock, and version expiry can keep a copy longer. Deleted live data can remain in backups until those copies expire. If we restore a backup, data that had been deleted from the live system can return and must be deleted again. Operational logs have their own windows: streamed application logs target seven days; security, flow, and disk-audit logs about 30 days; load-balancer logs about 90 days plus seven days for old versions; and CloudTrail about 365 days plus 90 days for old versions. We keep billing, tax, and dispute records for seven years after the charge, and longer if a dispute, audit, or law requires it. Visitor deletion and analytics purge do not erase those records.
Children
The service is not directed to children under 16. We do not knowingly collect personal information from children under 16, and we do not sell or share personal information of consumers we know to be under 16. If you believe a child has provided personal information, contact us and we will delete it.
Changes
We may update this policy by posting a new version and changing the date above. If a change materially reduces your rights, we will also email the account address.
Contact
Questions, California requests, and account privacy requests: our contact form, or by mail:
Triumph Interactive, Inc.
6805 Dwight Rowland Rd
#431
Willow Spring, NC 27592
Phone: (919) 591-0923