Treasure Data

Overview

Treasure Data is an enterprise Customer Data Platform (CDP) founded in 2011 and headquartered in Mountain View, California. Acquired by Arm Holdings in 2018, Treasure Data operates a large-scale cloud data warehouse built on Apache Hadoop and Presto, designed to ingest, unify, and activate behavioural data at petabyte scale. Enterprise organisations — particularly in retail, automotive, financial services, and consumer goods — use Treasure Data to build unified customer profiles from website behaviour, mobile app events, offline transaction records, and third-party data sources. The unified profiles are then used for audience segmentation, personalisation, and marketing campaign activation.

What This Script Does

Treasure Data's website SDK loads from cdn.treasuredata.com and initialises as td in the global JavaScript namespace. The SDK captures a comprehensive stream of behavioural events and sends them to Treasure Data's cloud ingestion endpoints.

Cookies and identifiers:

• td_uid — First-party persistent visitor identifier set on the host domain; created on first page load; expiry 2 years. This is the primary key used to stitch all behavioural events to a single visitor profile across sessions.
• td_ssc_id — Server-side cookie complement used when the td_uid is set via a server-side helper to extend effective cookie lifetime beyond ITP restrictions; expiry 1 year.
• _td — Secondary analytics identifier sometimes deployed alongside td_uid for session deduplication; expiry 30 minutes to 1 year depending on configuration.

Script filename and CDN: sdk.js loaded from cdn.treasuredata.com/sdk/latest/ or a version-locked path. Event data is POSTed to in.treasuredata.com/postback/v3/event/ using the host organisation's write API key.

Per-event data collected:

• Full page URL, referrer, and title
• UTM campaign parameters and traffic source attribution
• Click events with target element metadata (CSS selector, href)
• Form interaction events (field focus, submission — excluding sensitive field values by default)
• Page visibility and scroll depth
• Session duration and engagement score
• Device fingerprint components: user agent, screen dimensions, timezone, language
• IP address (used for geolocation and enrichment at the CDP layer)

Downstream activation: Collected data feeds Treasure Data's audience segmentation engine, enabling real-time segment membership evaluation and activation to downstream marketing platforms including DMP partners, advertising DSPs, and marketing automation systems (Salesforce Marketing Cloud, Adobe Campaign, Braze).

Consent & Compliance

Treasure Data is classified under the analytics category, though its downstream marketing activation capability means it often functions as a marketing tool. Under GDPR Article 6, the persistent td_uid identifier and behavioural event collection require a lawful basis. When data is used solely for internal analytics with no third-party marketing activation, legitimate interest may apply; when it feeds advertising audiences or personalisation, explicit consent is required. IAB TCF Purpose 1 (Store/access device), Purpose 7 (Measure content performance), and Purpose 9 (Apply market research) are relevant. Under ePrivacy, the td_uid persistent cookie requires prior consent. Treasure Data participates in the EU-US Data Privacy Framework. Under CCPA/CPRA, if behavioural data is shared with advertising platforms for audience targeting, this may constitute a "sale" or "sharing" of personal information.

Should You Block This Without Consent?

Conditional. If Treasure Data is used solely for first-party internal analytics, a legitimate interest argument may be sustainable with appropriate balancing tests. If data feeds third-party marketing activation or audience sharing, block until the user provides explicit analytics and/or marketing consent.