Unlocking Server-to-Server Personalization for a Fortune 500 Enterprise
A Fortune 500 tax preparation firm launched tamper-proof server-side pricing experiments — with durable identity, full Adobe Analytics reporting, and zero new infrastructure.
The challenge
The firm's digital team needed to A/B test product pricing on high-traffic landing pages using Adobe Target. Two fundamental constraints made client-side delivery unacceptable.
First, security: any personalization delivered via JavaScript is visible and modifiable by the end user. Browser developer tools allow inspection and alteration of DOM content, JavaScript variables, and network responses. For a pricing experiment, this is a disqualifying risk — test values cannot be exposed in the browser.
Second, identity: modern browser privacy policies (Safari ITP, Firefox ETP) limit client-side cookies to a maximum lifespan of 7 days. A visitor who returns after 10 days is treated as new, re-bucketed into a different test group, and potentially shown a different price. This breaks both test validity and customer trust.
Building server-side personalization internally would require provisioning infrastructure, engineering an integration with Adobe Target's API, implementing durable identity, and coordinating across development, IT, and security teams — months of effort before a single experiment could run.
The solution
Ingest Labs' data collection platform was already deployed at the edge of the firm's web properties, operating as a first-party data collection layer. Because the platform was already in the request path, already scaled to traffic volume, and already managing first-party identity, server-to-server personalization became an activation on existing infrastructure.
The platform's identity solution (MPID) sets a first-party, server-side-only cookie with an approximate two-year lifespan — across all browsers, including those with aggressive ITP policies. MPID is passed as a First-Party ID (FPID) to Adobe's Edge Network, anchoring Target's traffic allocation. Same visitor, same experience, every visit.
The entire personalization flow operates before the page reaches the browser: the visitor request arrives, MPID is resolved, Adobe Target is called server-to-server, and personalization decisions are injected into the response. The visitor's browser never makes a personalization request, never receives raw experiment data, and never has access to the test logic.
A two-datastream architecture preserves Analytics for Target (A4T) reporting without modification — the server-side datastream fetches Target decisions only, while the client-side datastream sends the analytics event with a display notification that stitches the decision back to Adobe Analytics.
Results
| Metric | Detail |
|---|---|
| Time to production | Days from initiation to live server-side pricing experiments |
| Infrastructure provisioned | None — activated on existing deployment |
| Identity durability | ~2 year cookie lifespan vs. 7-day ITP limit |
| Security posture | Pricing data never exposed to client-side JavaScript |
| Vendor migration impact | Zero — AT.js to Web SDK migration absorbed by platform |
| Analytics continuity | Full A4T reporting preserved via two-datastream architecture |
| Visitor consistency | Deterministic test bucketing via MPID/FPID — same visitor, same price, every visit |
Beyond a single experiment
What was built is not an experiment — it is a server-side personalization capability that scales horizontally. Adding a new experiment requires no infrastructure work, no new deployments, and no engineering coordination. The marginal cost of each additional experiment is configuration, not engineering.
The same pipeline that delivers a pricing test can simultaneously deliver homepage content variations, promotional offers, feature flags, or any other JSON-based personalization — all through the same request, the same identity, and the same analytics pipeline.
An enterprise unlocked a security-sensitive personalization use case in days, with no new infrastructure, no vendor lock-in, and full measurement continuity — and gained a platform that scales with their ambitions.
Ready to run server-side experiments?
Tamper-proof personalization, durable identity, and full analytics — live in days.