Email data, finally in the same room as everything else.
Conduit is a productized implementation from CdpHub's consulting practice. We stream every send, open, click, and bounce from Salesforce Marketing Cloud into Adobe Analytics with stitched identity — so email stops being a silo and starts being part of the cross-channel story.
The problem we solve
Email reporting lives in one tool. Everything else lives somewhere else.
Most enterprises running SFMC and Adobe settle for nightly FTP drops into Adobe's Data Sources. It works. It also quietly costs them the attribution, personalization, and cross-channel visibility that made them buy Adobe in the first place.
FTP Data Sources
Nightly file drops, manual reconciliation, fragile joins.
- Email events arrive 4–24 hours after they happen
- SFMC subscriber keys don't stitch to Adobe visitors — every join is manual
- Transfer failures detected hours late, with no automatic replay
- Email under-credited in multi-touch attribution models
- Every new report is a three-week ETL project
Conduit
API-native ingestion. Identity stitching. Self-healing reliability.
- Events in Adobe within minutes, via the Bulk Data Insertion API
- SubscriberKey mapped to customerID for cross-channel journey stitching
- Idempotent retries, dead-letter queue, high-water-mark state
- Email participates in Adobe's native attribution like any other channel
- New reports take an afternoon in Workspace — not a quarter of ETL
What Conduit delivers
Three capabilities, one pipeline.
Most email integrations ship data. Conduit ships analytical capability — the kind that shows up directly in your Workspace reports and your attribution models.
Identity stitching that actually stitches.
SFMC's SubscriberKey becomes Adobe's customerID, so every email send, open, and click lands on the same visitor profile as the web and mobile behavior that came before and after it. One person, one journey, every channel.
Attribution that credits email properly.
Once email events live in Adobe's hit stream, they flow through every attribution model the same way paid search and organic do. The gap between last-touch and multi-touch often reveals that email is doing 3–4× more work than your reports suggested.
Near real-time or batch. Same codebase.
One pipeline, two speeds. Your team runs it as a cron job for daily reporting when latency doesn't matter, or as a service for journey orchestration that needs current-state email engagement. Swap modes with a config flag — no second build.
The engagement
How we deliver Conduit.
A four-phase engagement that takes most clients from signed SOW to first production batch in 4–6 weeks, with CdpHub's consulting team embedded alongside your engineering.
Audit your SFMC and Adobe instances. Map SubscriberKey governance, inventory eVar and event slots, identify identity stitching prerequisites.
Define the SFMC-to-Adobe variable contract. Sign-off from your Adobe admin and SFMC admin on the mapping spec before any code ships.
Deploy the pipeline to your cloud environment. Wire observability, run parallel against existing FTP for validation, cut over when parity is confirmed.
Build the first three Workspace reports with your analytics team. Document the runbook. Hand off to your operations team with optional ongoing support.
We built Conduit because we kept implementing the same pipeline at every enterprise client. Productizing it means our consultants spend engagement time on the hard parts — identity governance, attribution modeling, report design — instead of re-writing the same OAuth lifecycle and retry logic.
What clients build
Reports that weren't possible before. Now they're an afternoon.
Once email events are in Adobe with stitched identity, a whole category of Workspace projects opens up. Here's a selection of what our clients build in the first 90 days after go-live.
Campaign leaderboard by revenue per send
Ranks every SendID by RPE, not open rate. The metric that actually drives budget decisions.
Email-influenced vs. email-only conversions
Segments buyers by email involvement. Usually reveals email is 3–4× more impactful than last-touch suggests.
Channel contribution dashboard
Email's share of credited revenue across linear, time-decay, and algorithmic models — over time.
Engagement tier segmentation
Hot / warm / lukewarm / cold, based on combined email and behavioral data. Pushes back to SFMC.
Journey performance by stage
Welcome series stage 1 vs. 2 vs. 3 — measured against actual site behavior, not just opens.
LTV by acquisition × email engagement
Customers who become email-engaged within 30 days typically have 2–5× the lifetime value.
Under the hood
A pipeline built to stay up.
Conduit is five stages, each designed around the idempotency and recovery guarantees that production data pipelines actually need — not the happy-path flow that demos well and breaks at 3am.
Idempotency keys on every BDIA request. High-water-mark state advances only on full success. Failed batches route to DLQ and replay on a cooldown — no operator intervention required.
OAuth 2.0 client-credentials on SFMC, JWT or OAuth S2S on Adobe. Secrets in your vault, never on disk. No PII beyond SubscriberKey leaves your perimeter — email addresses stay in SFMC.
Per-event-type counters, end-to-end latency, DLQ depth, token health. Emit as EMF, StatsD, or Prometheus — alerting on freshness, not just uptime.
