SYS.PROD // IH-00LiveMarketing analytics

iHealth Marketing

A marketing-attribution platform that tells a nine-brand group exactly which spend actually worked.

Client
iHealth Marketing
Sector
Marketing analytics
Live at
ihealth.marketing
Year
2023

The business

The internal analytics and attribution platform for a group of US direct-to-consumer health brands, consolidating advertising, storefront and lifecycle data across dozens of platforms into one reconciled view.

The brief

Built from first principles and load-bearing two and a half years on. It resolves customer journeys at 98% accuracy across an intricate multi-channel funnel — first-party cross-domain tracking, granular event attribution and identity resolution — then reconciles that against Shopify, Google Ads, Amazon, Meta, Klaviyo, Attentive and more. The unglamorous, real-money kind of analytics: not a dashboard that looks impressive, one that changes where the next dollar goes.

What we built

01

Identity resolution and cross-domain tracking

First-party tracking that follows a real person across domains and sessions, so a journey is stitched rather than guessed. This is the part that makes every number downstream trustworthy, and the part most tools skip.

02

Attribution that survives scrutiny

Per-channel order attribution across paid search, shopping, PMax, marketplace, social, email, SMS, reviews, organic and direct — reconciled daily, with the unattributed share shown rather than quietly redistributed.

03

Resilient ETL across dozens of platforms

Scheduled and event-driven pipelines that survive the APIs going down, changing shape, or silently backfilling. Reconciliation checks prove both sides agree instead of assuming it.

04

Cohort and new-customer reporting

Month-over-month new-customer economics per brand — acquisition cost, order value, conversion rate and repeat behaviour — over a multi-year window, which is the horizon these decisions are actually made on.

05

AI insights layer

An assistant sitting on the warehouse that summarises trends, flags anomalies and raises priority alerts in plain language, so the daily read does not depend on someone remembering to open the right tab.