From “channel” to the Customer Operating System
- Team Uniquon
- 3 set
- Tempo di lettura: 3 min
Sector: B2B services - Europe.
Starting point
January. A multi-country European group asks us to “re-think the interactions with customers”. Three working sessions with the CFO, CMO and CIO surface the real issues: slow quotations, fragmented onboarding, expensive support. The Board approves a single North-Star Metric: “Tasks completed first time” across three critical journeys — quotation, activation, and ticket resolution.
“We don’t want pageviews. We want customers finishing what they start.” — CFO
Diagnosis (10 days)
Event-based tracking revealed as-is gaps; three separate CMSs driving shadow IT; personalisation that wasn’t explainable; fragmented consent. Core Web Vitals were inconsistent — LCP above 4s for 40% of mobile sessions. Digital cost-to-serve was higher than forecast because half of sessions diverted to assisted channels.
Uniquon’s intervention (12 weeks)
Experience by task, not layout. We redesigned around jobs to be done: a single guided quotation page; stepwise onboarding with save-and-resume; ticket resolution that serves product-signed answers (not generic FAQs) with transparent handover to agents only when needed.
Composable stack. Headless CMS and a shared Design System; services exposed as reusable modules via APIs. A central Data & Consent Layer handles consent, identity and a real-time feature store. No superfluous data — minimisation and explainability by default.
Observable decisions. A journey-orchestration engine with clear rules and controlled experimentation. AI generates content variants and micro-flows, yet every decision is logged, reversible and measurable.
Engineering reliability. Public SLOs for performance and availability; feature flags for incremental releases; observability on Core Web Vitals and application errors. WCAG accessibility designed upfront.
Unified governance. A single DX steering forum — product, design, data, technology and privacy at the same table — with a fortnightly Board pack covering economic impact and risks.
Results at a glance (after 12 weeks) — single list
Time-to-Quote: from 14′20″ to 2′52″ (p95).
Task Success Rate (quote + onboarding + ticket): +21 percentage points.
Deflection to self-service with satisfaction ≥4/5: +18 percentage points.
Digital cost-to-serve: –19% (like-for-like volumes).
Conversion on purchase journeys: +11%.
First-party Consent Rate: +24%; Data Quality Score: +17%.
Core Web Vitals: LCP < 2.5s in 92% of sessions; Availability 99.95%.
Experiment velocity: 7 tests/week; 38% with statistically significant uplift.
What we actually built: a Customer Operating System
Not a new homepage, but a Customer Operating System: content and services as composable building blocks; a data layer unifying identity, consent and real-time decisioning; journeys that adapt to context with verifiable rules; telemetry that ties every change to an outcome (revenue, cost, satisfaction).
“The difference isn’t cosmetic. We now know why something works — and how to scale it.” — CIO
Board-level lessons
Outcomes over touchpoints. Customers “buy” completed tasks, not pages.
Composability beats project fatigue. Once built, modules are reused everywhere.
Privacy is an asset. First-party data with explicit consent makes AI useful and sustainable.
Public SLOs change culture. When performance is an explicit commitment, it stops being optional.
Next steps (Q2–Q3)
Extend the Operating System to partner channels and mobile app; activate machine-callable journeys for external AI agents; establish a permanent experiment catalogue; link financial KPIs to variable budget. Joint targets: –30% cost-to-serve at nine months, +15% conversion on core journeys, NPS +10.
“DX is not a facelift; it’s the plant that turns every interaction into margin.” — CEO
