Building the analytics foundation

How I designed Intermedia's first analytics platform — from naming convention to query builder, filters, and a chart system built from scratch.

Industry

Unified Cloud Communications · 2018

Role

Lead UX designer
Full ownership of Analytics

Stack

Sketch/Figma
Miro

Context

Partners reselling Intermedia's services had no analytics layer of their own. Every report, every performance check, every call dispute, it went through an admin. Managers couldn't see their own data. Competitors could offer this. Intermedia couldn't.

Outcome

We built the analytics platform from scratch. For the first time, senior roles had visibility into how their teams and services were performing. 60% of 125,000 customers adopted it.

125,000

Intermedia customers

60%

analytics adoption rate

The challenge

Intermedia had grown through acquisitions. Every service had its own name, its own dataset, its own terminology. AnyMeeting. Voice. Different metrics, different data structures, no unified language. Before designing anything, the product needed a common language.

Customers were direct about what was missing:

Customers challenge us about certain calls and we don't have the data to defend ourselves.

Line managers are asking for numbers and I have to ask developers to export metrics.

The report we have is hidden in the configuration portal and I keep trying to find it.

I proposed we stop using service brand names and rename everything to human language representing the entity. Not the product.

The filter system was the hardest problem

The core design problem: how do you let someone with no technical background query across different datasets?

Date range: what period
Report type: what service
Filters: what entities within that service

Tabular data needed a visual language

No colour system existed for data visualisation. I built one from scratch, derived from the brand identity.

The date picker evolved post-launch

Post-launch, customers needed to include archived reports. The component was revisited and extended — the three states show that evolution.

The shipped product

Launched as MVP with core report types. Each subsequent report was its own scoped project built on the same foundation.

What shipped and what it meant

60% of 125,000 customers adopted the platform

Developers no longer manually exporting metrics

A unified naming system across all services

A foundation every subsequent analytics project was built on

Reflection

I pushed for proper analytics instrumentation from day one. We tracked adoption through Google Analytics, enough to know 60% of customers used it but not enough to know how. Feature-level tracking was proposed and not resourced. In retrospect, that data would have accelerated every subsequent decision about what to build next.

Next Case Study

Building the analytics foundation