From Raw Analytics to a Decision-Making Engine
A global asset manager had GA4 running across two properties but no way to turn it into decisions. We built a six-dashboard system, from discovery to v1.0, in two weeks — then handed it off for the team to run themselves.
Client: A global asset management firm, spanning its flagship site and a secondary strategy-specific property
Stack: GA4, Looker Studio, Microsoft Clarity
Timeline: Discovery to v1.0 in a couple of weeks
Engagement: Analytics architecture & dashboard suite (one-time build, personal handoff, detailed leave-behind)
The problem
The client's marketing and editorial teams had Google Analytics 4 running across two properties — the firm's main site and a secondary site for one of its investment strategies — but raw GA4 data doesn't answer the questions a content or marketing team actually has: Which articles are working? What are people searching for? Do employee bios drive anyone toward a fund page, or are they a dead end? Which pieces of content have quietly gone stale?
One example: the team had a filter/search feature live on the site and genuinely didn't know if anyone was using it, or how. No visibility into what people were actually searching or filtering for — just a feature sitting there on faith.
Without a system built to answer those questions, every content decision was a guess dressed up as a hunch.
What we built
Large Fraction designed and shipped a six-dashboard analytics suite that turns GA4's raw event data into a decision-making system the marketing and editorial teams can run themselves — no analyst required to interpret it.
1. Insights Hub Performance — Splits content health into two views: overall performance (views, engagement rate, true engagement time) and top entry points (which pages are actual front doors to the site). Cross-referencing the two instantly separates "acquisition champions" (content pulling people in from outside) from "internal discovery" pieces (content people find once they're already on-site).
2. Thought Leadership Search & Filter Trends — Tracks active intent, not passive browsing: which topics, asset classes, and investment teams users deliberately filter or search for. A custom-built signal for where content demand is highest — and where the taxonomy doesn't match how users actually search (e.g., users typing "inflation" for content tagged "Macro Economics").
3. Employee Bio & Journey Analysis — Follows the "next step" after someone reads a staff bio: do they move to a fund page (high commercial intent), loop back into thought leadership (credibility building), or hit a dead end? Gives the team a clear, evidence-based case for adding next-step links to bio templates.
4 & 5. Content Decay Analysis (Main Site + Secondary Site) — An interactive, URL-level time series that overlays traffic against engagement rate for any piece of content. It distinguishes two very different problems that look identical from views alone: content that's still excellent but no longer being distributed, versus content that's genuinely gone stale and needs a real refresh.
6. Secondary Site PDF Leaderboard — Ranks downloadable assets by actual download volume — the highest-intent action GA4 tracks. Surfaces which whitepapers and reports are worth repurposing into LinkedIn posts, email sends, or short videos.
The approach
Each dashboard was built around a specific operational question, not a generic metrics dump — the difference between a report someone glances at once and a tool a team actually uses every week. Custom event tracking was engineered on top of native GA4 data (filter interactions, bio-to-next-page journeys, content decay over time) to capture the behaviors that matter but that GA4 doesn't surface out of the box.
Just as important: the suite shipped with a full strategic guide — not just definitions of what each metric means, but plain guidance on how to read combinations of signals and what action each pattern should trigger. The goal wasn't to hand over dashboards; it was to hand over judgment the team could exercise without waiting on an outside analyst.
The outcome
Version 1.0 is live in production, giving the marketing and editorial teams a consistent, shared view into how content, people pages, and downloadable assets are performing across both properties — replacing scattered, ad hoc GA4 digging with a system built to be checked, trusted, and acted on.
The architecture is built to evolve: as the team develops new questions or launches new content initiatives, the dashboard suite extends to answer them.
The engagement was a one-time build, not a retainer — the team was handed the dashboards along with a detailed leave-behind guide, so they can run and interpret the system themselves without waiting on an outside analyst.