Analytics Engineering

Jonathan Bridges

Analytics Engineer & Measurement Specialist

I build the systems that tell a business whether its marketing is actually making money. Tracking that fires the way it should. Pipelines that don't quietly break. Dashboards that answer the question an owner actually has, which is usually just "is this working?"

Selected Work

Building with AI

How I actually use AI: directed to do the reading and drafting, never trusted with a decision that has to be right, and verified before it ships. The method behind everything here, proven across three projects.

  • AI direction
  • Verification
  • Judgment
View case study →

Multi-Source Dashboards

Client-facing Looker Studio dashboards that pull analytics, Google Business, point-of-sale, and social into one view. Built around the questions an owner actually asks, and they flag what to do next.

  • Looker Studio
  • GA4
  • Multi-source
View case study →

The Command Center

A production system that audits how a business measures its marketing. It finds where the tracking is lying, and it won't report a number it can't stand behind.

  • Node.js
  • Headless render
  • Deterministic scoring
View case study →

Measurement Setup

GA4 and GTM set up so a site measures what matters. The standard events done right, plus custom events tuned to how each product actually works.

  • GA4
  • GTM
  • Custom events
View case study →
How I Think

I built an audit that scores how well a business measures its marketing. Most of that score comes from reading the business's website. But some sites can't be read. They block bots, or the URL is bad, or the server throws an error. The easy thing is to score those as a zero.

A zero is a lie there.

A zero from "I read the site and nothing's set up" looks exactly like a zero from "I couldn't read the site at all." Those mean opposite things, and one of them is a number I might put in front of a prospect and be completely wrong about. So I built the scoring to refuse. When it can't actually see a site, it doesn't guess a low number. It says "couldn't scan" and holds back the overall score.

That's the whole job to me. Anybody can make a tool spit out a number. Knowing when a number shouldn't be trusted, and building the thing to say so instead of faking it, is the part that takes judgment.

About

I'm an analytics engineer, and I got here through a psychology degree and a habit of building things. The psych part actually matters. Measurement is really about behavior, and a dashboard is only worth something if it changes what someone decides to do.

I run E11EVEN Analytics, a measurement studio. I take on work directly, and as a white-label partner for agencies that don't have analytics in-house. I built everything on this site while working a full-time job, which should tell you what I do with my own time.