When you walk into an integrated energy major to help out with a SharePoint migration, you expect a certain level of corporate order. You assume you're going to find hard hats, safety glasses, and everything neatly labeled. But if you pry open the door to the enterprise analytics infrastructure, you might find something that looks less like a high-tech control room and more like a teenager’s bedroom the day before report cards come out.
For a long time, the digital analytics landscape at one of Western Canada's largest integrated energy majors was essentially an orphan. There was no business owner, no technical owner—it was just this thing that existed in the background, humming along in the dark. Everybody in the corporate world loves the phrase "data-driven." It looks great on a PowerPoint slide. But actually getting the digital plumbing to work across a massive organization is a whole different ballgame.
Changing the Locks on the Digital Saloon
My first look under the hood was at Google Tag Manager (GTM) for the main public website. GTM is a fantastic tool, but if nobody is watching the door, it turns into a digital saloon with no bouncer.
Because nobody specifically owned the analytics governance, third-party vendors had been given access over the years to add their own custom code. And they just never left. We found dozens of vendor accounts that had been sitting there for up to a decade. Some of these folks hadn't logged in since the Vancouver Olympics, yet they still had the keys to execute unvetted, custom JavaScript on a publicly traded company’s primary domain. It’s a bit like giving a spare house key to a plumber who fixed your sink in 2012 and just hoping he doesn't come back to rearrange your furniture.
The first step wasn't making things pretty; it was a hard perimeter lockdown. We revoked the old vendor access, dismantled the single-point-of-failure administration model siloed within downstream operations, and built a proper, documented access list. You can't paint the walls until you know who has the keys to the front door.
The Junk Drawer and the Sledgehammer
Once the perimeter was secure, we had to look at the actual data inside Universal Analytics (UA). It was a hoarder's paradise. Over the years, people had created 30 to 40 different analytics views, each with highly specific, completely undocumented filters. It was the digital equivalent of that one junk drawer in your kitchen—the one with the dead AA batteries, a single chopstick, and a warranty for a toaster you threw out in 2015.
With Google officially sunsetting Universal Analytics, we had a choice to make. We could spend six months trying to untangle a decade of digital spaghetti, or we could do the healthy thing: apologize to the drywall and bring in the sledgehammer.
We went with a "nuke and pave" strategy. We built a pristine, strictly documented Google Analytics 4 setup from scratch. But we didn't just throw the past in the trash; we carefully packed up 10 years of historical enterprise web telemetry into Google BigQuery before we bulldozed the rusty garden shed.
The Tim Hortons Drive-Thru & The Big Pivot
Once we actually gave the analytics landscape some shape and governance, the business suddenly started seeing real value. But that popularity brought its own headaches. People started treating the analytics desk like a Tim Hortons drive-thru window. "Yeah, hi, could I get a medium demographic report with a side of bounce rates, please?"
To save my own sanity, I productized the service, building slick, corporate-approved Looker Studio templates and forcing a strict intake process. We finally had a beautiful engine that could accurately measure how the public was engaging with the company's bilingual corporate publications.
"An activist investor group rolled into town, installed a handpicked CEO, and completely changed the mandate. Overnight, the editorial team was gutted."
And then, the corporate reality shifted beneath our feet. Publishing frequency plummeted from a couple of rich, bilingual stories a month to maybe one story every four months. Suddenly, our analytics landscape wasn't about B2C content engagement anymore; it became entirely defensive. The executives didn't care who was reading stories; they wanted to know how investors viewed them.
Right next door, their downstream retail fuels and convenience division was a completely different world. Because they were running a true retail B2C environment—selling gas, managing loyalty points, driving car washes—they had a mature, highly capable analytics team that made total sense for their business model. They were insulated from the corporate whiplash because their digital metrics were directly tied to cash registers.
| Request Vector | What Stakeholders Asked For | Architectural Solution |
|---|---|---|
| Ad-Hoc Reporting | Unfiltered pageviews, bounce rates by device | Standardized Looker Studio self-serve templates |
| Editorial Tracking | Curiosity metrics for bilingual corporate stories | Strict intake gating tied to actual business outcomes |
| Stakeholder Access | Manual email spreadsheets / analyst gatekeeping | Tiered role-based provisioning (Authenticated vs. Link-shared) |
The Florida Keys and the Proxy Trap
This sudden defensive posture brought a new challenge: executive telemetry. During quarterly earnings releases, our Legal and Investor Relations (IR) teams needed reports practically yesterday. They wanted to know if specific activist investor groups—say, a boutique firm operating out of the Florida Keys—were casing our website right after an earnings call.
Google Analytics can't do that, so I sourced and deployed a specialized tool called Irwin IQ to de-anonymize the IP traffic and flag the exact financial institutions kicking our tires. We successfully built a high-tech intelligence radar.
But here is where I made my biggest mistake: I didn't stay in the room to present it.
I allowed my manager to act as the go-between. He took the intelligence we generated, walked into the executive boardrooms, and the heavy oak doors clicked shut. I had zero visibility into what was actually being discussed, what the executives were asking, or how the data was being interpreted.
Instead of a healthy feedback loop, the intelligence just vanished into a black hole. Later, my manager would emerge from these high-stakes meetings and ask for things that made absolutely no strategic sense—saying something like, "Hey, the executive team really needs to know our bounce rates in Guam."
| Telemetry Layer | Google Analytics 4 (GA4) | Irwin IQ (Specialized IR) |
|---|---|---|
| Core Function | Aggregate consumer behavioral event tracking | Institutional B2B IP de-anonymization |
| Privacy Boundary | Strictly anonymized (no entity identification) | Maps corporate network CIDR blocks to firm names |
| Primary Use Case | Web traffic volume, campaign attribution, UX funnels | Detecting activist investors during earnings calls |
Because I wasn't in the room to hear the actual business question that led to that bizarre request, I couldn't course-correct or offer a better metric. And because my manager didn't fully grasp the mechanics of the intelligence pipeline we had just built, he couldn't translate the true value of the Irwin IQ data to the IR team. We were playing a broken game of corporate telephone.
Fixing enterprise analytics isn't just about cleaning up Google Tag Manager or surviving an activist takeover. It’s about realizing that if you architect the solution, you have to be in the room to present it. If you let a middle manager act as a blind filter between you and the executives, you aren't a digital architect anymore—you’re just a guy in the basement pulling bounce rates for Guam.