Jeff Cai

Case studies.

This is a selection of case studies, the shape of each problem, and how I helped.
Names are omitted due to NDAs.

Dashboarding

Consolidating fragmented reporting into a single trusted source.

Franchiser, 40+ locations

Each of more than 40 individual locations tracked leads and named campaigns differently, so there was no reliable way to compare performance, or even know how many leads the franchise was generating as a whole. Rebuilt the event taxonomy from scratch and consolidated every location into one unified data model.

One trustworthy source of truth, now used daily by leadership, internal teams, and individual locations to make decisions.

Retailer, live performance dashboard

Given carte blanche by the client, trusted to do what was best for the business rather than working from a brief. Set the measurement strategy, chose the KPIs, and designed a live dashboard with drill-down into category and marketing performance from the ground up, including the underlying connectors and data feeds.

Leadership went from no real view into their data to an interactive year-over-year report they could explore themselves.

Post-mortem analysis

Diagnosing what broke in a campaign, launch, or reporting pipeline after the fact.

Multi-banner retail group

Quarterly revenue looked soft right after significant site changes, a new consent banner, and a new email system all landed in the same quarter, while SEO and paid media were shifting at the same time. Every channel was confounded together, so no single change could be cleanly blamed for the dip, and leadership needed to know whether to pull back spend or hold the course.

Isolating one cause was not possible, so tracked SEO, paid media, email, and direct back to baseline across every banner to build the recovery narrative instead. Confirmed it as a transition effect, not lost demand, and presented it to the C-suite at head office, who held spend steady into next quarter.

Tagging & tracking

Fixing the foundation so everything built on top of it can be trusted.

Lead-generation company, Quebec

The consent management setup was not blocking any data by default for Quebec traffic, which meant the client was not compliant with the province's privacy law and was exposed to a fine. Mislabeled ad tags were also leaking or suppressing the wrong signals, and lead-capture forms were firing tracking events even when submissions failed or were incomplete.

Rebuilt the tagging and consent setup end to end, including cross-domain tracking into their booking platform, so lead attribution by channel finally holds up.

Supplements company, $600K/mo ad spend

Event tracking was built as long concatenated strings instead of proper dimensions, the ad platform was not connected to receive conversion data, and the lead form was silently stripping UTM parameters before submission, all while spending $600K a month on ads with no reliable way to see what was working. Consent handling also needed to be brought up to California's requirements.

Rebuilt the tracking and form implementation so attribution flows cleanly into their CRM, with server-side tagging now underway as the next layer.

How can I help?