case study· Seed · Toronto

Sole business hire across 400+ customers.

Joined a seed-stage product team as the only business-side hire. Split the platform along user lines, automated a support backlog, and shipped the dashboards that became the shared source of truth.

83%
SDR setup time cut (2h → 20m)
8+ hrs/wk
support-review effort eliminated
40%
reduction in manual reporting
400+
customers analyzed
01 · problem

Where it started.

Spector's early product was built for engineers, but growth came through business-side SDRs who didn't have engineering context. Setup took 2 hours. Support tickets were piling up. Leadership had no single pane of glass for product usage, customer health, and support load.

02 · approach

How I worked it.

01

Usage analysis in BigQuery + Postgres

Mined the product DB + event stream to find where non-technical users dropped off vs. where engineering users dug in. The data made the split obvious — two user types needed two different onboarding surfaces.

02

Platform split: business vs. engineering

Spec'd and championed a fork of the onboarding flow — streamlined configuration for SDRs, advanced surface preserved for engineering. SDR setup dropped from 2 hours to 20 minutes, an 83% cut.

03

Automated support classification

Designed an auto-classification pipeline using tags, clustering, and templated routing. Processed 400+ items — support tickets, internal notes, and product feedback — and eliminated 8+ hrs/week of manual review, while surfacing recurring themes that fed the roadmap.

04

Power BI as source of truth

Built dashboards stitching product usage, customer health, and support metrics. Replaced a patchwork of weekly slide decks — manual reporting effort fell ~40%. Exec, product, and CS all made decisions off the same board.

03 · outcome

What shipped.

  • SDR setup time: 2 hours → 20 minutes (83% reduction).
  • Support-review workload: 8+ hrs/week → 0, with roadmap themes surfaced as a byproduct.
  • Manual reporting effort across teams: down ~40%.
  • Shipped the first cross-functional dashboards the team actually trusted.
stack
BigQueryPostgresPower BIPythonSQL

Want one like this for your team?

I'm looking for analytics engineer / operator roles where this kind of work is the job.