Portfolio · Fractional Chief of Staff engagement Building Metrics From Scratch for a US SaaS A US SaaS company had just raised its Series A with no reliable view of its own numbers. In 12 weeks, I built the metrics, dashboards and reporting it needed to run the business and the board with confidence.

The challenge

Data lived across Stripe, HubSpot, Zendesk and a tangle of spreadsheets. There was no agreed north star, each team tracked different things, and nobody could confidently answer basic questions about churn or activation. Every Monday the CEO spent half a morning compiling a board update that still felt unreliable.

Approach

I worked as the CEO's right hand, starting by mapping every data source, report and unanswered question. Then I brought the CEO and team leads together to agree one north star metric and the metric tree beneath it, connected the data into a single source of truth, and set up a weekly rhythm so every metric had an owner and a clear next step when it moved.

---

Key deliverables

• North star and metric tree: one shared definition of success, broken down by team
• Single source of truth: Stripe, HubSpot and Zendesk connected, with no more manual pulls
• Three live dashboards: company, team and board level, all from the same data
• Automated board report: built overnight and ready for review on Monday morning
• Reporting rhythm and ownership: who owns each metric, how often it is reviewed, and what happens when it moves

---

Inside the build: the automated board report

The workflow that gave the CEO her Monday mornings back. The weekly update used to mean pulling numbers from four tools by hand. Now the report builds itself overnight and waits for a quick review.

1. 1
   Data syncs overnightStripe · HubSpot · Zendesk

   Revenue, pipeline and support activity are pulled automatically every night, so nobody exports anything.

2. 2
   Metrics are calculated in one placeMake · Looker

   The data lands in one model where MRR, churn, activation and the north star are calculated the same way every time.

3. 3
   AI drafts the commentaryClaude API

   The AI compares this week with last week and the targets, explains what moved, and flags anything outside its target range.

4. 4
   The CEO reviewsGoogle Docs

   On Monday morning the draft report is ready. The CEO checks the numbers, adjusts the commentary and approves it.

5. 5
   The board receives itEmail

   The approved report goes to the board with live dashboard links for anyone who wants to dig deeper.

---

Results

---

Skills applied