Data Driven Design
Numbers open and close the work.
They define the problem, measure the hypothesis and prove the result. Without an agreed metric, the design is not done.
- +117% in credit mix (Carnê Digital · CDC dashboard · 2022–2023)
- 33 → 4 clicks to credit (Carnê Digital · funnel)
- 100% organic traffic (Colbe · product analytics)
The funnel before the screen
The dashboard showed the drop, not the reason. I crossed the step where people vanished with what they said in interviews.
Hypothesis, metric and stop criterion agreed before the first prototype.
Growing without paid media
At Colbe every event is born with a business question. No orphan events, no dashboard nobody opens.
I decide what to build and what to cut by looking at retention, not volume.
What must not get worse
Every metric that goes up pushes another one. While credit mix went up, what could not drop was agreed beforehand.
Every hypothesis ships with the metric it chases and the one it protects.
- Define One metric the whole team understands.
- Instrument If it is not in the event, it did not happen.
- Read Funnel, cohort, segment.
- Decide Reading becomes priority.
- Prove Before and after, with a criterion.
Tracking plan
Every event with the business question it answers. No orphan events.
Privacy
No name, email or typed text goes to GA4.
Funnel
Start, cases, contact, email copied. Where people stop, I see.
Quality
Core Web Vitals measured on the visit, live accessibility and SEO audits.
Success
Rate of reaching contact and conversations booked per week.