I define the hypothesis before any code.
What we believe will improve, for whom, and why, with the success metric and baseline agreed before design starts.
GROWTH EXPERIMENTS & ACTIVATION · SERVICE 02
Product experiments that get more users to activate and convert.
I design, test, and measure product changes so you can decide with data what to scale before you build it.
Leandro Balbian · Growth Product Manager & Product Designer · 8 years in B2B SaaS and marketplaces
- FOR
- Heads of Product, CPOs, Growth teams, and founders
- DURATION
- 2 weeks of design + a measurement window
- FORMAT
- 3 live sessions + async work
- YOU LEAVE WITH
- A data-backed call: scale, iterate, or kill
02 / THE PROBLEM
Users show up, but they drop off before they activate.
The demand is already there: traffic comes in and users sign up. But somewhere in the flow, intent breaks, and the team ships changes without knowing which one moved the metric.
- Onboarding loses users before they see the value.
- Forms or critical steps have drop-off.
- Dead ends cut intent short.
- Features ship without a clear way to know whether they improved anything.
THE RULE
Measure to decide, not to decorate the experiment.
03 / HOW I DO IT
How I turn an idea into a measurable bet.
I design and prototype the intervention.
The flow, content, or change to test, in a prototype your team can test in days.
I measure and decide.
At the end, we compare against the baseline and make a data-backed call: scale, iterate, or kill.
04 / WHERE I STEP IN
I step in at the moments that decide whether a user moves forward.
ONBOARDING
First use
Users reach the moment they understand the value, fast.
ACTIVATION
Key action
More users completing the action that predicts they'll stay.
CONVERSION
Critical steps
Forms, checkout, or lead capture with less drop-off.
RECOVERY
Lost intent
Flows that rescue users when the main path fails.
I measure with GA4, Microsoft Clarity, and Metabase, or with whatever tools your team already uses.
05 / WHAT THE SPRINT LOOKS LIKE
Two weeks of design, a measurement window, and a data-backed decision.
BEFORE WE START
Starting point
WHAT I DO
I review the leak you've already identified and the tracking you have.
WITH YOUR TEAM
Read access and the metric we want to move.
WEEK 1
Define
WHAT I DO
I write the hypothesis, the success metric, and the baseline.
WITH YOUR TEAM
Kickoff session (60 min).
WEEK 2
Design
WHAT I DO
I design the flow and the prototype, and spec the events to track.
WITH YOUR TEAM
Prototype review with product and engineering (60 min).
MEASUREMENT
Decide
WHAT I DO
I read the results against the baseline and put together the readout.
WITH YOUR TEAM
Decision session: scale, iterate, or kill (60 min).
Important: the measurement window depends on your traffic, and we set it together in week 1. I recommend a PM and someone from engineering join all three sessions.
06 / WHAT YOU GET
Four deliverables, and what you can do with each one.
Experiment brief
FORMAT
1-page document
WHAT YOU DO WITH IT
Align product, design, and engineering on what's being tested, for whom, and with which metric.
Flow and prototype
FORMAT
Figma file + clickable prototype
WHAT YOU DO WITH IT
Your team tests it with users or takes it into development without reinterpreting it.
Measurement spec
FORMAT
Events and decision metric
WHAT YOU DO WITH IT
Data or engineering implements tracking knowing exactly what to measure.
Readout and decision
FORMAT
Results against the baseline
WHAT YOU DO WITH IT
Decide whether to scale, iterate, or kill, and what to test next.
You don't always need an A/B test: depending on traffic and risk, it can be a gradual rollout or a prototype validated before development.
07 / BEFORE AND AFTER
Before the experiment, the diagnosis. After it, the full product.
Each step builds on the last: the diagnosis defines what to test, the experiment validates it, and the product scales what worked, without re-explaining the problem from scratch.
- 01 Diagnosis
- 02 Experiment (you are here)
- 03 Product
01 · PRODUCT & GROWTH DIAGNOSIS
Find what to test before you design.
If it's still not clear where conversion leaks, the diagnosis finds the cause and the priority.
03 · COMPLEX PRODUCT DESIGN
Scale what worked.
When the experiment validates the bet, I design the full solution: flows, roles, a design system, and handoff.
08 / RELATED CASE STUDIES
The method, applied to real cases.
GROWTH
Intent recovery
A recovery flow that keeps intent alive when inventory isn't available. Proposal ready for validation.
GROWTH DECISION STUDY
Ghosts
A prioritized rescue queue for customers with real intent who never reached a broker.
PRODUCT DESIGN
ABLA
Sensitive flows with multiple entry points and continuity between actions.
09 / NEXT STEP
If users show up but don't move forward, let's design what to test.
Give me a quick rundown of where the flow breaks and I'll reply with a first idea of what to test.
Does it include paid media campaigns?
No. The focus is improving the product experience to convert the demand you already have.
Do you always need an A/B test?
No. The technique depends on traffic, risk, and operational capacity. It can be a rollout, a working prototype, or a test before full development.
Who implements the tracking?
I define the signal and the spec with your team. Data or engineering implements tracking where needed.
How long is the measurement window?
It depends on your traffic and the size of the change we expect to see. We set it together in week 1, before launch.