I pulled a product back from the brink of shutting down: churn from 17% to 7.8% and break-even 2 months ahead of schedule

Key results

• Monthly churn from 17% to 7.8%, a 54% drop
• Involuntary churn from 30% to 12% of total churn, through billing optimization
• Monthly revenue loss from R$47k to R$18k, about R$348k/year recovered
• LTV/CAC ratio from 0.5 to 3.2
• Break-even in month 11, against a 13 month deadline that meant break-even or shut down
• Customer lifetime from 5.9 to 12.8 months, and NPS from 12 to 42

Context and background

The product

3DGym was a mobile app built for physical therapists and orthopedists to manage patient treatments, create personalized exercise programs, and track patient progress remotely. The platform let healthcare professionals:

• Create and customize rehabilitation exercise programs
• Monitor patient adherence and progress
• Communicate with patients through in-app messaging
• Access a library of professionally filmed exercise demonstrations
• Track treatment outcomes and adjust protocols as needed

Monetization model

3DGym ran on recurring subscriptions across two plans: a premium plan above R$99/month and an entry plan at R$34.90/month. With that mix, the average ticket was R$69 when I joined and reached R$73 by the time I left, reflecting the plan mix and the introduction of the annual plan and the Professional tier over the course of the initiative.

Target audience

The product served two profiles of healthcare professionals: physical therapists and orthopedists. That segmentation by profession resurfaced in the root-cause analysis of churn: physical therapists typically had lower credit card limits and renewals concentrated at month end, which made them more vulnerable to payment failures.

That difference shaped onboarding and billing cycle decisions throughout the initiative.

My role

As Head of Products & Engineering, I was responsible for product strategy and roadmap definition, leading product discovery and delivery processes, managing the engineering and design teams, establishing data infrastructure and analytics practices, coordinating with the Customer Success, Marketing, and Support teams, and presenting metrics and strategic recommendations to the founders. The role required balancing strategic decision-making with hands-on execution across product management, data analysis, user research, and occasional design work.

Obsession with the problem

Finding the root cause

My first priority was understanding whether the business challenge came from product-market fit, acquisition, operating costs, or retention problems. Through systematic analysis, I determined that churn was the main blocker to profitability.

Discovery process

1. Analyzed cohort retention curves using Firebase Analytics
2. Conducted exit interviews with churned users (15+ conversations)
3. Reviewed customer support tickets to identify friction points
4. Mapped user journeys to understand drop-off patterns
5. Segmented the user base by profession (physical therapists vs. orthopedists)
6. Examined payment failure rates and billing cycle patterns

1. Involuntary churn (30% of total churn)

The data revealed that nearly a third of all churn was involuntary: users who wanted to keep using the product but ran into payment failures. Deeper analysis showed that renewals happened at month end, when physical therapists' credit cards often hit their limits; physical therapists typically had lower credit limits than orthopedists; the app store grace period was very short (3 to 5 days); there was no proactive communication about payment failures; and users often only discovered their subscription had lapsed when trying to access the app.

2. Poor onboarding experience

Users struggled to grasp the product's value proposition during their first sessions: 65% of trial users added fewer than 2 patients in their first month, average time to the first meaningful action was 8 or more days, there was no guided setup or activation milestones, users got full feature access with no context or progressive disclosure, and there were no clear success metrics communicated to users.

3. Low feature adoption

Analysis of engaged users versus churned users revealed critical behavioral differences:

4. Technical quality issues

User interviews and support tickets revealed frustration with app crashes on older devices (particularly Android 7.0 and earlier), forced updates that blocked access until updating, slow video loading for exercise demonstrations, and sync issues between the mobile app and web dashboard.

The challenge and the constraint

When I joined 3DGym as Head of Products & Engineering, the company faced a critical business challenge: despite significant investment in user acquisition, the product was losing customers at an unsustainable rate. The company needed to reach break-even in 13 months or face possible shutdown.

Initial assessment

• Monthly churn rate: 17%
• High customer acquisition cost (CAC of about R$800) against an average ticket of R$69, resulting in negative lifetime value (LTV)
• Limited engineering resources: 2 devs and 1 PO, all outsourced through a consultancy
• No metrics infrastructure in place
• Unclear understanding of user behavior patterns
• Product still not at break-even after multiple years in market

Strategic approach

With limited engineering resources and a 13 month deadline, I needed to maximize impact fast. I used a modified RICE framework weighted for speed: Reach (how many users affected), Impact (how much this reduces churn), Confidence (how sure we are it will work), Effort (engineering time required), and Speed to Value (how quickly we'll see results).

That led to a three-phase approach:

Prioritization: what got cut

In the first two phases, I deliberately left out of scope: more elaborate engagement loops (gamification, referral program, push notifications), pricing changes (annual plan, Professional tier), and predictive churn modeling. Under the modified RICE model, those items had lower reach and speed to value given the 13 month deadline, or depended on a more stable retention base before they were worth investing in. Systematic A/B testing of pricing also didn't make the cut within the deadline, and the price adjustment ended up being more ad hoc, in phase 3.

Hard calls

Attacking involuntary churn before onboarding. Onboarding had greater long-term impact potential on activation and LTV, but needed more design and engineering time before showing results. Involuntary churn was faster to fix and bought time and credibility for the rest of the roadmap, at the cost of delaying the more structural activation problem by a few months.

Restructuring the outsourced engineering team. I moved from 2 devs and 1 PO outsourced through a consultancy to a single, more experienced internal senior dev at a lower cost than the outsourced setup. The trade reduced raw execution capacity in favor of more experience and direct alignment with product strategy, which in turn demanded even more disciplined roadmap prioritization.

Solution implementation

Phase 1: quick wins (months 1 to 3)

Payment optimization. The fastest path to reducing churn was addressing involuntary payment failures. I extended the grace period from 5 to 14 days on iOS and 21 days on Android, implemented email notifications on payment failure (days 1, 7, and 12), added in-app prompts to update payment information, shifted the billing cycle for physical therapists to mid-month (when credit limits reset), and created a special payment plan for the high-risk cohort.

Implementation: 2 weeks, minimal engineering effort.

Critical stability fixes. In parallel, I fixed the crash affecting Android 7.0 users (15% of the base), made app updates optional with a grace period, optimized video loading with progressive download, and added offline mode for the exercise library.

Implementation: 4 weeks of engineering.

Phase 2: onboarding optimization (months 4 to 7)

With the bleeding stopped, I focused on getting users to the activation threshold (6+ active patients). Before redesigning onboarding, I ran 12 moderated test sessions with new users, analysis of activation patterns of successful users, competitive analysis of similar health apps, and research to understand therapist workflows.

Solution: progressive onboarding with milestone-based activation. A welcome flow asking about practice type and biggest challenges, quick setup of the first patient profile in under 2 minutes, a first guided exercise program with templates, milestone celebrations at 1, 3, and 6 patients, and compressed time-to-value by moving critical actions into the first session. I also built role-based templates, empty-state guidance, a visible progress checklist for the first 30 days, and a drip email campaign reinforcing activation milestones.

Implementation: 6 weeks of design and engineering.

Churn for users completing the onboarding checklist: 9%, versus 19% for those who didn't.

Phase 3: product quality and engagement (months 8 to 13)

The final phase focused on removing remaining friction points and building engagement loops: weekly progress summaries for therapists, push notifications for completed exercises, in-app gamification, and a referral program. On pricing, I introduced an annual plan with a 15% discount, created a "Professional" tier for high-volume practices, and implemented a usage-based trial. On quality of life, I cut app size by 40%, improved library search, added bulk actions, and built integrations with clinic management systems.

Implementation: 5 months of iterative improvements.

Results and impact

Key learnings

Applicable frameworks and methods

• RICE prioritization to evaluate and sequence initiatives.
• Cohort analysis to understand retention patterns across user segments.
• Jobs-to-be-Done to understand therapist workflows and pain points.
• Activation metrics to identify leading indicators of long-term retention.
• Funnel analysis to map the user journey from acquisition to activation to retention.
• Root cause analysis (5 whys) to diagnose the causes of involuntary churn.

Conclusion

Cutting churn from 17% to under 8% required a holistic approach, combining fast tactical wins with long-term strategic improvements. The key was maintaining disciplined prioritization in a resource-constrained environment, letting data drive decisions, and deeply understanding user behavior.

This experience reinforced that solving complex product problems requires more than technical fixes: it requires a comprehensive view spanning product, engineering, customer success, and business strategy. By focusing on understanding the "why" behind user behavior and systematically addressing root causes, we turned a struggling product into a sustainable, profitable business.

The frameworks and approaches developed during this initiative have since been applied in subsequent roles, consistently delivering measurable improvements in retention, activation, and product-market fit.

Appendix: tools and technologies used

A note on the data: the financial figures and metrics in this case have been sanitized and approximated for confidentiality. The proportions, direction, and order of magnitude of the results are faithful to what was observed.