Skip to content

Choose which categories you allow. Necessary cookies can't be turned off.

INDYA · B2C SaaS · Health tech

Changing sports nutrition from the product side

Role
Product Lead
Period
-
Sector
B2C SaaS · Health tech
Team size
From 13 to 27 people
Reporting
CPO and CTO

01 — In a minute

01. Growth built on the lifecycle, not on features

  • Growth strategy built on the lifecycle, not on features. I co-defined the three levers (activation, engagement and retention) and the roadmap and the team's OKRs were organised around them.
  • A user research practice where there was none. Interviews, surveys and post-churn analysis on a continuous cadence: INDYA had a very good data structure and very little conversation with its users.
  • Monthly churn from 16% to 10%. Several fronts at once to understand and tackle both voluntary and involuntary churn: product work combined with customer success.
  • First-month activation: +28%. Onboarding, personalisation and value comprehension: so a new user understands what the app is for before the motivation that made them download it wears off.
  • Pricing redesign with A/B testing. I unified plans and removed entry barriers: +13% conversion and +5% ARPU, without touching retention.
  • More consistent delivery. From Trello to Jira, agile ceremonies actually run as such, explicit delivery and approval criteria, sprints with a single focus, and retrospectives to improve as a team.

02 — The story

02. From a service's software to a product

Plenty of data, very little conversation

INDYA had already come a long way when I joined, in January 2022. It had started as a nutrition clinic in Valencia: the founders saw they could deliver a better service if they monitored their clients' day to day, and commissioned the first version of the app from an external agency. It worked (it proved the service improves when the nutritionist sees the day to day) and on the back of that they hired an internal team and closed the first round, which put the app in the hands of the first clients outside Valencia and unconnected to the physical clinic. That is where INDYA stops being a service's software and becomes a product.

I joined with the brief of professionalising it (product, methodology, data and scalability) to turn it into the reference nutrition app. It had customers, it monetised and it had careful design, and that is exactly what made it hard to see what was missing: a goal system that did not order decisions, biweekly sprints that did not fit in two weeks, a high level of bugs and (most striking in a company with a very good data structure) very little conversation with users.

How I worked there

Focus before cadence. I inherited biweekly sprints that existed but never closed. What I changed was not the length: a sprint with fifteen open initiatives is not a sprint, it is a wish list with a date. We moved from Trello to Jira and, with the change, task management gained explicit criteria (definition of ready and definition of done), which is what lets you discuss whether something is finished without discussing it with whoever built it.

The retro as team maintenance, not as a verdict on the sprint. I introduced them as soon as friction between design and engineering became recurring. The rule that made them useful: we do not talk about what went well or badly, or about who; we talk about what each person will bring to the product next time.

From install data to product data. With the CPO pushing, we moved from Firebase's basic metrics to Amplitude. The effect was not having charts: it was that we stopped looking only at business metrics and started genuinely understanding the user.

And talking to users, which was the thing nobody did. I came in through the churn door (a problem with an owner opens a door that a research initiative does not) and out of it came four permanent inputs: post-churn (why they leave), power users (what keeps them and what they need), new users (what activates them and what blocks them) and general surveys for the pulse.

Apple, Gasol and the apps where the user already lived

The pairing with the product designer was one of the pieces behind the jump, and it did not take long to be noticed from outside: Apple spotted the app and brought us into Apple App Store Foundations, its mentoring programme. We came out of it as one of the first companies in Europe to ship widgets (a feature Android had had for years and that was just launching on iOS) and featured on the App Store: for the user, sorting the day's nutrition without opening the app; for us, visibility money cannot buy.

And we integrated where the user already was: Apple Health, Health Connect, Strava and TrainingPeaks. The direct benefit was not having to log the same thing twice, but the big one was for the service: the nutrition team began planning on real performance, not on what the client remembered.

All of this marks a path of work done well, and that pays off: with clients, and in the kind of partner who comes to you. Shortly after, none other than Pau Gasol became a shareholder. What mattered was not the amount: it was what it meant to a client who trains.

Pressing play loads the video from YouTube, under their terms.

03 — The case

03. Multiplying check-ins

I could tell you here how we raised activation or how we brought churn down, which are the cases you expect from a SaaS. I prefer a less standard one, because it shows better how I work: the lever was not in the problem's metric, it was two steps earlier.

Looking for what the most active users did differently, we filtered in Amplitude for those with more than five months and over 50% daily activity. What surfaced were check-ins: each user marks how they did on the activities their nutritionist plans (workouts and meals) on a five-level scale. It works partly as gamification, but above all it is the data the nutritionist uses to adjust the plan.

And two parallels appeared. Users who never activated mostly never touched the feature. And among those who did use it, usage dropped two to four weeks before cancelling. That turns a feature into a leading indicator: it does not explain why they leave, it warns that they are about to.

Before touching anything, the questions: were we explaining well what they are for? were five states enough? was marking them easy? did being consistent give anything back? The interviews confirmed almost all of them had a point, including from users who already did it daily and wanted to contribute more than a percentage: notes, photos.

So we made two minimal changes, to learn: explaining why they matter (to the user and to the nutritionist) and adding a daily counter of completed activities. The goal was not to move the number, it was to find out whether explaining the value and giving back a signal was enough. The answer was immediate: 32% more check-ins.

With the learning validated we stopped asking and started pushing: Complete in a minute (an end-of-day wizard), post-miss notification (if a day was left empty, the next one takes you to the wizard), Sentiment (optional tags that tell the nutritionist which food works before training) and a monthly report with compliance, best workouts and a ranking by sport. Three months in, it was no longer 34% of users logging their check-ins but 52%, and total volume had multiplied by 2.2: more users reached for the feature, and those already using it used it more.

Results

  • +32%

    check-ins from two minimal changes alone: explaining why they matter and giving back a daily signal

  • ×2.2

    total check-in volume three months later, with the share of users logging them going from 34% to 52%

  • 2-4 weeks

    of warning a drop in usage gives before a cancellation

More new users activated, better retention past the three-month mark, better information to plan with, and early detection of users at risk: all from looking at a behaviour nobody was watching.

04 — Takeaways

04. Four things I take from two years

  1. The lever was not in the problem's metric, it was two steps earlier. Attacking churn head-on would have produced a list of reasons; looking at what the people who stayed did differently produced a behaviour we could move and that warned us weeks in advance. Since then I look for the leading indicator first, not the cause.
  2. Well-run retros are the cheapest thing a team can do for its own speed. Not because of what gets fixed in each one, but because they turn friction into something said out loud instead of accumulated. The condition is that nobody talks about who is to blame.
  3. Integrating where your user already lives is worth more than asking them to come to you. Apple Health, Strava and TrainingPeaks added no features: they removed the work of logging the same thing twice, and along the way gave the nutritionist the data they did not have.
  4. Cash flow is the monster that devours startups. It is one of the metrics that needs the closest watch: when it drifts from the plan, it can kill an otherwise great product.