← Back to Innovator Stories
ProductEngineeringLeadership

The Hardest Decisions Were Not Technical

The more tools I explored, the harder the decisions became. Not because there weren’t options — but because there were too many.

6 min read05 Apr 2026, Sun

The Moment

As the product evolved, I found that the harder decisions were not about building features. They were about choosing what not to build - or more precisely, what not to depend on. There was a point where I started looking at external APIs more seriously. On the surface, this seemed like the obvious next step. Why build everything when there are already powerful services available? But once I went deeper, the reality was more nuanced.

The Tension

The first area was location. Using Google Maps and Google Places felt like the natural choice. The data is rich, the experience is familiar, and it solves many problems immediately. But evaluating it more closely revealed trade-offs. The second area was flight data. At first, I assumed there would be a clear best option. Instead, I found a spectrum of trade-offs. The third area was weather. This felt straightforward at first, but simply adding weather data did not automatically improve the product. Across all three, the challenge was not technical capability. It was deciding what was truly needed, what introduced unnecessary dependency, and what would quietly shape the product in ways I might not want.

What I Did and Changed

I approached each area as an evaluation rather than an integration exercise. For Google APIs, I used them selectively where they improved search and place accuracy, but avoided over-dependence where they would define my product structure. For flights, I evaluated rich aviation APIs, travel platform APIs, and lighter options, and realised there was no perfect choice — only trade-offs between capability, complexity, and cost. For weather, I shifted from thinking of it as a feature to treating it as contextual support embedded in the right place. The key change was moving from asking “what can I integrate?” to asking “what actually improves the experience?”

Google API Evaluation

API / OptionStrengthTrade-offSummary
Google Places APIAccurate search and autocompleteCost and usage constraintsBest used for input, not as core data structure
Google Maps JavaScript APIRich visual experienceAdds complexity and dependencyUseful selectively, not essential everywhere
Google Static Maps APILightweight map previewsLimited interactivityGood balance for simple use cases
Google Geocoding APIReliable location conversionAdditional dependencyBest used as a supporting layer

Flight API Evaluation

API / OptionStrengthTrade-offSummary
Rich Aviation APIsDetailed data, full coverageExpensive, complexToo much for what the product actually needs
Travel Platform APIsEnd-to-end ecosystemSetup overhead, constraintsPowerful, but introduces friction
Lightweight APIsEasy to startLimited reliability or depthGood for testing, not always for scaling

Weather API / Weather Approach Evaluation

API / OptionStrengthTrade-offSummary
Full Forecast DataComprehensive informationToo much for users to processMore data does not mean more value
Simplified Weather APIs (for example Open-Meteo)Lightweight, sufficient accuracyLess detailedEnough for decision-making context
Standalone Weather FeatureEasy to implementDisconnect from user flowInformation without context adds effort

The Insight

The best solution is not always the most powerful one. It is often the one that fits the experience you are trying to create.

Broad Reflection

External platforms are not neutral. When you use them, you also adopt some of their assumptions, constraints, and structure. The real discipline lies in deciding where to leverage them and where to retain control. Often, the hardest decisions are not about what you can build. They are about what you choose not to depend on.

Related Stories

LeadershipProduct

The Builder’s Shift: When Thinking Turns Into Doing

I thought I was just trying something small. I didn’t expect it to change how I think about building.

4 min read05 Apr 2026, Sun
Read Story →
EngineeringLeadership

Alignment Over Complexity: Why Systems Fail Quietly

Everything was working. Until I realised I couldn’t trust what I was seeing.

4 min read05 Apr 2026, Sun
Read Story →
ProductDesignLeadership

Why Good Products Feel Simple (Even When They’re Not)

We had everything built. And yet, something still felt wrong.

4 min read05 Apr 2026, Sun
Read Story →