By: Sarah Mwendwa

Picture this: You’ve just launched a platform with a flawless frontend, a seamless backend, and perfect APIs. Then your client logs in to their new admin dashboard and asks the inevitable question: “So… how many people are actually using this thing?” Welcome to the world of analytics, where numbers become narratives and dashboards answer the strategic questions your stakeholders need to ask.

The Core Value of Analytics

Here’s the truth: Analytics are not fundamentally about the numbers; they’re about confidence. When you invest heavily in building a platform, you want to rest easy, knowing your investment is yielding returns. Are sign-ups increasing? Are users engaging with the core features? Is the platform thriving or gathering digital dust?

Analytics serve as the heartbeat monitor of any platform. They transform abstract hope (“We have users!”) into concrete facts (“We have 1,247 active users, 342 of whom logged in yesterday”). They turn initial anxiety into clear insight and unanswered questions into actionable strategies. For founders, product managers, or anyone invested in the platform’s success, analytics are not merely nice-to-haves—they represent the difference between proactively steering a ship and drifting in the dark.

Not All Metrics Tell the Same Story

Analytics are stratified, and each layer reveals a different part of the platform’s story.

Consider a gaming platform. You might initially see “1,000 registered users” —a seemingly decent headline number. Upon digging deeper, you find only 200 logged in this week. Further inspection shows that the 50 users who created teams are logging in 5 times as often as solo players. The headline number suggested “we’re doing fine,” but the deeper data revealed that “teams are our secret sauce”.

Platform analytics work across three distinct levels:

• Level 1: The Surface: “We have 1,000 users.” This is a foundational count, but it doesn’t clarify whether the user base is active or growing.
• Level 2: The Behaviour: “We have 1,000 users, but only 200 logged in this week. 50 of them created teams, and 15 made transactions.” This level pinpoints what users are actually doing.
• Level 3: The Pattern: “Active users are up 15% month-over-month, team creation happens mostly on Tuesdays, and users who join teams are three times more likely to stick around.” Here, you move past what happened to understanding the why and predicting what might happen next.

The key is discerning which level of insight is required to answer a specific question.

The Strategic Framework for Choosing Analytics

To determine which metrics to track and build, use this four-point mental framework:

1. The “Coffee Meeting Test”: Imagine meeting with a key investor or executive. The numbers you need to present to sound confident and informed are your priority metrics. If you can’t answer “How’s the platform doing?” with verifiable figures, a fundamental requirement has been missed.
2. The “So What?” Filter: For every potential metric, ask, “So what?”. Tracking page views, for example, is meaningless unless it informs a specific decision. Tracking daily active users, however, tells you if engagement is growing or shrinking, which informs marketing spend and feature priorities. If a metric doesn’t lead to a clear “So what?” conclusion, it’s likely not worth developing yet.
3. The “Pain Point” Detector: Listen closely to what stakeholders and users repeatedly complain about or ask for. “I wish I knew which skills are most common in our talent pool…” immediately identifies a clear analytics requirement. Stakeholders often write your requirements; all you need to do is listen carefully.
4. The “Decision Support” Principle: Every effective analytic must help someone make a decision. User growth metrics guide marketing investment; feature usage guides development prioritisation; transaction trends guide pricing adjustments. If a number doesn’t support a decision, it’s merely data floating in space.

Making Analytics Actionable

The art of analytics isn’t just about calculation; it’s about presenting data in a way that makes sense to the human brain and drives action.

• Nail the Basics: Before building complex funnels and cohort analyses, ensure your foundation is solid: reliably track total users, transactions, and core actions.
• Context is Everything: “1,247 users” is just a number. “1,247 users (up 23% from last month)” is a story. Giving context—by comparing to last week, last month, a target, or a benchmark—transforms metrics into actionable insight.
• Show Trends Over Time: Almost every analytic increases in value when viewed over time. A single day’s number is a snapshot; comparing this week to last week reveals a trend. Always allow users to zoom in and out across different timeframes (day, week, month, all-time).
• Keep it Scannable: Your dashboard should answer “Is everything okay?” in approximately 10 seconds. Put the three most critical metrics front and centre—if everything is highlighted, nothing stands out. The top section must be your “health at a glance” zone.
• Slice Your Data (“Don’t Forget the ‘Who'”): The most valuable insights come from slicing data by region, user type, team, or any dimension relevant to your platform. “Overall engagement is down” is worrying; “Engagement is down, but only in one region” is a specific problem that can be fixed.
• Prioritise Simplicity: Resist the temptation to build the fanciest dashboard imaginable. Start simple, make it reliable, and only layer on complexity once you understand what people truly need. You can always add more; it’s difficult to subtract confusion once it’s present.

The Bottom Line

Analytics are storytelling tools disguised as numbers. Every metric you build must answer a question, solve a problem, or support a key decision. The true art is not in calculating complex figures, but in choosing which numbers actually matter.

Next time a client asks for analytics, think about the story they need to tell, the questions keeping them up at night, and the decisions they need to make. Then, build the dashboard that empowers them to do exactly that.

Happy building, and may your APIs always return 200s.