Building analytics products for someone else's customers

A different kind of product role

GoodData makes an embeddable analytics platform, the engine and visualization layer that SaaS companies drop into their own products when they need reporting and dashboarding capabilities without building from scratch. My job was to work directly with customers to define, design, and launch their analytics products on the platform, delivered within 3–6 month windows, with up to four engagements running simultaneously.

The customer base spanned wildly different industries and company sizes, from the largest financial services companies to ten-person startups. What they shared: they already had a successful core product, and they needed analytics as a key differentiator or a table stakes feature.

The skill this role demanded above all others was speed of understanding. I developed my practice to be able to rapidly develop genuine domain fluency, identify the analytical questions that actually drove decisions for a specific set of users, and translate that into a scoped, prioritized MVP solution fast enough to matter.

A discovery framework built for analytics products

When I arrived, the services team ran customer discovery using a workshop template built around Nir Eyal's Hooked framework. I disagreed with the premise, but I adored the workshop format. Hooked is a consumer engagement model designed to create behavioral habits, useful for social apps, poorly suited to B2B analytics products where users aren't trying to form habits, they're trying to answer specific business questions and make decisions. Applying a hook-based model to a fraud analyst or an autobody repair shop owner misses what actually matters to them.

I replaced it with my own framework, which I called Questions to Be Answered, a discovery approach built specifically for analytics products.

Alongside the discovery framework, I built a best practices guide for analytics product design centered on keeping users in a flow state, covering data visualization principles, techniques for mapping how users move between dashboards, and design patterns that made insight-finding feel natural rather than effortful. Combined with project plans, feature prioritization templates, definition of done criteria, and go-to-market materials, the goal was a repeatable, teachable practice for analytics product management that the team could carry forward.

Travel Spend Analytics

Tourism bureaus and destination marketing organizations used this dashboard suite to track international visitor spending by category and origin market, informing budget allocation and campaign targeting. Designed so that marketing directors could extract strategic insight without requiring analytical training.

1

Overview Dashboard

Year-to-date and monthly spend trends by visitor origin and market segment, with year-over-year change indicators. The entry point for most users, designed to surface the most decision-relevant signals immediately.

2

Country Detail View

Selecting an origin market surfaces year-over-year trends, card count, transaction volume, and a spend-per-card distribution, giving marketers context to prioritize high-value visitor segments over raw volume.

3

Multi-Country Comparison

Users could select up to five origin markets simultaneously to compare spending trajectories and category mix, designed for campaign planning across competing target audiences.

Across two continents

One dimension of this role that shaped me in ways I didn't fully appreciate at the time was the constant collaboration between GoodData's San Francisco and Prague teams. I was one of the few people working fluidly across both, bridging time zones, communication styles, and organizational cultures on live customer projects simultaneously.

It gave me an early fluency in distributed team dynamics that has informed how I've led and collaborated ever since, and it planted the seed for my eventual move to Prague.