The Seven Principles for Making Data Platform Engineering Work

I recently revisited John Gottman’s book “The Seven Principles for Making Marriage Work.”

While the book focuses on relationships, it struck me how many of those principles also apply to building successful Data Platforms and Platform Engineering organizations.

Just like a strong marriage, a successful platform doesn’t happen by accident. It requires trust, communication, shared goals, and continuous care.

Over the past few years working on large-scale Data Platforms, I’ve realized that technology alone doesn’t make a platform successful — relationships with the teams who use it do.

1. Build a Deep Understanding of Your Users

In strong relationships, partners take the time to truly understand each other.

Platform teams must do the same.

Your users include data engineers, analysts, ML engineers, and product teams, each with different workflows and expectations. If platform teams build solutions based only on infrastructure perspectives, adoption will always be limited.

The most successful platforms are built by deeply understanding:

• How teams build pipelines
• How analysts explore data
• How ML teams train models
• Where friction exists in daily workflows

A platform succeeds when it solves real developer problems, not just technical architecture problems.

2. Invest in Developer Experience

Great relationships thrive on small daily investments.

The same applies to Developer Experience (DX).

Engineers shouldn’t spend hours figuring out how to deploy pipelines, request access, or configure infrastructure. Platforms should provide simple, repeatable workflows that remove friction.

This means building:

• Golden paths for common workflows
• Self-service onboarding for new teams
• Templates and automation for pipelines and environments
• Internal developer portals that make discovery easy

When developer experience improves, innovation accelerates across the organization.

3. Turn Toward Feedback Instead of Ignoring It

In relationships, ignoring feedback leads to frustration.

Platform teams face the same challenge. Users will always have feedback about tools, workflows, or limitations.

The worst thing a platform team can do is dismiss it.

Instead, successful platform teams treat feedback as a signal for improvement.

This means:

• Listening closely to user pain points
• Observing platform usage patterns
• Running regular feedback sessions and office hours
• Iterating quickly based on real-world usage

Platforms evolve through continuous collaboration with their users.

4. Let Platform Teams Influence Architecture

Healthy relationships respect influence from both partners.

Similarly, platform teams must have a voice in technical architecture and engineering standards across the organization.

Without this influence, organizations end up with:

• Fragmented tooling
• Inconsistent pipelines
• Duplicate infrastructure
• Governance gaps

Platform teams should help guide decisions around:

• Data standards
• Pipeline architecture
• Storage formats
• Observability practices
• Governance frameworks

The goal isn’t to control teams — it’s to create consistency that enables scale.

5. Solve the Problems That Matter

Not every problem needs solving immediately.

But platform teams should focus on the highest friction problems that slow organizations down.

These often include:

• Complex pipeline deployment
• Difficult data discovery
• Lack of observability
• Governance challenges
• Cost inefficiencies

When platform teams solve these foundational problems, they unlock productivity across dozens or even hundreds of teams.

6. Create Shared Meaning Across Teams

Great relationships thrive when partners share a common vision.

Similarly, data platforms succeed when teams share a common understanding of data.

Without shared meaning, organizations struggle with:

• Conflicting metric definitions
• Inconsistent data models
• Lack of trust in analytics

Platforms must create shared meaning through:

• Semantic layers and metrics definitions
• Metadata and data catalogs
• Clear data product ownership
• Governance frameworks

When teams share a common understanding of data, decision-making becomes dramatically faster.

7. Build Trust Through Reliability

Trust is the foundation of every strong relationship.

In Data Platforms, trust is built through reliability.

If pipelines break frequently, dashboards show inconsistent numbers, or costs spiral unexpectedly, trust erodes quickly.

Platform teams build trust by ensuring:

• Reliable pipelines
• Strong observability
• Transparent cost visibility
• Clear governance and data quality standards

When users trust the platform, they build on it with confidence.

Final Thought

At the end of the day, a Data Platform is not just technology — it is a relationship between the platform team and the teams who depend on it.

Just like strong relationships, successful platforms require:

• Understanding
• Trust
• Communication
• Continuous improvement

Technology may power the platform, but relationships determine whether it truly succeeds.