What is a tech stack?

What a marketing stack typically contains

• Measurement — analytics, attribution, an MMP where apps are involved.
• Buying — DSPs, networks, platform advertising accounts.
• Data — a warehouse, ETL, a CRM or customer data platform.
• Engagement — email, push, in-app messaging, marketing automation.
• Tag management, such as GTM.

How stacks bloat

By addition without subtraction. Every tool arrives to solve a specific problem and none of them leave when the problem does. Contracts renew automatically, integrations stay wired, and the SDKs stay in the app.

The costs compound quietly: overlapping functionality across three tools, conflicting numbers that nobody can reconcile, app size and startup time, and a data model where the same customer exists differently in five systems.

Most organisations discover on audit that they are paying for several tools nobody has opened in a year.

What a good stack looks like

Fewer tools, each with a clear owner and a defined role, integrated around one source of truth for the metrics that matter.

The single most valuable property is that the numbers agree. A stack where the CRM, the warehouse and the advertising platforms report different revenue figures produces meetings about which number is right rather than decisions about what to do — and that cost exceeds any tool's licence fee.

The AI-era additions — measurement harnesses, agent integrations, MCP endpoints — are increasingly built rather than bought, which is what the marketing engineer role exists to do.