MarTech stack triage: when to consolidate, when to add, when to rip out

If a marketing team has been operating for more than four years, the MarTech stack is no longer designed; it's accreted. Tools were bought for a campaign, kept for a quarter, forgotten for a year. Here's how to triage the result without either over-consolidating or over-spending.

The three categories every tool falls into

Run an audit and put every tool in one of three boxes:

• Workflow critical: someone uses it weekly, it's tied to a revenue process, and removing it breaks a campaign. Keep, regardless of cost.
• Workflow optional: someone uses it monthly, the output is nice-to-have, and a more expensive tool covers 80% of its function. Negotiate or consolidate.
• Workflow theatre: licensed because it was free with another tool, or because a vendor pitched it well in 2024, and nobody can name the workflow it supports. Cancel.

The mistake most consolidation projects make

Procurement-led MarTech consolidations optimise for licence cost and end up breaking workflows. The team rips out a niche tool that "the all-in-one platform also does" — and discovers a quarter later that the all-in-one's version of that feature is two years behind and half-functional. The savings on licences cost more in marketer time than they recovered.

The fix: do the audit workflow-first, not licence-first. Start from the campaign, work backwards to the tools. If two tools both support the same workflow well, then consolidate. If one tool barely supports a workflow that's critical, the niche tool stays.

The three buys worth making

For most mid-stage growth teams, three categories of tool consistently under-spend relative to their value:

• Server-side event tracking and CAPI infrastructure. iOS attribution is broken; client-side cookies are degrading. Server-side is no longer a nice-to-have.
• A real CDP, even a small one. The cost of segmenting customers properly is dwarfed by the cost of not segmenting them.
• An LLM-backed content production tool with a brand voice guard. Volume is the lever; brand consistency is the constraint. Both need tooling.

The two-year cycle

MarTech stacks need triage every 24 months. The vendor landscape shifts faster than that, but workflow change at the team level is slower. Triage cycles longer than two years compound the accretion problem. Triage cycles shorter than 18 months introduce churn the team can't absorb.

Run the audit. Categorise every tool. Kill the theatre. Re-buy where the workflow is exposed. The point isn't a smaller stack — it's a stack that maps to the campaigns the team actually runs.