Most stacks are not bloated because someone bought bad tools. They are bloated because nobody ever ran the math on what to turn off. This is the math.
What this isA repeatable audit that turns your full martech inventory into a defensible keep, cut, or consolidate decision for every tool, sequenced into a phased rollout.
You give itYour tool list with costs, owners, and admin access (or read-only billing plus usage exports).
You get backA scored inventory, a one-word decision per tool, an overlap map, and a phased consolidation plan ordered by renewal date and switching risk.
The steps
- Build the raw inventory before judging anything. One row per tool: name, category, annual cost (list price), internal owner, contract renewal date, and notice-period. Pull this from finance and expense reports, not memory. Catch the shadow tools on someone's personal card. You cannot rationalize what is not on the sheet.
- Measure real usage, and mark the gaps honestly. For each tool get one usage number that means something: active vs licensed seats, records processed, sends, API calls, last-login dates. If a tool exposes no usage data and you cannot infer it, write UNKNOWN. Do not estimate. A guessed 60% that is really 10% kills the audit's credibility.
- Map the data flows and find the overlaps. Draw who sends data to whom, the sync direction, and the system of record for each object. Two things fall out: integration debt (the tool nobody can remove because others read from it) and functional overlap (group tools by the job they do; any group with two or more is a consolidation candidate).
- Score every tool on the rubric, 1 to 5 each: Usage (adoption vs what you pay for), Business Value (does removing it break a revenue or reporting workflow), Overlap (high overlap scores low, it is redundant), Cost vs value, and Switching cost (high switching cost scores high, sequence it carefully). The point is that the score comes from the inventory, not the loudest opinion in the room.
- Assign one decision per tool, driven by the score. KEEP: high usage, high value, low overlap. CUT: low usage, low value, high overlap, or duplicate spend. CONSOLIDATE: its job is covered by a tool you are keeping. MEASURE-BEFORE-DECIDING: any tool with UNKNOWN usage gets a 30-day instrumentation task, then a re-score. Never guess utilization to force a verdict.
- For each consolidation group, pick the survivor and prove it covers the work. Do not pick on price alone. List the specific features and data each loser provides, then confirm the survivor handles them. If the missing piece is load-bearing (a compliance field, a key integration), that gap becomes a migration task or the consolidation does not happen.
- Sequence the plan by renewal date and risk, not by savings size. Cancel CUT tools at their next notice deadline (a missed renewal costs a full year). Order CONSOLIDATE moves so low-integration tools go first and systems of record go last. Never run two migrations that touch the same data object in the same phase.
- Lock in the result and set the recheck. For every cancellation, confirm it went through before the deadline, export the data first, and kill the integrations feeding it. Re-run the audit on a cadence and require a renewal date, owner, and usage source on day one for any new tool. Sprawl returns the moment the sheet goes stale.
Template to start from
INVENTORY + SCORECARD (one row per tool) Tool | Category | Annual Cost | Owner | Renewal Date | Notice-Period (days) Usage Metric | Usage Value (or UNKNOWN) | Data In | Data Out | System of Record For SCORE (1-5 each): Usage | Business Value | Overlap (low = redundant) | Cost-vs-Value | Switching Cost | Total DECISION: KEEP / CUT / CONSOLIDATE / MEASURE-BEFORE-DECIDING Consolidate into: ___ Phase: ___ Cancel-by date: ___ Annual $ freed: ___ CONSOLIDATION GROUP (one per overlapping job) Job: ___ Tools in group: ___ Survivor: ___ Reason: ___ Features/data the survivor must absorb: ___ Gap (load-bearing? y/n): ___ -> migration task or no-go PHASED PLAN Phase | Tools | Action | Owner | Trigger date (renewal/notice) | Risk | Rollback | $ freed