What GTM stacks actually look like
Operator benchmarks from 100,000 modeled scans — not vendor surveys, not analyst reports, not self-reported executive panels.
Every number on this page comes from running the StackSwap scan engine against 100,000 realistic GTM stack scenarios sampled from 12 common team archetypes. The data reflects what the engine actually produces — no cherry-picking, no vendor bias, no testimonials.
Median monthly recoverable waste for teams with 26–50 (largest modeled segment):
$8,680/mo
That's $104,160 per year, in the median case.
Across 100,000 modeled scans with realistic GTM stacks.
Methodology: every stack was generated from realistic templates matching how mid-market GTM teams actually buy, then scanned through the same engine that powers StackScan. Details at the bottom of this page.
Smaller teams have less waste because they have less stack. Mid-market teams (16–50 people) are where the waste curve peaks — enough complexity to create overlap, not enough to justify enterprise-grade cleanup.
| Team size | Median recoverable/mo | Median annual | 95th percentile |
|---|---|---|---|
| 1–5 people | $0/mo | $0/yr | $2,129/mo |
| 6–15 people | $80/mo | $960/yr | $3,388/mo |
| 16–25 people | $5,120/mo | $61,440/yr | $11,363/mo |
| 26–50 people | $8,680/mo | $104,160/yr | $19,326/mo |
| 51–100 people | $21,410/mo | $256,920/yr | $39,680/mo |
| 101–200 people | $25,890/mo | $310,680/yr | $65,795/mo |
| 201–500 people | $76,280/mo | $915,360/yr | $129,612/mo |
These numbers reflect teams with enough stack complexity to be worth scanning — they're not averages across every GTM team everywhere.
The scan engine consistently identifies the same patterns across thousands of stacks. Legacy enrichment and sequencing tools get replaced most often. AI-native alternatives dominate the replacement list.
Most replaced
Tools scanned stacks cut most often
- 100%54,381
- 84%51,663
- 100%18,740
- 100%12,978
- 100%9,060
Most recommended
AI-native tools the engine suggests as swaps
- 46%73,121
- 38%60,303
- 8%12,978
- 6%9,060
- 3%5,090
Across 100,000 scanned stacks, most tools are worth keeping. About 1 in 3 tools in a typical stack gets a change recommendation.
Teams upgrade AI capability without increasing spend
Savings come from replacing expensive tools, not radical cuts
Modeled across 100,000 stacks from 12 GTM team archetypes
Most common overlap pairs (refreshed v1.0.0)
The 100,000-stack refresh adds overlap-pair tracking. Each row is a tool pair the engine flagged as redundant in some percentage of stacks, with modeled annual recovery from picking one as canonical. Bigger sample = tighter percentages. Methodology.
| Pair | % of stacks | Median annual recovery |
|---|---|---|
| HubSpot Marketing Hub + Salesforce | 29.97% | $1,800/yr |
| Clari + Gong | 23.79% | $1,200/yr |
| Apollo.io + ZoomInfo | 20.49% | $2,400/yr |
| Outreach + Salesloft | 18.74% | $1,200/yr |
| Apollo.io + Outreach | 17.75% | $1,200/yr |
| Linear + Notion | 15.2% | $960/yr |
| Clearbit + ZoomInfo | 13.99% | $3,600/yr |
| Chorus + Gong | 12.98% | $1,200/yr |
| HubSpot + Marketo | 0.76% | $6,000/yr |
| HubSpot Marketing Hub + Marketo | 11.91% | $17,280/yr |
Recovery values are deterministic per pair (engine assigns a fixed annual save from the OVERLAPS table in data/tools.json). Variance lives in the detection rate, not the per-pair amount. Per-pair pages have the full consolidation analysis.
Methodology
Every number on this page comes from running the StackSwap scan engine against 100,000 synthetic GTM stack scenarios. Scenarios are generated from 12 realistic team archetypes (mid-market sales-led, growth-stage multi-channel, early-stage founder-led, dev-tools PLG, enterprise RevOps, AI-native modern, and others), with controlled variation in team size, industry, and tool selection. Each archetype includes core tools, common additions, optional tools, and realistic "legacy drift" — the tools teams accumulate over time that create overlap.
The scans use the same engine (scanStack()) that powers the public StackScan product. The distribution of verdicts, savings, and replacement patterns you see here is what real teams see when they run a scan against a similarly-shaped stack. The engine is deterministic — given the same inputs, it produces the same outputs.
This isn't industry data. It isn't a customer survey. It isn't analyst opinion. It's the output of running a specific scan engine against 100,000 realistic stacks. If you want to know what your stack would produce, run your own scan — it takes 60 seconds.
Refresh v1.0.0 (12 archetypes, seed 42): The current dataset is fully reproducible — run SIM_SEED=42 npm run simulate:100k to regenerate bit-identical aggregates. Full pipeline + limitations documented at /methodology.