81.66% of modeled B2B SaaS GTM stacks contain at least one redundant tool pair.

Median team is sitting on $93,240/yr in modeled recoverable spend. We modeled 100,000 stacks across 12 operator archetypes. Open methodology, reproducible from one command.

81.66%

of modeled stacks contain at least one redundant tool pair

$93,240

median annual recoverable spend per stack

100%

of stacks containing Outreach are flagged for replacement

96.05%

of stacks containing Mailchimp flag it as redundant

$33,000

median annual overlap if Marketo and Pardot share a stack

Methodology in three sentences

100,000 synthetic GTM stacks are generated by sampling tool combinations from 12 operator archetypes (early-stage B2B SaaS, mid-market sales-led, enterprise RevOps, AI-native modern team, and 8 others) weighted by realistic population prevalence. Each stack runs through scanStack() — the same pure scoring function that powers the public StackScan tool — producing per-tool verdicts, overlap detections, and modeled recovery. Same seed + same engine = bit-identical aggregates; the dataset is committed in the open repo and runs in ~753 seconds.

Why synthetic, not customer data

StackSwap is pre-revenue. Citing customer scans we don't have would be misleading. Synthetic stacks let us span the full operator distribution — including patterns a real customer base would underweight — and the methodology is fully reproducible. The numbers should be read as “modeled directional truth,” not empirical fact, and the limitations are documented at /methodology.