Build vs Buy
An Agentic TMS Decision Calculator
Four paths. One unbiased model.
- Status quo. Keep what you have and absorb the drift.
- Extend and augment. Bolt decisioning onto your existing TMS.
- Build new & maintain. Build composable decisioning in-house.
- Buy agentic-native. Buy an architecture designed for the problem shape.
Run yours
Adjust any input. Math recalculates live.
How each path is priced.
Status quo. Annual bleed = your input × annual freight spend × horizon. The slider is yours to fill.
Extend and augment. Indicative $500K–$1.2M over the horizon, biased by your current TMS state.
Build new & maintain. Team payroll × horizon, using Robert Half's 2026 medians — $142K base for software engineers, $170,750 for AI/ML — fully-loaded at 1.35×. Expected adds a 50% buffer; high applies 2.5×, calibrated to BCG's 2024 finding on large-tech program overruns.
Buy composable. Indicative $1M–$4M over the horizon, modulated by complexity. The defensible number is your RFP shortlist.
If you have better numbers, use yours.
What the numbers tend to say
- Status quo rarely surfaces as the winner once a defensible bleed rate is entered. Until then, the calculator reports it as unpriced.
- Extend surfaces when complexity is low and current TMS state is modern or composable. Bolt-on integration cost is indicative — your real number depends on architecture.
- Build surfaces when IT capacity sits at the dedicated-or-larger end, time horizon is 5 years, and risk tolerance is aggressive. The 'expected' case already applies a 50% buffer per BCG 2024 data.
- Buy surfaces as the default math winner for most enterprise retail inputs. Vendor subscription is indicative — your real number is your RFP shortlist.
The math will not always say buy. When it does not, the calculator will not pretend it did.
The Build path's “expected” (+50% buffer) and “high” (2.5x) risk bands are calibrated to BCG's October 2024 finding that more than two-thirds of large-scale tech programs miss scope, time, or budget. Build payroll math anchors to the Robert Half 2026 Technology Salary Guide — $142K median for software engineers, $170,750 for AI/ML engineers, with a 1.35 fully-loaded multiplier per Robert Half methodology.
The transformation-failure context is anchored to Bain's 2024 survey finding that 88% of business transformations fail to achieve their original ambitions. Total transportation cost is referenced against APQC's cross-industry median of $8.33 per $1,000 revenue. Status-quo bleed has no defensible Tier-1 coefficient and is taken from your own input; bolt-on integration and enterprise TMS subscription ranges are indicative — defensible numbers come from your own RFP quotes.
