the business case
the business case · hyper engineering, in budget terms

Make the maths work without the headcount.

You have a backlog you can’t staff against, a hiring market that doesn’t want to be hired into, and a board that still wants the roadmap on schedule. The line item won’t move. The output has to.

This page is the case for nmux as a budget decision — the numbers, the model, the worked example, and a calculator you can run against your own team in the next ninety seconds.

↳ the bottom line
2.2× throughput per engineer-month (expected · modeled · 3.4× upside) ~ 11 wk payback period (modeled · 12-eng team) 4–8 hires deferred (year one · modeled)
↳ modeled projections · documented assumptions · run your team's numbers below
§ 01 — the budget line that won’t move

Engineering is your single largest variable cost. It’s also the one you can’t scale.

$220K+
fully-loaded annual cost per senior engineer
↳ salary, equity, benefits, infra share
5–9 mo
time-to-productivity for a senior hire
↳ from offer signed to first owned ship
1 in 4
open reqs you'll close to plan
↳ industry benchmark · 2024 talent survey

You can’t hire your way out of this. Not at this rate, not at this price, and not on the timeline your roadmap promises. The only path that scales is output per engineer.

§ 02 — the leverage

What you’re actually buying. Leverage on the line you already have.

nmux doesn’t replace your engineers — it’s an exo-suit they wear. It removes the tax they pay on the hours they already work. Three places it shows up on the line.

01

Headcount leverage

before Hire 4 engineers to ship 4 engineers' worth.
with nmux Hire 0. Ship 4 engineers' worth from the team you have.
↳ modeled: 12-eng team absorbs scope of 18 in one quarter at expected adoption
02

Time-to-value

before 12-month SaaS rollout, 4-quarter procurement cycle, 3 vendors to integrate.
with nmux Single static binary. Pilot in 2 weeks. Production in 6.
↳ modeled deployment: 41 days, pilot to production
03

The defensible budget

before Soft promises about "productivity gains" you can't substantiate.
with nmux A worked TCO model your CFO can sign, with deterministic outputs they can audit.
↳ every run reproducible · sha-256 verifiable
wondering how nmux compares to harnesses and platforms on the line? see the comparison table →
§ 03 — the calculator

Run your team through it. Get a number you can defend.

Choose a scenario, pick a framing, adjust your numbers. The model uses an adoption ramp for year one and steady-state math for year two+. Every assumption is visible — export for your CFO.

live · roi model · v2 ↳ all figures in $K USD · annual basis
scenario
framing
↳ team basics
Engineering team size 12engineers
Fully-loaded cost per engineer 220$K / year
Features shipped per quarter today 8features
license tier
↳ cost side
One-time setup, integration, training 25$K
Senior eng time reviewing AI output 10%
↳ growth framing · your numbers
Features producing measurable impact 30%
Average annual value of one successful feature 150$K
↳ outputs
Features added (annual)
year 1
96
steady
0
Blended multiplier 2.0× yr1 → 2.2× steady
year 1 steady
Annual value $0K $0K
Net benefit $0K $0K
ROI $0K $0K
Payback $0K
cost breakdown annual
License $144K
Implementation (one-time) $25K
Review overhead $0K
Year 1 total $0K
Steady-state $0K
↳ adoption ramp: months 1–2 at 1.2×, months 3–6 at partial uplift, months 7–12 at steady state · growth framing assumes customer's own success rate and value-per-feature numbers · efficiency framing counts avoided future hires, not layoffs · model documented in TCO PDF
§ 04 — worked example · series-b saas shape

A 14-engineer team. One quarter. What the model projects.

A worked example derived from the calculator above — expected scenario, 14-engineer team, $220K average loaded cost. The shape your finance team will see. Plug your own numbers in §03 to model your team.

↳ ledger · q3 · 14 engineers · modeled $220K avg loaded · series-b saas shape
Engineering line (quarter) $770K ↳ 14 × $220K / 4
Features shipped — baseline 9 ↳ team's prior 4-quarter avg
Features shipped — with nmux 31 ↳ model output · 3.4× · expected scenario
New features delivered +22 ↳ absorbed scope from Q4 + Q1
Hires deferred (year-end) 5 ↳ $1.10M annual cost avoided
nmux license (annual) $144K ↳ Team tier · 25 seats
Year-one net benefit +$1.36M ↳ $1.10M hires + $400K added value − $144K license
Payback period 9.5 wk ↳ license cost / monthly value
↳ what you'll defend in your next board review
22 features of additional scope, absorbed this quarter ↳ from Q4 + Q1 roadmap $1.36M year-one net benefit ↳ hires + value − license Payback in under three months ↳ 9.5 weeks · documented model A model your CFO can challenge — and re-run ↳ every assumption visible
↳ run the model on your team's actual shape Request a walkthrough
§ 05 — proof points

The line items behind the line.

01

No SaaS in your supply chain

Single static binary in your environment. No vendor lock-in to negotiate later.

02

Audit-ready by default

Every artifact, every routing decision, every agent action — SHA-256 verified, in your repo.

03

Predictable in finance speak

Output per pipeline run, not per developer mood. Forecasting becomes possible again.

04

Standards as code, not aspiration

Engineering standards become enforceable gates — enforced or the pipeline doesn't advance.

05

Throughput compounds

Each run accumulates an ADR + spec + test bundle. Future runs reuse it. The team gets sharper.

06

No telemetry leaving

Code, prompts, and outputs all stay in your network. Your IP is yours, no asterisk.

§ 05b — what your finance team will ask

The questions your CFO will bring back.

We’ve sat through enough budget reviews to know which questions come up. Here they are, with the short version of each answer.

q · 01

What if we already have Copilot or Claude Code?

Keep them. nmux is the layer above — orchestrates them into pipelines, enforces your standards, produces audit trails. Your engineers don't change their tools; you change what those tools do at the team level.

q · 02

What's the risk if nmux goes away?

Single static binary. Uninstall and nothing breaks. Engineers fall back to the tools they were already using. No vendor lock-in to negotiate later, no migration project if you ever switch.

q · 03

How does this scale beyond 14 engineers?

License scales by tier. The model in §03 runs to 120-engineer teams. Pipeline architecture is parallel by design — your scaling limit is the compute capacity you provision, not the license seat count.

q · 04

What's the variance on these numbers?

Conservative 1.5×, expected 2.2×, optimistic 3.4× — every assumption visible in the calculator above. Plug your team's actual shape in. Documented adoption ramp, no hidden multipliers.

§ 06 — your move

Run the model on your team’s shape.

30 minutes with one of our engineers. We’ll plug your team’s numbers into the calculator, walk through the assumptions, and answer the questions your finance team will ask. No slides, no follow-up sequences.

↳ also reading? see the technical case nmux © mmxxv · the exo-suit for your hyper engineering team