Isaac Mavis
For more than fifteen years, Isaac Mavis has built the pricing, revenue management, and AI systems that decide whether commercial businesses actually make money. At Allegiant he built the data science function from nothing and the systems that priced more than a billion dollars in annual airfare. He has owned commercial and revenue strategy at Priceline, Avelo, and Sun Country, and taken more than one company from no AI capability at all to production machine learning in a matter of months. A Chicago Booth MBA with degrees in mechanical and aerospace engineering, he is the rare operator who can design the model, build the system, and sell the result to the board.
Margin House is that work, now offered to a small number of clients at a time.
Start a conversationWhat we build
Commercial systems
Pricing engines, revenue management workflows, and network optimization fitted to a specific business. Not slideware. Working systems the commercial team can operate and trust.
AI and machine learning
Zero-to-one ML where there was none, forecasting and decision automation in production, and GenAI and LLM tools aimed at the few places they actually pay off. Built, deployed, and explained to the people who fund them.
Built to be run
Every engagement leaves something the client owns and can extend, with the founder personally accountable for the work. We will even price on results alone: no result, no fee. No layers, no surprise staffing, no extraction.
Selected work
Read the full track record →Revenue management, built from scratch
A machine learning revenue management system designed from nothing, the pricing brain behind more than a billion dollars in annual airfare.
The best margins in airline history
A modular network planning suite that solved capacity, scheduling, and crew pairing together, helping Allegiant post some of the highest operating margins in commercial aviation.
From launch to first profits
A custom in-house revenue management system that lifted revenue an estimated 14% and helped drive the airline's first profitable quarters.
Recent thinking
All thinking →What AI actually does in revenue management
An honest account of where AI earns its keep in RM, and where the vendor pitch outruns reality.
Why your best RM analysts keep leaving
Why your sharpest analysts leave, and the margin hiding in the work they were doing.
The network planning decision no one fully owns
How frequency decisions get made without the full P&L in the room, and what it costs.