Stop Planning Routes in a Spreadsheet That Assumes Your Competitors Do Nothing

Stop Planning Routes in a Spreadsheet That Assumes Your Competitors Do Nothing

Every route launch begins the same way. Someone builds a spreadsheet. It has a demand estimate, a fare assumption, a cost line, and a ramp-up curve. The spreadsheet says the route breaks even in month nine. Leadership approves it. An aircraft gets committed.

Then the market does what markets do. A competitor adds a frequency two weeks after your launch. Fuel moves 12%. The demand stimulation you modelled at 15% comes in at 6%. Month nine arrives and the route is nowhere near breakeven, and everyone acts surprised.

Nobody should be surprised. The spreadsheet was never a forecast. It was one hand-picked scenario, and usually an optimistic one, because the person who built it wanted the route approved.

The single-scenario problem

A spreadsheet forecast makes one giant hidden assumption: that the world holds still. Your fare assumption stays valid. Competitors don't respond. Demand behaves. Costs stay put.

In aviation, every one of those assumptions fails routinely. The question is never whether your base case is wrong. It's wrong. The question is: how wrong can it be before the route stops making sense? A spreadsheet can't answer that. It gives you a point estimate when what you need is a distribution.

This matters most for the carriers with the least room for error. A ten-aircraft regional operator betting one airframe on a new route is betting 10% of its capacity. If the route underperforms for a year while sunk-cost thinking keeps it alive, that can be the difference between raising the next round and not.

What simulation changes

Instead of asking "what does this route earn in our base case?", simulation asks "what does this route earn across a thousand plausible futures?"

Concretely, a Monte Carlo approach to route evaluation runs the same launch through a large number of scenarios where the uncertain inputs actually vary: demand materialisation, fare environment, fuel price, ramp-up speed, and, critically, competitor behaviour. The output isn't a single breakeven month. It's a distribution:

• The route is profitable in 68% of simulated futures
• Median breakeven: month 11, not month 9
• In the bottom quartile of outcomes, the dominant cause is a competitor frequency response within 90 days
• Downside cases cluster around one specific assumption, which tells you exactly what to watch after launch

That last point is underrated. Simulation doesn't just grade the decision, it tells you which assumption your route lives or dies on. If 80% of your failure scenarios involve a competitor retiming their morning departure, you have a tripwire to monitor from day one, and a pre-thought-out response instead of a panicked one.

Competitors are agents, not weather

The step beyond Monte Carlo on inputs is modelling competitors as decision-makers. A rival carrier isn't a random variable like fuel price. It responds to what you do, following its own logic: defend hubs, protect corporate share, avoid tit-for-tat on thin routes.

Multi-agent simulation encodes those behavioural profiles and lets the competitive dynamics play out. You launch. The simulated competitor evaluates whether your route threatens anything it cares about. It responds, or doesn't. You respond to the response. Run that game thousands of times and you get something no spreadsheet contains: an honest picture of the fight, not just the route.

We built SimLab, AvioIQ's simulation engine, around this idea, running route and network decisions through multi-agent Monte Carlo batteries before a single aircraft moves. The consistent lesson from that work: the routes that look best in a static spreadsheet and the routes that survive competitive simulation are not the same list. Static analysis systematically favours routes into contested markets, because it prices the market as it exists today, before your entry provokes anyone.

"We don't have the data for that"

The usual objection. It's less true than it sounds. Indian carriers have DGCA traffic data, published schedules, observable fares, and airport-level statistics. US carriers have Form 41 and DB1B. The raw material for calibrating demand and cost models at departure level exists in the public domain. What's been missing is the reconstruction and simulation machinery on top, priced for someone who isn't Emirates.

The honest caveat: simulation doesn't predict the future either. Nothing does. What it does is replace false precision ("breakeven month nine") with calibrated uncertainty ("seven in ten futures work, and here's what kills the other three"). That's a different quality of information to put in front of a board, an investor, or your own conscience before committing an aircraft.

Your competitors are not a fixed line in a spreadsheet. Stop planning as if they are.

SimLab is the multi-agent simulation layer of AvioIQ, built by Aviation Oasis. It stress-tests route and network decisions against thousands of competitive scenarios, calibrated on validated departure-level economics.