Revenue Architect Is Not A New Job

When demand softens on a route, something familiar tends to happen across most commercial organizations: the pricing analyst lowers the fares, revenue management opens more inventory, and marketing authorizes additional spend to stimulate demand. Three separate functions, three responses to what might be the same underlying problem, with no one necessarily asking what the underlying problem is or whether all three responses are working toward the same outcome. Melissa Skluzacek calls this the double dip, sometimes the triple, and she sees it as the clearest indicator of how far most commercial organizations still are from a different way of working.

The revenue management function sits at a consequential intersection: observing what network planning, scheduling, pricing, marketing, and digital are doing, then watching whether any of it translates into bookings and revenue at the level of individual flights and days to departure. Few other functions have that view. In organizations Melissa describes as more mature, analysts are beginning to use it differently: not just asking what lever they can pull, but what is causing the problem in the first place (a pricing issue, a demand issue, or a competitive one) and then working across functions to address it rather than each team solving their own narrow version of it.

The constraint has always been time. Analysts carry responsibility for hundreds or thousands of flights, and there is a ceiling on how much any one person can see clearly at that volume. This is where Melissa sees AI’s most direct contribution: compressing the distance between raw data and usable insight, so that more of the analyst’s day goes toward solving the problem rather than locating it. “The future revenue architect is someone who spends less time hunting for information and more time solving problems and coordinating across teams.” The point is to free up judgment, not to sideline the person exercising it.

Revenue management has always carried a tension between science and something harder to articulate. Two experienced analysts can sit with the same data, work through the same inputs, and arrive at slightly different conclusions, and both might be defensible. Markets have personalities; years of reading a specific set of routes produces a kind of intuition that does not transfer cleanly into a model feature. “Revenue management is both an art and a science. We haven’t gotten rid of the art, even with all this technology and AI.” The practical implication is that a good commercial framework needs to account for this rather than try to resolve it. A common analytical starting point from which individual judgment can operate is what the industry has not built consistently, yet this is what the revenue architect role depends on.

What Melissa wants to see more of is what she calls an experimental orientation. Before taking an action, state what you expect to happen, how you will measure whether it worked, and what would cause you to change course. This matters especially because revenue management is not made up of a handful of large decisions; it is hundreds of small ones over the life of a flight. “An analyst could make an intervention today, another one next week, and react to a competitor move two weeks later, and it really becomes difficult to untangle what drove the final outcome.” Structuring decisions this way, she says, is how intuition gets calibrated over time rather than remaining purely informal. This is one of the things AI can support, by making patterns in past decisions visible rather than leaving them to memory.

There is a measurement gap underneath all of this. Airlines measure outcomes such as seat revenue, yield, load factor, ancillary, and those metrics are essential because they reflect what the business produces. What they do not reveal is whether the revenue management decisions inside them were good ones. “AI isn’t creating a measurement problem; it’s exposing one we’ve always had.” The case she makes is not about what technology should be built to address this, but about what discipline the function has been deferring: being clear, before an action is taken, about what success would look like and how you would know.

The theoretical optimum sits permanently out of reach. This is simply a fact of life in airline revenue management. “We do not know what the optimal revenue is for any sort of flight. Human purchases are irrational and there are so many different variables. We’re just trying to get as close as we can to this theoretical optimum.” The RM team is not observing the market from the outside; their decisions are part of what the market is, shaping the booking curve that subsequent decisions will have to read. The analyst who understands that (the one who is accountable for the prices customers see and can explain why those prices are what they are) is the one who can work well alongside AI rather than defer to it.

Which brings the story back to where it started. The double dip, the triple dip, three functions each solving their own version of a problem that might have one root cause. This is the structural condition the revenue architect role is designed to address. Not by owning every function, but by having the vantage point and the cross-functional fluency to see pricing, demand, competition, distribution, and ancillary as a single commercial problem rather than five separate ones. Whether AI created that role or simply made it more urgent is almost beside the point. The need has been visible for a long time. The question is which organizations will build the conditions for it to exist.