When you buy online, the logic is simple: define what matters to you, then compare a manageable set of relevant options. Flight shopping is the opposite. You’re thrown into a maze of routes, prices, ancillaries, and often disappearing deals long before anyone has really understood what “good” looks like for your trip.
That isn’t just bad UX; it’s a structural problem in how airline products are priced and distributed.
The fare is not just a price tag
Airline fares were built for a world where airlines had one blunt instrument to manage demand: the price attached to a seat on a specific route on a specific date. To make that work at global scale, the industry invented a highly structured “language” of fare rules and codes so computers could interpret what each price actually buys.
The fare is not just an amount; it is a set of rules and conditions (often hundreds) that need to be met so the “price tag” is valid. Taken together, these rule categories act like a technical grammar or programming language for pricing, but are opaque to almost every traveler.
The search space is computationally explosive
Flight shopping is a massive, constantly changing search problem. For even a “simple” round trip, the number of valid flight combinations can range from ten thousand to over one hundred million, and for each combination there can be tens of thousands to trillions of possible fare combinations once rules are applied.
That means no shopping system can truly “try everything” for every search. They must rely on heuristics and caching, inevitably trading completeness for speed.
Why we never see all the possibilities
Given these constraints, modern flight search systems do not really explore all pricing. Instead, they generate and cache sets of “likely” itineraries and prices based on recent demand patterns and rule structures, cutting off travel options that look “unpromising” long before every flight combination has been evaluated.
The net effect for travelers is that what you see in a shopping screen is not all the options, but a curated subset of options that were relatively easy and cost-effective for the system to compute, cache, and distribute.
New distribution, old experience
The industry’s answer to legacy constraints was IATA’s New Distribution Capability (NDC), an XML-based standard that lets airlines construct offers themselves and push richer content and ancillaries into indirect channels. In theory, NDC should allow airlines to personalize offers, bundle services dynamically, and move beyond the static fare-filing logic. In practice, NDC has been layered onto the existing distribution stack rather than replacing it.
So, from a traveler’s perspective, NDC has not yet delivered a visibly different shopping experience: most still start with origin, destination, and date, then face a wall of prices and schedules with limited evidence of true personalization.
Living in a sea of cached, expiring results
On top of this structural complexity sits an operational reality: much of what we see in flight search is pre-computed and stale by the time we act on it. To keep response times acceptable, retailers and intermediaries rely heavily on cached availability and pre-priced itineraries, only validating prices and inventory against airline systems when the traveler clicks through to book.
From the traveler’s point of view, this feels like a bait-and-switch: you invest time sifting through pages of options only to discover, several clicks later, that the outcome was never guaranteed.
Why the usual “shop then filter” model fails
Taken together, these factors explain why flight shopping feels so unlike other e-commerce. When the traveler is pulled straight into an environment with nothing more than origin, destination, and date, we’re asking them to solve a problem that even large-scale optimization engines can only approximate.
This is why adding more choice, more filters, or more content into the same old grid does not fundamentally fix the experience; it simply moves more of the burden onto the traveler and their patience.
Start with “what matters”, then compare only what’s relevant
If flight shopping is structurally different, the process needs to be structurally different too. Instead of throwing every combination of schedule, price and attributes at travelers and asking them to figure it out, we should invert the flow: resolve intent first, then surface only the offers that are truly relevant.
In concrete terms, that means understanding the value of attributes like total travel time, tolerance for connections, flexibility, and other requirements before generating a meaningful shortlist.
The way out is not to show everything faster, but to show far less (but more relevant), because we finally took the time to understand what “good” looks like before we started to search.