I've read several posts on "Married Segments" across various blogs and there seems to be a misconception about what they are and how they are used in the industry. Most of the discussions I've read about "Married Segments" are actually talking about the difference between Local (single segment) Availability and Flow (connecting flight) Availability.

As most folks on here know, fares are tied to various inventory codes which are limited to 26 (A thru Z) by industry standard. Some carriers have enahnced their systems to allow for additional codes, but those can be used for internal purposes only (they don't transimt between other airlines or the GDSs).

Given that fares aren't generally priced on an equivalent revenue/mile system from market to market, selling a seat to a customer flying in one market is actually more valuable than selling the same seat to a customer flying in a different market. Let's use an example:

DCA-ATL with a Q fare of $200 OW
ATL-PHX with a Q fare of $300 OW


If DCA-PHX had a Q fare of $500 OW AND there was equal probability of demand from one passenger for DCA-PHX as for one passenger on DCA-ATL and one passenger on ATL-PHX, the Q seat has the same value to the airline regardless of which passengers buy the Q seat on the two flights.

Where things get interesting is when the DCA-PHX Q fare is $700. As such, this passenger is more valuable and the DCA-PHX Flow Availability in Q will be higher than the Local Availability in Q for DCA-ATL and ATL-PHX.

When the concept of Flow and Local Availability was first introduced in the 1980s, many travel agents and airline employees figured out that they could book connecting flights which had higher availability and then cancel the unwanted segment(s) to obtain a seat at a fare in a market that wasn't actually available. For example, booking DCA-ATL-PHX then cancelling ATL-PHX would allow you to then price the DCA-ATL fare that wasn't actually availabile.

To address this loophole, the industry implemented Married Segments which, as their name implies, means that they are coupled. If you book DCA-ATL-PHX, the segments are married to one another and cancelling one generally can't be done without cancelling both.

Of course, it is also possible that the example above could be reversed where the DCA-PHX Q fare is only $400 (less than the sum of the two local segments). In this case, Flow Availability in Q for DCA-PHX would be lower than the Local Q Availability for the DCA-ATL and ATL-PHX segments.

Again, many agents tried to game the system by booking DCA-ATL first then booking ATL-PHX as a second transaction, but in the same PNR. This would then allow them to obtain seat to price the $400 fare that wasn't available in DCA-PHX. To address this problem, the industry developed an additional solution known as Journey Data.

With Journey Data, any previously booked segments in the PNR are passed along with newly requested segments to a carrier's inventory system. The inventory system then detemines if the new segment forms a connection with any of the previously booked segments. If so, the avaialibilty for the new segment will be based on the Flow Availability for the connection, not the segement's Local Availability.

So if you book DCA-ATL and then request ATL-PHX, the system will see DCA-ATL is a connectio to ATL-PHX and the ATL-PHX Q availability would be based on the Flow Availability for DCA-PHX. Likewise, if you book ATL-PHX first then try to add DCA-ATL in front of it, the DCA-ATL availability will be based on DCA-PHX.

In instances where the inventory system allows the new segment to be booked, it will set a "New Marriage Condition" on the previously booked segment(s) that was deemed to be part of the connection, thus protecting the Flow Availability decision.

Hope that helps.