Toward the end of my tenure at the airline I arrived at three conclusions:
Ads do not create demand for travel
Airlines compete primarily on price and schedule, and (to a lesser degree) experience; demand for the product is baked into the market
If an airline is selling tickets via online travel agents (Expedia) the emphasis is almost entirely on price and schedule
Southwest airlines does not sell via OTAs, therefore their annual ad budget is usually as much as every other domestic airline combined
Since sales are primarily factors of price and schedule, there is a slim (and probably inconsequential) margin of customers willing to pay more for a differentiated experience
The most effective way to do this would require lockstep cooperation with the revenue management team (i.e. deliberately list fares at higher prices than other airlines in test markets, with the thought being advertising the differentiated experience would sell the higher priced fares and fill the seats)
These people generally want nothing to do with you as a marketer
Ads could be used to cross-sell or upsell existing customers on additional products
To further explore #3 above, I started experimenting with custom audiences on Meta.
This is where I describe what is possibly a privacy crime? I don't think it was several years ago, who know now, I'm not a lawyer.
After years of experiments, I failed to find ways to scale the entire budget into ads that would consistently generate incremental revenue. However, the system in place to upsell premium products to existing customers, or cross sell vacation packages, was exclusively email based. Given the airline's promotional email open rates (ones that included merchandising for premium economy seats) were pretty abysmal (sub 40%), it left room for us to conduct a little experiment:
Objective: Determine if showing premium economy ads to customers who had purchased economy tickets with departure dates in the next two weeks increases the upgrade rate to premium economy seats
Method: Generate a custom audience list, randomize the list into test and control groups. Upload the test group into Meta as a custom audience. With reach as the goal setting,
Outcomes and notes:
The test did not generate lift to premium economy sales in the test group
75% of the total list was successfully matched into Meta
90% of the matched users were reached in less than 2 days
Given the abysmal open rates of the promotional emails, this is a relatively inexpensive way to merchandise product to people who otherwise wouldn’t see it
Running reach campaigns against custom audiences is cheap
You could be a business with tens of millions of customers and it would still be cheap
Our creative was terrible and there was no offer
DGAF, was just trying to get something live
Site merchandising and simplifying the path to purchase are the only ways to make this work.
As an aside here, I find it funny that you join up with a company like an airline and are given free reign to use the product whenever you want (free standby flights are dope). As such you are part of the larger mechanism of selling things, and should thus be experts in that function. Like, you should know why is the airline better than others, how do you buy stuff from that airline, is that a good and easy experience. But the only way to book a flight as an employee is through a service that is entirely separate from the customer facing experience (whattup myidtravel freaks). How is anyone supposed to know if the purchase flow is any good (or if it’s flat out broken) if you haven’t used it in years? But I digress...
In conducting this experiment I discovered that there was no simple way for a customer to self-service upgrade to a premium economy seat after purchasing an economy ticket.
The page on the site that merchandises the premium economy product literally did not have an exit link to complete the upgrade. Merchandising for this product was included in every itinerary email, and those units linked out to this page, and there is no good way to sell someone this product on the website. No one I spoke to had any idea this was a problem.
Suffice it to say that I am not surprised that this test did not generate lift, but I remain convinced that it could. Not only that, but I think most large companies are leaving a ton of revenue on the table by delegating cross sell and upsell responsibilities to CRM teams and leaving them email and other owned channels as the only resources they have to accomplish their goals.