Selling Points: How Predictive Analytics Can Drive Ancillary Revenue


Illustration by Angelica Geisse

APEX Insight: Airline retailers are following in the footsteps of brands like Amazon and Target by using predictive analytics to offer a better customer experience – while making food fly off the galley shelves.

Target, the department store, knew a teenager was pregnant before her father did by analyzing patterns in her purchasing behavior. The retail giant had hired statistician Andrew Pole to be its director of Guest Data Management to ramp up its predictive analytics department in 2002. Over the next eight years, Target’s revenues grew from $44 billion to $67 billion due, in part, to the “heightened focus on items and categories that appeal to specific guest segments, such as mom and baby,” according to the company’s former CEO, Gregg Steinhafel. 

The retail industry has been using data in sophisticated ways to predict and influence consumer behavior for decades. Now companies like Black Swan Data, gateretail and Guestlogix are applying a similar strategy to help airlines generate ancillary revenue and provide a more personalized experience for passengers.

“This isn’t reinventing the wheel – it’s been done by Amazon, but now we can start bringing this into the in-flight environment” – Patrick Prefontaine, Black Swan Data.

“Let’s say every time you finish a sad movie, you drink a certain soda,” suggests Patrick Prefontaine, Black Swan’s chief commercial officer of Avionics. “The platform will recognize that you’ve started watching a sad movie. When it finishes, there should be a button that pops up offering you that specific beverage. This isn’t reinventing the wheel – it’s been done by Amazon, but now we can start bringing this into the in-flight environment.”


Prefontaine is referring to Black Swan’s Connected Traveler Solutions vision, which enables airlines to understand consumer behavior patterns based on data that is captured during passenger transactions. “It could be selecting a certain film, or the way they’re exploring the content,” says Gary Townsend, Black Swan’s chief technology officer. “Anytime a passenger interacts with the system, we’re starting to learn a little bit more about their behavior and mapping that to a persona.”

The platform is being developed in partnership with in-flight retailer gateretail and its parent company, gategroup. It can track passenger interactions with the graphical user interface on seatback screens and with applications on travelers’ personal devices. 

“Anytime a passenger interacts with the system, we’re starting to learn a little bit more about their behavior and mapping that to a persona” – Gary Townsend, Black Swan Data.

While travelers could choose to log in to receive a fully individualized experience, they can also be filed into a persona based on their behavior without being identified. “You can imagine a plane on a particular route having 15 to 20 different personas on board,” Townsend says. “One is a wine connoisseur who enjoys sharp cheddar cheese and action-based movies, while another loves diet cola, turkey sandwiches and comedies.” The idea is that a passenger’s in-flight experience is shaped by the persona that’s assigned based on that individual’s behavior.


“Without visibility into what passengers are actually doing during the day of their journey, airlines are essentially selling merchandise and hoping that passengers happen to like the catalog,” says Jamie Dinsmore, senior vice-president of Global Sales for Guestlogix. “Airlines already have the data; they just need the retail analytics side to put their data to productive use.”

Guestlogix’s airline commerce platform, which consists of a management console, a flight attendant point-of-sale system and a passenger application, is currently flying on various airlines, including Swoop. The management console includes a retail analytics engine for mobile ordering, which launched in April; it’s what enables personalized, data-driven campaigns based on machine learning. 

During the product research phase of building the platform, Guestlogix found that 94 percent of airline professionals were confident the new system would improve passenger experience and 86 percent believed it would increase ancillary revenue.

“Without visibility into what passengers are actually doing during the day of their journey, airlines are essentially selling merchandise and hoping that passengers happen to like the catalog” – Jamie Dinsmore, Guestlogix.

“The platform allows airlines to put their passengers at the heart of every decision, and to utilize their data to become true retailers across the entire journey, not just in the sky,” Dinsmore says. “Passengers can revise their meal order, for example, while waiting in line for the lavatory.”

When luxury brand Burberry started using in-store tablets a few years ago, retail sales shot up 15 percent year-on-year, and orders taken on iPads in store accounted for more than a quarter of digital sales made during the period. “With tablets, retail associates have visibility into information such as online inventory, who’s walking into the store, what customers have bought in the past, what they recently searched for and whether they are a loyalty program member,” Dinsmore explains. “Imagine what tablets could do for flight attendants.”

Selling Points: Predictive Analytics Using Passenger Data
A visualization of data left behind by passenger activity, which Black Swan calls Passenger DNA.


A recent article published in Harvard Business Review extolled Amazon’s ability to anticipate what customers want, calling Amazon Prime “the gold standard for one-to-one customer marketing at scale, thanks to innovations in machine learning.” Part of the e-commerce innovator’s strategy is exploring how to apply data to create operational efficiencies.  

In December 2013, Amazon obtained a patent for anticipatory shipping – a delivery method that begins before customers complete a purchase. It allows the online giant to predict purchases based on factors like previous orders, product searches, shopping cart contents and length of time the cursor hovers over an item. According to the patent, the packages wait at hubs or on trucks until the order is completed.

In addition to surprising and delighting passengers by forecasting their future purchases, machine-learning algorithms can simultaneously help airlines stock inventory more efficiently. According to Steve King, CEO of Black Swan Data, the goal of its partnership with gategroup is twofold: enable an expanded product offering that is better aligned with passenger expectations and reduce waste through advanced demand forecasting.

“The solution leverages large data sets and advanced analytics to give customers choices that create a unique experience without increasing wastage or fuel use through overloading,” says Simon de Montfort Walker, gategroup’s chief technology officer and president of eGate Solutions. “[We] integrate these real-time optimizations into our catering units and our supply chain.”


By analyzing social and historical data, Black Swan has informed product development at multinational companies like PepsiCo. The data company’s Trendscope tool identifies popular ingredients and predicts which ones will scale in the next six to 12 months. The algorithm is trained on seven years of data and is able to make predictions with 90 percent accuracy up to six months ahead, according to Townsend.

“We can forecast six months out what kind of products people will want to buy,” Townsend says. “Airlines would benefit significantly by knowing and understanding what kind of food and duty-free products they will need to stock their aircraft with months ahead  of time.”

PepsiCo used Trendscope to develop new flavors of Sensations chips. The products have yet to launch, but received exceptionally high scores when tested with a group of consumers, according to James Howarth, former strategic insights director at PepsiCo. 

Perhaps passengers ordering a soda after a sad movie could do with some chips, too… Just a suggestion.

“Selling Points” was originally published in the 8.4 August/September issue of APEX Experience magazine.