There is an ancillary revolution in the hospitality industry. A recent Oracle study found that 81% of hoteliers surveyed expect a big service model shift between now and 2025, and 49% strongly agreed that special amenities and upgrades are critical to their revenue strategy.
People often liken this shift to airlines, but perhaps a better analogy is retail. Retailers – from big box to social media sites – have trained consumers to not just accept upsell or ancillary offers, but to look out for them. For hoteliers, this is a good thing. These companies, and others, make it possible for hoteliers to position upselling as a means of creating a better experience for their guests, not just hawking products and services.
But how are these retail companies upselling so successfully? And how can hoteliers emulate it? Machine learning. This form of artificial intelligence enables hotels to capture and apply data to present relevant offers to guests in real time and maximizing total revenue per guest. This type of data isn't the usual sort that hoteliers tend to fall back on. Historical data, especially in post-COVID times, is less valuable than it once was. What's more valuable is data that illuminates why the guest is taking this trip right now, so a deep understanding of reservation data is critical – market segment, source of business, rate code, day-of-week check-in, etc.
As guests move through the booking and pre-arrival phase, every additional guest interaction is valuable data. Did they click on the loyalty offer on the booking engine confirmation web page? Did they ignore an upsell offer for a suite on the confirmation email? Did they click on an early check-in offer on the pre-arrival email? Did they pre-register on their mobile device?
As a system begins to receive data during the reservation life cycle, personalization – even for guests who have never booked with your hotel before – becomes possible. Personalization at this level identifies different things about a guest all at the same time to understand what combination of attributes that guest values, at what price, and at what point in their journey the guest will be most likely to upgrade or request services or products.
When a guest checks in, the system can take into consideration what was already offered to the guest earlier in the reservation life cycle and how the guest reacted. The closest way to get to true personalization is using these buying signals specific to an individual guest. With that knowledge, the system can present recommendations to the front desk/reception agent in real time to inform their upsell offers for each guest.
Every guest wants a different experience. Some are willing to spend on a suite, some on food and beverage, and some on other offerings. What hotels offer will vary widely; a hotel with an attached water park doesn't want guests to leave the property, so will offer park-view rooms, or pet passes, or pool-side cabanas. A city-center property attached to a mall may have partnerships with retailers or entertainment providers in the mall or offer marked-up shipping services for all those packages bought at those retailing partners. A hotel in a location with lots of outdoor activities may become a reseller for local tour and activities providers.
One of the greatest values of machine learning is the feedback loop; that is, writing the guest interaction with the offer back into the data so the machine can learn what guests accept – and just as importantly, what they reject – to make better offer decisions in the future. The more buying signals the system can receive, the more targeted merchandising offers can become.
Ancillary vs. upselling vs. cross-selling – it's all the same: revenue generated by offering guests products and services relevant to their stay. The only way to generate sustained and forecastable incremental revenue is to automate it using machine learning.