In the world of travel shoppers, personalisation is the key to the future. Many travellers are looking for a personalised travel experience as soon as they begin researching for their next trip. A recent Bitkom study showed that a large majority of all users welcomes AI-driven recommendations to their travel planning.
This is more than a user expectation. It’s an opportunity for online travel agencies to provide a booking experience tailored to users’ needs by providing personalised recommendations, content or services. Engaging individually with each user has proven to contribute directly to user satisfaction and to be a main driver of conversion.
Luckily, AI-driven technology has reached a level of development that enables digital travel shops to provide automated personalisation at scale.
The traveller’s buying journey often begins weeks or months before they travel – when they’re inspired to research for a trip. Even at an early stage of the travel booking journey, travel agencies can begin to provide a more personalised experience on their sites. Within the first few clicks on a travel portal, a user already provides valuable clues about his current travel plans and stage in the booking funnel.
A visitor does not need to log in for this – they just need to give their (GDPR compliant) consent to allow their data to be used to personalise and continuously improve the services provided. The data collected is anonymised, but still allows provision of a tailored experience specific to the user’s current needs.
So let us have a look at the key user insights that can be predicted in order to generate individual shopping experiences. Put together, these insights form a predictive user profile – which is where every personalisation use-case starts. Additionally, we will also share some statistics on how accurately those insights can be predicted.
What is Their Budget?
What amount does a traveller want to spend on his next trip? What is his budget range? What offers can you select in his range?
If a traveller has a limited budget and a particular destination in mind – such as a week on the beach in Spain up to 700 EUR – showing “Recommended Hotels” out of his price range isn’t beneficial. He could easily get the impression the products are generally too expensive for him and leave the website to look for cheaper offers elsewhere.
If those “Recommended Hotels” fall within the top or middle of his range, however, he’s more likely to continue searching – and eventually buy on your site.
If a user does not yet have a specific budget in mind, it’s beneficial for your company to showcase diversely priced offers. This provides him the opportunity to find out which price and value level suits his wishes and wallet.
You can also decide to spend more resources on guiding high-value users to a booking. This could be, for example, retargeting him across platforms, prompting a call with a travel agent or simply offering him tailored vouchers.
What are Their Interests?
This is the most important aspect of a user profile. If a visitor to your site is not presented relevant offers related to their interests, they’ll quickly leave in search of a better site. “Interest” should be one of the most detailed categories within your user profiles.
What is the site visitor interested in? Are they just browsing, or do they already have very specific interests?
As a traveller spends more time on a site, their user profile becomes more and more refined. The key is to start tailoring the results as soon as possible, even if they aren’t highly refined yet.
Within a few clicks, it may become clear that a particular user is searching for a trip for themselves, a group or a family. This impacts what will be of interest to them – and it may differentiate significantly from their last booking(s).
As soon as you know the details of a trip (family or solo, beach or mountains …), you can start showing more personalised offers to keep the visitor engaged.
Are They Engaged?
The key to guiding a visitor to a booking is engagement. If he isn’t exploring, reading offer descriptions, etc., he is probably not getting his needs met.
It is therefore essential to know at any time during a session, how likely a site visitor is to take another step on your site? Is his engagement rate increasing as he views more? Is he receiving the optimal engagement options?
If you can observe that the engagement rate is dropping or a visitor is receiving the wrong options, you can start changing what is presented to a visitor to keep them on the site. If he starts searching for beach vacations and later focuses on hiking options, you want to adjust the offers to make sure he always sees relevant destinations. Self-learning algorithms are the method of choice for this task.
In the graph below, you can see how our “Churn Classifier” performs on users who have visited an offer list at least once on an exemplary travel portal. We predict the churn risk for every user at any page view.
You can see how many page views (indicating sustained engagement) users had after we classified them in one of three classes of churn likelihood. The users with the lowest number of subsequent page views are the users for whom we predicted the highest churn risk.
What is the Intent?
How close is the user to booking? Are they just looking to get inspired? Are they searching for availabilities of a specific offer or comparing prices in order to book in this session? If you know the intent of your user, you can determine the most relevant offer or service to present next.
For example, a user is visiting your site looking for inspiration. While you continue to show them offers and travel options that match the search profile, you don’t want to waste resources in trying to convert them at this time. They are at the wrong stage to effectively offer discounts or vouchers. Also showing disruptive pop-ups urging them to contact a call centre to book when they aren’t ready can drive them off your site.
If a user is comparing different offers for a specific trip, they might be closer to booking. If they continuously find offers that are not available anymore or for which prices have increased significantly, the user may get frustrated. At this point, a user has a high churn risk. A pop-up offering the help of an agent at a call center could help keep the user interested and eventually save the sale.
The graph below shows how our “Booking Classifier” performs in practice on another travel portal. Without such a classifier you would need to treat every user the same. But how many more options would you have, if you knew which users are the 13,6% of your audience who just need a little nudge to make them a converted and happy customer?
What to learn from past purchases?
What has this user booked so far? Is he likely to buy again? Is he looking for upsells or does he want to read the offer description again? If a user has booked with a travel company in the past, he already knows what the experience of using that particular company is like. If he returns to the site, he is likely to book again.
The user already has a profile from the previous interactions on the site. This allows us to show personalised offers from the beginning of the search. Even if he is looking for a different travel experience than a previous booking, such as a solo holiday instead of a family holiday, some personal characteristics can be reused to create offer lists that relate to his interests.
This does not mean you should show him just similar offers to the ones he already booked. This so called “filter-bubble” is a pitfall of some AI-driven approaches as it limits the diversity of the offer portfolio taken into account for a particular user. But certain characteristics, like an expressed interest in hiking or weekend-trips, can serve as starting points to create personalised offers for the current trip.
Each of these predictions for an individual user are key elements of the buyer’s journey and can be used to move them further along the booking funnel.
Are you interested in better understanding your users to automatically improve their experience on your website and increase sales? Let us show you how profiles can help you in a personal demo. Drop us an email via the contact form below.