E-commerce hotel chatbots: what hoteliers want to know

This article answers the top 2 questions hoteliers always ask about e-commerce hotel chatbots :

1- What do customers ask your chatbot?

2- How to measure your chatbot ROI?

What do customers really ask a hotel chatbot?

Hotel chatbots are usually available on the hotel website, Facebook Messenger, and sometimes on WhatsApp or the hotel app (if there is one).

Most interactions happen before the stay

In 2018:

— 85% of interactions happened even prior to the reservation.
— 98% of interactions happened before check-in.

Chatbots are mainly used by customers that are in the discovery phase and are still hesitant to book your property. Despite being available during the stay, at this stage, the real potential of hotel chatbots is pre-stay and the focus is clearly on sales.

Most customers’ questions are basic and repetitive

While doing demos I have seen hoteliers ask: “Tell me a joke”, “Can I book a tuk tuk”, “Do you have umbrellas in the rooms”, “Do you have Cambodian cuisine”. I have never seen real customers ask our hotel chatbot these questions. The only specific restaurant requirement that we observed was vegan food.

On your website, people are not looking to try the limits of your bot or to have fun. They have a real question that is preventing them from booking a room. Subsequently, they want an answer. Most of the time, the information is available somewhere on your hotel website, but customers just do not make an effort to search anymore. They expect you to deliver the information instantly upon request.

Reassurance is also a key driver for people when they interact with a hotel chatbot  

People come with their own contexts and fears. They do not know your hotel and are going to spend money, so they want to get it right.

— Customers can be fully aware that the check-in is from 2 PM but still ask about check-in time because they plan to arrive very late at night and are afraid that the front desk will be closed.
— You can display beautiful pictures of your swimming pool but people are still going to ask if it is heated, if kids can access the swimming pool etc. In fact, the more you show your swimming pool (or any feature) on your website, the more you will see questions about it in the chatbot conversations.

While there are variations among hotels, Jupiter Lisboa Hotel in Lisbon is very representative of what we see in the hotels that are part of our data set. This hotel has a restaurant, spa, swimming pool and parking. For boutique hotels with no restaurant, people still ask the same questions and expect an answer about restaurants around the hotel. So, people have basic needs and will always need to have them covered. Hence, your e-commerce hotel chatbot should be ready to answer.

Our demo bot simulates a boutique hotel in Paris. Feel welcome to try it here: https://demo.quicktext.im/zoe1/

How to measure your chatbot ROI?

The success rate is the volume of conversations where the user got the information he wanted or could successfully process his request to book, cancel, modify his reservation or more.

The objectives of e-commerce hotel chatbots are to

1. Filter recurring questions to avoid unnecessary calls and emails.
2. Increase conversion.

1- Filter simple questions to avoid unnecessary calls and emails

Right after a city hotel implements our chatbot the results usually look like this example:

• 76% of conversations are handled successfully by the chatbot.
• Some conversations need to be optimized. Here you need to have a clear understanding that 50% of chatbot mistakes are not related to AI but to incomplete predefined answers provided by you, over which you have total control. By taking the time to optimize your chatbot’s predefined answers you can increase the rate of successful conversations from 76% to 88%. (More details here.)
• The 12% of misunderstandings should be dealt with by your AI provider through a machine learning process.

2- Increase conversion

Booking engines are a fine way to book for many customers. In other cases, when a demand is a bit out of the usual “2 persons, 1 room” (flexible dates, group requests…), people tend to have difficulties using your booking engine. Chatbots, however, will know what to answer in order to reassure and capture the lead.

Our hotel chatbot is able to take standard reservations as well as group requests and reservations with flexible dates (example: “1 weekend in June”). This makes it a lot easier for customers to enter the conversion funnel. Of course, the chatbot can’t send this lead to your booking engine as it won’t be able to handle non-standard requests. So, a follow-up will be necessary. If you do, the financial reward will be quite handsome. Absalon Hotel is an excellent example of the ROI you can expect.

On average, the chatbot of Absalon Hotel in Copenhagen handles 184 conversations per month and gets 64 room reservation requests. Of these, 30% are converted, leading to 99 360€ of extra sales per annum.

The follow-up process is easy. Furthermore, some automation tools such as Hotel ResBot can make it really simple to follow-up by email on incoming requests. Hotel ResBot allows you to create email templates. The AI recognizes if the request is a reservation/modification/cancellation, pre-writes the e-mail, inserts the available offers and rates and your team just has to verify it and press send.

If you want to learn more about e-commerce hotel chatbots and how they can help you boost your direct sales, please visit our website or feel welcome to contact Benjamin Devisme at bde@quicktext.im.