Make Tonight Unforgettable with HotelTonight

In need of a last-minute hotel room for upcoming travel? How HotelTonight is using machine learning to disrupt the online travel booking industry.

Founded in 2010, HotelTonight (“HT” or the “Company”) has been revolutionizing the online travel industry through its cutting-edge mobile app that offers consumers access to last-minute hotel deals at discounted prices. HT’s value proposition is compelling and addresses multiple trends within the travel and hospitality industry, including:

  • Increase in Spontaneous Travel. Travel (particularly among millennials) is moving towards last-minute and mobile booking. 50% of hotel bookings are made within three days prior to check-in and 43% of accommodations are booked via mobile applications [1,2]
  • Distressed Inventory. According to the Company, ~40% of hotel rooms are unoccupied on any given night, and hotels are not very willing to discount in order to preserve their rates [3].
  • Curated Selection. Hotels have struggled to capture a captive audience with legacy online travel booking providers that offer hundreds of potential lodging options.

Historically, incumbent online travel booking companies (i.e. Expedia, Priceline, TripAdvisor, and offered relatively antiquated travel solutions that did not address these issues. As such, HT’s founder and co-CEO, Sam Shank, an experienced travel industry entrepreneur, launched HotelTonight, a last-minute hotel booking mobile app that connects savvy, on-the-go consumers with a curated list of hotel rooms for all budgets (from Basic to High-Roller). HT works directly with hotels to sell unsold hotel room inventory (which would otherwise go unsold), helping drive increased demand and revenue.

HotelTonight’s mobile app leverages machine learning to power a streamlined and dynamic user interface where consumers can book last-minute hotel rooms at steep discounts. The Company has demonstrated significant growth, expanding rapidly to serve hundreds of destinations across 35+ countries with 15,000+ partner hotels [4].

Is Hotel Tonight’s business model sustainable? Or will the Company inevitably be swallowed up by the online travel booking behemoths?

The online hotel booking industry is hyper-competitive. HotelTonight faces extreme competition from legacy online booking providers as well as several mobile, data-driven start-ups that have emerged in recent years. As a result, HT’s future growth and profitability might be at risk due to shrinking margins as customer acquisition costs rise with the increased competition.

In my opinion, HotelTonight’s data-driven approach, rooted in machine learning has enabled the Company’s growth and should continue to serve as a barrier to entry. Over the last 7+ years, the Company has worked to perfect its ecosystem of algorithms to drive real-time analytics and personalization. In particular, HT utilizes several tools that help increase customer engagement and lifetime value, including:

  • HT Pros: best-in-class mobile concierge service, primarily targeted for high-value users.
  • HT Perks: loyalty program that provides users with discounts based on tier (1-9+); another avenue for hotels to target user segments (based on spend, preferences, etc.). HT understands its users and know which types of hotels they like to book at, so hotels can use this to target customers
  • Geo Rate: HT is able to target customers based on their geography and offer different prices to different customers to drive incremental bookings/revenue for its hotel partners

As a result, I believe that HotelTonight has created a niche within the spontaneous, last-minute hotel booking market and should continue to benefit from secular tailwinds, particularly among millenials. According to Shank, millennials are “fundamentally different as a generation, which extends to how they travel and how they really travel for business. What they really want is value. They’re really interested in spending a little bit more to get a whole lot more versus previous generations who just want to spend the absolute least amount of money as possible.” [5].  Importantly, in September 2017, the Company increased the booking window from 7 days to 100 days. Moreover, HT just began offering a desktop website for consumers to book hotels, in contrast to its historical mobile-only approach. These moves could help the Company offset increased competition and fuel future growth with an increased reach and addressable market size.

Going-forward, HotelTonight needs to focus on driving increased brand awareness and customer loyalty.  With the democratization of data, the travel industry is undergoing a digital transformation as consumers have become smarter than ever about their travel decisions. Importantly, 57% of U.S. travelers feel that brands should tailor their information based on personal preferences or past behaviors [6]. As such, HT will need to continue to leverage data and machine learning to drive personalized, high-quality hotel rooms at great rates, and become further entrenched within its customer base.

However, the crucial question remains: As machine learning becomes more and more ubiquitous, will HotelTonight be able to maintain its competitive edge? Are the Company’s existing barriers to entry strong enough to drive sustainable growth in the future? 

(781 words)


  1. “Digital Marketing Trends For Travel And Hospitality: Engaging An Audience On The Run | Adobe Blog”. 2018. Adobe Blog.
  2. “From Search Engine To Booking Engine: Sojern’S 2017 Hotel Report”. 2018. Sojern.Com.
  3. 2018. Usatoday.Com.
  4. “Hoteltonight Essentially Pivots To Become Hotel Whenever”. 2018. Skift.
  5. Castillo, Michelle. 2018. “When Millennials Get A Hotel Room, How It Looks On Social Media Matters More Than The Price”. CNBC.
  6. “2018 Travel Trends To Help Shape Your Marketing – Think With Google”. 2018. Think With Google.



Customization vs. Scale – How BMW Imagines the Future of Cars


Man or Machine? Does AI have a place in Venture Capital?

Student comments on Make Tonight Unforgettable with HotelTonight

  1. I do not think that Hotel Tonight will be able to maintain a competitive advantage. Machine learning is now part of almost every Fortune 500s 5 year plan. Expedia, AirBnB etc. are all using machine learning today and Hotel Tonight, as far as I can tell, is only marginally ahead of those competitors if at all. Hotel Tonight’s saving grace, in my opinion, will be if they’re acquired by a larger competitor for their data science talent or platform, which is becoming a common way to acquire talent in a space with so few qualified workers.

  2. I think HotelTonight will be able to maintain a competitive advantage if it can reach the right customers and be top of mind at the time of hotel booking. While I am impressed by their use of ML and AI to offer services/special rates once a user is already in the app, I think they need to use their consumer insights to attract and maintain customers. I wonder if HT could leverage the data it has on customer retention/churn to create a targeted marketing campaign to build a base of loyal users such that its service becomes indispensable.

  3. HotelTonight’s competitors advantage, as you point out, is their user experience. I’m less convinced that machine learning in and of itself is their competitive advantage since hotel inventory access is commoditized. Also as HotelTonight becomes less niche by expanding into wider booking windows and a desktop experience, it become less differentiated than the other larger hotel engines. Which begs the question, what sets it apart. Customer loyalty feels hard to achieve in a price sensitive market especially since the Brand isn’t connected with the end hotel chain itself that a user may feel loyal to.

  4. Machine learning capabilities are being commoditized rapidly. Unless HotelTonight finds a way to lead technological development in the field, I strongly believe that their competitive advantage will erode. They’ve done a remarkable job growing their user base by exploiting an overlooked problem, but what stops the other travel booking giants from copying HT’s features overnight? Either they make a compelling customer retention play, or they extend their technological advantage. Otherwise they should sell the company to one of the larger competitors.

  5. As Hotel Tonight continues to expand and acquire more customers, it will need to add more hotels/rooms as well. In that process, Hotel Tonight is going to become very comparable to other websites such as Expedia, Orbitz, etc. Also, customers that have spontaneous trips tend to book for shorter vacations (long weekend vs. 2 week vacation) and therefore generate less revenue. HT needs to find a way to appeal to frequent and established travelers. Therefore, it won’t be able to keep up its competitive advantage. The value proposition of Hotel Tonight currently works because it can afford to be niche and exclusive; however, to scale HT will need to add other services or follow the trend of established companies.

  6. Interesting essay! I believe Hotel Tonight can maintain a competitive advantage. I am a frequent user, and I appreciate how easy it is to use and how it is geared toward last-minute travelers. Even though the other services are useful and may have similar machine learning chops, they are not solving this specific problem as intently. Over time, they have branched out into more traditional products as well, such as status discounts.

  7. Thanks for sharing your view on Hotel Tonight and how the company plays in the machine learning area. In my opinion, the competitive advantage for Hotel Tonight is actually its value and customer experience/ease of use. I was actually quite surprised to learn that they extended their booking window from 7 days to 100 days, as I think that decreases the value proposition from the perspective of a customer. For example, when I log onto the app and book last minute, I feel like I’m getting a great deal because the hotel wants to increase utilization and, because of that, I have access to a room at unique price point. By increasing the booking window to 100 days, I think it transforms Hotel Tonight into more of a traditional online booking platform. With regards to machine learning, I think this provides the company with an advantage in customization and targeted advertising/pricing, but it is not enough to differentiate the company from its competitors. In addition, I think there is a fine line to walk in the hospitality industry with machine learning and customer preferences to ensure that people feel they are getting what they want, but are also able to maintain their living space as a safe and private space.

  8. I have used HotelTonight a few times so it was very interesting to read about how they are using machine learning. If I were them, I would also be worried about other competitive threats like Airbnb or even credit card reward programs with hotel partnerships which are dramatically impacting the hotel category. Also, the large players like Expedia are getting smarter and smarter with their data, so HotelTonight will need to continue innovating and advancing quickly in order to stay top of mind with consumers.

  9. Thanks for a fascinating article and overview of hotel tonight! I have to agree with Noah here that I’m concerned about the long-term durability of the hotels tonight model. It reminds me a lot of Gilt Group – the members only discount fashion site. One piece of data I want to see is how the hotel tonight prices actually compare with those of Expedia or My guess would be that most hotels at this point use dynamic pricing – e.g. it will be cheaper to book that day – not just through hotels tonight, but all sites. I feel like moving from mobile-only to include a desktop compliment just shows how undifferentiated they’ve had to become. What can they do differently from their competitors? What can they provide their suppliers (hotels) that others cannot?

  10. Thanks for the sharing! However, similar to the comments above, I’m more skeptical about whether Hotel Tonight has really built up the entry barriers and competitive advantages with machine learning. Hotel booking has been a data-driven industry with floating pricing across various platforms, and demand forecast has always been a key part of most of the booking platfroms. So, for me, what makes Hotel Tonight different should be that they’re adding “location” data as an input to their model. However, this competitive advantage doesn’t play that much a role once they shift to website and to support a longer period for “pre-booking”. Therefore, I’m quite skeptical about whether they are able to be competitive on the laptop war.

  11. Hotel Tonight has helped me in a many hours of need so a big fan! But I do agree that the company needs to be vigilant in finding new ways to expand its competitive advantage. As mentioned by several learned scholars in earlier comments, everyone in the hotel/accommodation/travel industry is heavily gearing toward machine-learning driven digital strategy. Two of my favorites – SPG and Airbnb – have apps that do HT’s core functionality as good or better than it does. Perhaps HT’s real edge will be in the non-big brand name boutiques around the world, or in innovating a more professionalized alternative to Airbnb (e.g., for executive/serviced apartments).

Leave a comment