Olset uses Big Data to pick the three best hotels for youNewsBy Sean O'Neil | September 16, 2013Share This article was originally published on Debuting in open beta today, Olset is a virtual agent that detects when a traveler may have an upcoming trip and autosuggests three hotels based on either their preferences or their Expedia, Priceline, and Orbitz booking history.Once the site learns a traveler’s preferences or booking history, it matches those criteria against its database of hotel attributes. Then it suggests three hotels that may be a good match.Olset (pronounced "all set") bases its recommendations on a Big-Data analysis of what's posted online about hotels. It is constantly categorizing available structured and unstructured information about hotels, such as on TripAdvisor.This analysis allows it to personalize its suggestions, the company claims. For instance, if a user says it's important for him or her to have a fitness center, Olset won't suggest a hotel that has gotten negative reviews for its fitness center.Unique propositionMany personality- and preference-based hotel-matching services have been tried by other companies, such as Hotels.co.uk, in the past, without much success.Olset claims to do something different in a few ways: It expects most of its business to come from the unmanaged business travel market, rather than primarily from consumers.It also has partnerships with virtual personal agents and OTAs, which may be more effective in attracting customers than relying on marketing."The magic is in the on-boarding process"A user signs up by e-mail or through Facebook Connect. He or she can then manually set their preferences for hotel amenities, ratings, etc.Optionally, users can also enter their account information for their Expedia, Priceline, and/or Orbitz accounts. Olset uses a customer's credentials to log in on their behalf, suck down their data, and analyze their past hotel stays -- to determine the types of properties they prefer.After it learns a user's preferences, it is ready for you to request hotel options. For any destination and travel dates, it can deliver three relevant hotel options ("closest", "best fit", and "cheapest").Each of the three hotels comes with a one-click booking option that deep-links to the partner online travel agency's (OTA's) relevant transaction webpage.Integration with digital calendarsOlset also works with Outlook, Google Calendar, iOS's Calendar, and Android's calendars.For instance, if you add an entry on your digital calendar saying that you'll be in a particular town for a meeting, Olset will e-mail you with three suggestions for hotels to stay at based on your preferences.Olset recognizes the calendar entry as a potential trip, and writing the calendar note triggers its action, as long as you add it as attending the event by including its email address as an invite.Your trip description can be vague, such as "downtown dallas tx", or it can be specific, such as a street address.If you send a meeting invite from your calendar you will receive the email with three options.Promising partnershipsOlset is using Expedia Affiliate Network (EAN) for hotel inventory, which gives it 140,000 bookable hotel rooms worldwide.It has signed a partnership with GetGoing to help that opaque-booking tool improve its hotel conversions. GetGoing will let its customers use the Olset virtual agent right on its site to save time and optimize their booking process. (Tnooz has profiled GetGoing before.)Olset launches its combined solution with task-management app Any.Do. The partnership will enable Any.Do users to automatically book hotels when they put a task on their Do list. (Last week, productivity blog Lifehacker picked Any.Do as its favorite "to-do app".)Future potential partnershipsNext year, Olset plans to add to its service direct booking links for major hotel chains, such as Hilton, Marriott, and IHG, to cater to users who prefer book directly because they want to either earn or redeem loyalty points.It's also interested in adding opaque booking options, private rentals (Airbnb, HomeAway), and extended-stay lodging options.Street cred as a "Big Data" start-upIt's hip to call yourself a Big Data start-up, but Olset has more credibility to do so than most. Olset wowed crowds at the big data-focused O'Reilly Strata confab earlier in 2013, and it was named "one of the 10 coolest big data startups of 2013" by IT trade publication CRN.Olset's co-founder and CEO Gadi Bashvitz, has 15 years of business development, sales, product management, and marketing experience, including a $250M IPO at Verint and a $200M exit at Merced where he was a vice president and, separately, closed a $15M deal.CTO Amyn Sheriff was the lead architect for a few of TellMe’s products and stayed with Microsoft after its $800M acquisition. Angela Potter is the UX Design and development person.Kunal Verma, the lead scientist (working part-time), is an NLP and algorithms expert with a PHD in CS.Today Olset goes into open beta, having been in invitation-only mode since May 2013. The start-up has been bootstrapped and is now starting its seed funding round.Q&A with Gadi Bashvitz, CEO:Why should people or companies use your startup?We save travelers time in the booking process and find them the most relevant hotel room much more quickly.We will know if a hotel has firm beds, is quiet or has a gym with TVs and whether the Wifi is fast or not. We then learn a traveler’s preferences and have the ability to correctly match the two.Market size? Hotel bookings revenue was $271 billion in 2012. OLSET is initially focusing on unmanaged business travelers which represent 16% of the hotel market, or $43 billion and the 3 million users of smart virtual agents.Both segments have frequent travelers and are underserved in the travel industry. Annual booking commissions for these two segments exceed $5 billion.We are starting with hotels and will later expand to other products including rental cars and local activities.Competition?Competition comes from a few directions.1. OTAs which have a stronghold on marketing and SEO with large budgets.2. Large companies like Google trying to get into the travel market but have not done anything revolutionary.3. Emerging OTAs like Rocketmiles and Hipmunk, trying to gain traction by slightly changing the booking experience.Olset is uniquely positioned to disrupt this market due to deep personalization and the proactive booking approach.Revenue model and strategy for profitability?Our business model is straightforward. We generate revenue from hotel bookings. We make an average of $11 for every night booked, or an average of $26 per trip.Unlike many travel sites we are focused on the step in the funnel and the product (hotel bookings) that makes money from Day 1.In the future we will add other revenue channels including rental car bookings (average of $2.5 per trip), partner referral fees and advertising. Olset hopes to earn solid commissions for referrals to OTAs by offering pre-qualified referrals.Olset hopes to expand from hotel bookings to flights, too, eventually.Tell us how and why you founded the company?In the last decade I had the “joy” of flying 2.1 million miles and staying almost 1,000 nights in hotels.Over the past 6 years I ran business development and international operations for Merced Systems (acquired by NICE systems in 2012 for ~$200Mn) where we focused on big data and unstructured data analytics.In doing all this travel it was clear that travel booking and the actual travel experience is still broken and we are not using the latest big data technologies available to improve the booking process and travel experience.I have always had a passion for travel and I have visited more than 100 countries and 49 of the US states.After the acquisition of Merced I decided to combine the 2 passions for travel and big data to come up with a much better way to book travel.What is the strategy for raising awareness and getting customers/users?We have a 4-pronged approach disrupting the current traveler acquisition model. We partner with companies who reach users at the point they initiate hotel booking.1. Virtual agent apps which have 3M users and a 20% CMGR to proactively reach travelers. Any.Do is our first partnership.2. Smart calendar apps.3. OTAs to improve conversion. GetGoing is our first partnership.4. We will reach travelers directly through PR and viral marketing.How did your initial idea evolve? Were there changes/any pivots along the way?We learned quite a bit over the last few months. The three key lessons were:1. Our original process was to collect quite a bit of information about the traveler’s preferences. Even though this required very little work from the traveller it was still hard to have people take action. Thus we changed the process and while you have the option to provide us with a lot of information from the get-go, you can also start without providing any information and OLSET will learn your preferences over time.2. Originally we were thinking of going to market by ourselves. After all this is an awesome idea and why wouldn’t travellers immediately use it? But we received a lot of interest from potential partners and have decided to take multiple approaches to market.3. The ability to work with partners, predict the need for booking and completely circumvent the need to search for bookings is a very powerful proposition.Where do you see yourselves in 3 years time?In three years, we expect to be the de-facto travel booking provider for unmanaged business travellers and users of virtual agent / smart calendar apps.We also expect to expand our partnerships with OTAs and managed business travel providers and have our algorithm and process integrated into their solutions.The key things we expect to prove is that travellers will use virtual travel agents and that both the accuracy and conversion rates of a virtual agent are higher than cases where they are not used. We expect to be able to book hotels, rental cars, local activities and flights.What is wrong with the travel, tourism and hospitality industry that requires another startup to help it out?It takes way too long to complete a booking and you can’t really find the information that is actually important for you. All this information is hidden in unstructured data and it needs to be exposed.Many hotel providers and OTAs talk about big data but they are only using it to target their marketing and not actually better serve travellers.We know if a hotel which claims to have a pool, really has a pool in which a guest can swim or if it is a 9-foot puddle.Tnooz view: Olset has the potential to win accolades at upcoming travel trade conferences, such as PhoCusWright's Travel Innovation Conference in November. The big X factor is if consumers will be happy with the hotels that Olset recommends. Are the hotel suggestions accurate and relevant? Another unknown is if consumers will cotton to the idea of receiving only three hotel choices. On the one hand, there's established behavioral economics theory that shows that reducing consumer's choice can actually increase conversions in travel. On the other hand, consumers may have become trained to see a broad selection of pricing by OTAs and metasearch sites and not trust a limited selection without that context. In a test search for Paris hotels, it claimed the cheapest hotel meeting this reporter's criteria for dates a month in advance was $242 a night. Given that rate, I might want to adjust my criteria to get more affordable results. (I'm a cheapskate.) But I could be reluctant to spin the wheel again to get three more recommendations. The siren song of metasearch or an OTA might call to me. Another issue: Olset's user experience isn't seamless yet. For example, once you complete a booking you receive two confirmation emails: one from Olset and one from Expedia. You can ignore the one from Expedia, but the double records could confuse many users. All that said, the service is one of the most inventive we've seen for consumers since the debut of GetGoing, an opaque booking site that's a Y-Combinator alumnus. It's a good sign that Olset has teamed up with GetGoing and Any.Do virtual assistant services. Partnerships like those could efficiently enable user acquisition. As the start-up grows, we recommend that the foundersShare this quote NB:TLabs Showcase is part of the wider TLabs project from Tnooz.