Chatbots – automated bots that let you interact with a service or brand via messaging apps or SMS – haven’t yet become breakout hits, as a group. But that hasn’t stopped investors from pouring in $8 million into a hotel booking startup called SnapTravel, which lets users find and book rooms via SMS texts and Facebook Messenger.
The $8 million Series A round was led by iNovia Capital, a backer of Luxury Retreats which just exited to Airbnb for around $200 million. The round also included participation from seed round investors Lightbank, Bee Partners, and Hedgewood, along with strategic investor and Expedia board member Peter Kern, among others.
iNovia’s Chris Arsenault is joining SnapTravel’s board with the close of the round.
The startup itself was founded in spring 2016 by tech entrepreneurs Hussein Fazal, whose prior company AdParlor grew to $100+ million in revenue, then sold to AdKnowledge back in 2011; and Henry Shi, who previously built uMentioned and worked at Google, where he helped launch YouTube Music Insights.
Unlike a number of chatbots on the market today, SnapTravel isn’t a fully automated solution – arguably, to its advantage. The “bot” part of the service aids with onboarding and can answer users’ simpler questions, but when things get tricky, human agents are able to step in. For example, if you booked a non-refundable hotel but your flight gets cancelled due to weather, an agent would respond to your questions and try to help, rather than leaving you to deal with an unsympathetic bot.
You can also ask the agents for help with questions bots can’t handle – like if the hotel is pet-friendly, or if you can get adjoining rooms, among other things.
10 – Request talking to human agent
1 – Get Started
11 – conversation special requests
11 – conversation – special requests
3 – hotel in mind
2 – city date question
5 – searching for you
4 – budget question
7 – view all deals
6 – presenting deals
8 – selecting room
9 – map view
“Customers want a travel agent experience, but they didn’t want to call; they didn’t want to walk into a store,” explains Shi. “They just wanted to email us or message us,” he says, of when the company began testing ideas. That led them to realize that delivering the hotel booking experience could work over chat.
SnapTravel’s human agents effectively operate as modern-day travel agents. They’ll call hotels for you, as needed, and even phone to confirm your check-in and attempt to negotiate a free upgrade on your behalf.
Shi says that hotels honor SnapTravel agents’ special requests around 70 to 80 percent of the time, and they are able to upgrade customers to better rooms around 25 percent of the time.
Though still an early stage startup, SnapTravel has made a bit of a name for itself due to its deals and its “app-less” business model.
The company filters through thousands of hotels from sources, then leverages machine learning algorithms to narrow those down to the top options, based on factors like price, location, quality and overall value.
As customers continue to use SnapTravel to make their bookings, the service learns more about individual preferences – like if you want free Wi-Fi or free breakfast, for instance – and then takes those into consideration for its future recommendations.
SnapTravel doesn’t necessarily have any deals you can’t find elsewhere, but it can feel like it, at times. As its own website explains, it has access to the same unsold hotel inventory that others do. The service’s hotel inventory comes both directly from hotels themselves, as well as via partners like Expedia, Booking.com, Amadeus, Hotels.com, and Sabre (pending).
But while agreements between hotels and some hotel booking services disallow them from posting their super-low rates publicly, SnapTravel is able to work around that restriction because it’s only displaying prices through private, 1-on-1 chats – not publicly on the web.
But SnapTravel is not always the place for the lowest priced deal or the most options, some reviewers have found.
However, lower cost is not SnapTravel’s only draw, Shi points out.
“We’re 100 percent free to the customer…the way we make money is the commission on the hotels,” he says. “But we’re building a relationship with the customer because we’re chatting with them over a channel that’s very intimate, like SMS or Messenger. For us, it’s a conversation – and that conversation doesn’t stop when you make a purchase. It goes from search to purchase to upgrade to even post-checkout,” Shi explains.
At Facebook’s f8 developer conference earlier this year, SnapTravel announced that it had already achieved over $1 million in revenue on Facebook Messenger. It has since surpassed that, and launched its service on SMS and chat app Viber.
It eventually wants to reach users worldwide through other platforms, as well, like WeChat and Line.
SnapTravel’s app-less biz model
The company believes that chatbots hold potential – especially in the case of Facebook Messenger, in terms of reach. (Messenger topped 1.2 billion monthly users worldwide in April, for example.) SnapTravel doesn’t need to get anyone to download its app in order to work.
The team appreciates that SnapTravel can advertise to users in-app through things like Sponsored Messages, and retarget them directly in their messaging inbox when they fail to complete a transaction. In that latter case, SnapTravel says it has seen a 30 percent jump in conversions when retargeting in this way on Messenger.
“Being able to re-engage customers is critical for any commerce company. Typically, with an [online travel agency] when you come to the website and bounce, they have to follow you around the internet using display ads,” notes Shi. “That’s a very noisy and ineffective channel,” he adds.
“For us, being able to use Sponsored Messages on Facebook has been super effective. We’re seeing double-digit lift from these re-engagement efforts,” Shi says.
With the additional funding, SnapTravel says it will invest further in its Natural Language Processing (NLP) capabilities in order to reduce the reliance on human agents as its scales. In this envisioned future, that would mean an automated bot would be able to understand a conversational input from a user, like “I’m going to New York next week Wednesday for a night, and need a room at the usual spot,” for example.
Without needing a human to translate that into a booking request, SnapTravel would then send the customer to the booking form for the right hotel and take their payment through Messenger.
Currently, SnapTravel is a dozen or so product engineers based in Toronto, in addition to an operations company stateside. With the new funds, the startup plans to expand to a team of 20 to 25.