How HotelTonight killed its frenetic vibe and figured out how to scale

By Carmel DeAmicis , written on July 15, 2013

From The News Desk

This is the first story in a series I'm writing about the people doing the grunt work in the trenches of the startup world. Not the famous founders or the rock star six figure coders. The rest of us.

There are plenty of worker bees toiling all over Silicon Valley who don't get any of the fame or the glory but keep the innovation machine humming. Behind every great innovator is 1, 10, 100 (or if you're Sergey Brin, 38,000) hustlers, working to make that dream happen. Depending on which employees I wind up tagging along with, I might be cooking up free company meals for engineers, learning Excel over the shoulders of data entry guys, maybe even running a TaskRabbit job or two. As someone new to the Valley, it's a good way to learn the ecosystem from the roots up.

Last Friday, the goal was to witness the madness at HotelTonight, which lets users book accommodations through mobile devices, listening in on calls as the startup's staff scrambled to cut discounts with hotel managers before the app's "featured" deals went live at noon. The process was described by HotelTonight's CEO Sam Shank to Sarah Lacy as not unlike a bustling old fashioned newsroom with a small army of resourceful employees manning the phones to make sure cities across the US and Europe all had hotel rooms ready to go a noon everyday for that night.

I was excited to watch it all unfold. In my mind, I upped the ante, imagining HotelTonight  like a old school Wall Street trading floor. "Buy!! Sell!!! No Buy!!!!"

And….I was wrong. As it turns out, a year ago HotelTonight invested in technology that transformed the way the company worked and opened the door for it scaling internationally.

The startup no longer operates like a newsroom - it's a little closer to being a Zen Buddhist monastery. Unlike their original, crowded office where employees worked on top of each other, the company now has a giant, open workspace. I suspected if I stood at one end and shouted "hello!" my voice would echo back at me "ello…ello….ello…." The noon deadline was ticking down, but phones weren't ringing off the hook and conversations were being held in mild-mannered near whispers. The air didn't smell like stressful, sweaty employees - it smelled like people who had gotten eight hours of sleep and ate fruit for breakfast.

Who can we thank for my less than thrilling first day assigned to this series? The almighty Silicon Valley algorithm.

Back in the early days, market managers did all the grunt work of finding hotels to partner with and haggling over pricing. "You'd practically have a phone in each hand and be working your emails with your toes to make sure we had inventory for launch," co-founder Jared Simon said. "12 pm rolls around and you breath an enormous sigh of relief."

But in July 2012 everything changed - HotelTonight developed an algorithm that automated the system and lessened workplace pressure dramatically. Hotels now input their own prices and the algorithm chooses the best "deals," based on price range per type of hotel (luxury, mid range, etc). The startup has established enough hotel partnerships in  each market and gotten popular enough among the general public that hotels now jockey to be the chosen "featured" ones. They see it as their best chance at getting their vacant rooms booked the same day.

There is still a role for the humans, but it's one that you could handle while sipping tea, based on what I saw. I would be lying if I said I wasn't disappointed. I was meant to be shadowing a HotelTonight manager in their fast-paced, stressful negotiations. Instead, I followed Sam Selby who has the emotional temperament of a meditating turtle - that is to say, nothing much rattles him. Selby is the sort of fellow who looks like he'd get on well with anyone. He's attractive in a Macy's catalog sort of way, quick to smile but thoughtful in a gentle, media-coached way.

First up on Selby's agenda for the day was calling back a hotel manager in LA that Selby speaks to often. The manager runs a limited service hotel, so he likes to make sure he's undercutting everyone else since he doesn't have as much to offer. Their conversation was short, sweet, and casual. Selby took a look at the discounts other hotels were offering in the area - via a simple interface that displayed realtime room openings - and suggested the manager cut his price. The two chatted about how bookings this weekend looked surprisingly rough, Selby asked briefly "how's everything else going?," and they got off the phone.


As he ended the non-eventful conversation and hung up, I realized I didn't have much of a story.

Grunt work at HotelTonight – what co-founder Simon calls "putting out the fires" – has all but disappeared. Now, market manager Selby can spend his mornings as a consultant, telling hotel representatives the day's market demand, and his afternoons as a strategist, researching new markets HotelTonight should be tapping into. That sounds much nicer for him, but isn't exactly fodder for a gripping Pando narrative.

I did my best as a reporter to draw Selby out and find out about the man behind the hotel-finding mission. I utterly failed. In fact, he gave me the most upbeat, sanitized answers to every question I asked that I half expected a PR person step out from behind a Selby mask halfway through and shout "surprise!" The most negative thing Selby said about his job (either before or after the new algorithm) was that he misses "the sunrise," which he used to see when he'd wake up at the crack of dawn to get the East Coast deals in the pre-algorithm days. Well-played, sir…well-played indeed.

Simon laughed when I told him I couldn't crack Selby's surface and get the real story. "Yeah, Sam's a low blood pressure type of guy. But did he tell you about the time he shut down Vegas?" In fact, Selby had, and the anecdote is an excellent explainer for how the new algorithm has changed HotelTonight culture and made the startup scalable internationally.

HotelTonight has a feature called high roller, where hotels can display discounted same-day penthouse suites that they didn't fill. Because HotelTonight didn't have an automated system before the algorithm, employees has to enable or disable this feature by hand based on whether such luxury rooms were available.

In the midst of a busy work week, Selby thought he was disabling the featured penthouse room for Vegas, but in fact disabled the entire Las Vegas HotelTonight market….and didn't notice. In other words, he hit the wrong button and for that day all Las Vegas deals disappeared from the HotelTonight app. "You can imagine the stress levels in the office when something like that happens," Simon said. "The only reason he did that was because everything was so manual back then." Selby calls it his "most terrifying moment on the job."

As the Vegas mistake shows, the startup would be struggling to expand internationally if its system was still so vulnerable to human error. HotelTonight needed its market managers to devote a bigger portion of their time to the big-picture business strategy, not the day-to-day details. The algorithm allowed that to happen, and now HotelTonight has an office in London covering the European market, and 39 international cities on its app.

I stayed at the startup until 6 pm, and as the afternoon ticked on the office got a little louder. Someone started riding a scooter through the kitchen, a game of toss the beanbag began, and a group of 25 or so employees wheeled suitcases out the door for a weekend trip to Vegas together. Before the algorithm, they used to spend weekend mornings on the phone, finding featured room deals. Co-founder Simon had an overarching sense of guilt that he was ruining these guys' lives.

But now? He told me, "I'm happy you're bored, frankly. That means we can launch another hundred cities."