✕ BUSINESS INSIDER › Inside Wall Street and Silicon Valley's effort to reinvent the IPO process with Airbnb and DoorDash, and why they still ended up leaving billions of dollars on the table DAKIN CAMPBELL DECEMBER 18, 2020 airbnb cofounders Nathan Blecharczyk, Joe Gebbia, Brian Chesky Airbnb founders, from left, Nathan Blecharczyk, Joe Gebbia, and Brian Chesky on November 17, 2016. Mike Windle/Getty Images This story is available exclusively to Business Insider subscribers. Become an Insider and start reading now. Airbnb and DoorDash last week used a relatively new process for going public that borrowed elements from a traditional IPO and an auction. The process was pioneered by Unity Software in September. Despite their use of the so-called hybrid auction process, both companies saw their stocks surge in the first day of trading. Since then, Airbnb has risen 1.6% while DoorDash has slumped 19%. Business Insider talked to nearly a dozen people involved in one or more of the three IPOs, and found that Airbnb and DoorDash set initial ranges too low, showed an unwillingness to push prices higher, and faced strong retail buying interest. Visit Business Insider's homepage for more stories. Eye-popping initial public offerings have electrified what's a normally quiet season for public debuts — and stoked fresh criticism from those who say the process is broken and leaves too much money on the table for those raising fresh funds. The hand-wringing continued last week when Airbnb and DoorDash sold shares to the public and promptly watched their stock prices surge, ending the first day of trading up 113% and 86%, respectively. In the days that have followed, the stocks have risen 1.6% and dropped 19%, respectively. Pricing an IPO can be a delicate balance, with the company and its bankers looking to find the sweet spot between minimizing first-day gains and not pushing the price so high that the stock tanks that day or in the following weeks or months. There need to be enough investors who've bought in at the IPO willing to sell to create a market, but not so many that the price gets pushed down. The ideal scenario, according to bankers and executives, is a modest increase of between 15 and 30% that rewards IPO investors but doesn't leave a lot of money on the table for companies. But this year, at least 19 IPOs have doubled in price in the first day of trading, leading some market watchers to start to throw the word "bubble" around. Roblox, a company that was lining up an IPO in the final days of 2020, is suddenly hitting the brakes, saying it needs more time to strike the right balance for the company, shareholders, and staff. The stock-price gains for Airbnb and DoorDash were even more surprising, given that they were using a relatively new process designed to avoid price pops. The process, pioneered by the video-game-software company Unity Software when it went public in September and known as a hybrid auction, borrows elements from a traditional IPO and an auction. When Unity's share price rose a relatively restrained 31% on the first day it began trading, the startup community seized on the process as the latest in a string of imperfect but emerging solutions to a problem that has dogged the IPO market. Business Insider spoke with nearly a dozen people with roles in one or more of the deals in this story, most of whom spoke on the condition of anonymity to preserve their industry relationships. Representatives for Airbnb and DoorDash and their underwriters declined to comment for this story. What follows is an examination of the process that Unity pioneered and how the Airbnb and DoorDash IPOs were different. Those differences, including a reluctance to push the IPO price higher, supply-and-demand imbalances, and the influence of retail investors, set the stage for their first-day pops. Unity was the first to try out the hybrid approach Unity Software Chief Financial Officer Kim Jabal. Unity When Unity CEO John Riccitiello, Chief Financial Officer Kim Jabal, and their team set out to design their process, they wanted to accomplish three main objectives: get a better understanding of investor demand by collecting orders in both size and price, play an active hand in allocating the shares, and give employees a chance to sell shares during the first trading day. Jabal, an employee at Google when the company sold shares using a Dutch auction in its 2004 IPO, had perhaps a better understanding of the pros and cons of the IPO process than other startup execs. Lise Buyer, an independent IPO advisor who worked with Jabal at Google and has dozens of IPOs under her belt, had been working with Unity since July 2019. Unity — working with Buyer and bankers at Goldman Sachs, Credit Suisse, and William Blair — set out to design something different than the traditional IPO, where big investors typically commit to buying a certain number of shares at the market price. It often works like this: A fund manager will tell underwriters that it's willing to buy 5 million shares wherever they might be in the price range that the bankers have set. The manager doesn't usually say how the order would change at different price points, or how many fewer shares they'd be willing to buy at higher prices. That means the company and its bankers can be left in the dark in terms of how much demand there is for the stock at prices higher than the range. One way to find out is to raise the range, but Securities and Exchange Commission rules say a company can't increase it by more than 20% without refiling its paperwork. The setup can often lead to overly conservative, or low, IPO prices and contribute to first-day pops, so Unity wanted to do something different. But doing so risked turning off investors by asking them for more information than they were typically open to providing, so the company worked with Goldman to build a confidential portal for investors to input their offers directly. That tool meant investors weren't required to share their orders with salespeople and gave some a sense of security because they knew their orders were being kept confidential and not being used to guide other investors. Only a few bankers at Goldman and a few more at Credit Suisse were permitted to see the entire order book, Jabal said. "We wanted people to put in what they really felt it was worth and what they were actually willing to pay," she said. "And we just felt like that was the only way to get a true view of the actual demand curve of what people were really willing to pay." 'We want to give it to these midsized funds' Class V Group's Lise Buyer. Lise Buyer Once investor interest was collected, Unity turned to the allocation process. Keeping control, as the company wanted, meant taking a traditional auction or direct listing off the table. Both require a binding allocation of shares to whomever is willing to pay the highest price. "This was the difference between us and the traditional auction," she said. "We maintained our discretion over the allocations." Jabal started keeping tabs on investors about 18 months before the company went public. Her team built a spreadsheet that had rows for every investor, complete with notes about recent meetings, their ownership of other tech or gaming stocks, and, perhaps most importantly, how often they sold their investments, a metric known as turnover. That measurement is important for public-market hopefuls looking to establish a stable base of investors. They don't want investors who will quickly move in and out of their shares, which creates price volatility. The spreadsheet gave Jabal and her team more power when it came time to make the final allocation decisions with their bankers, she said. "We had all this data," Jabal said. "We could look at the data and say — in the final hour when the bankers were like, 'We need to give this much to these guys' — we could say, 'No, we want to give it to these midsized funds. Because look at their turnover data; they actually don't sell.'" Unity started by handing out shares on a pro-rata basis by price and size, before moving some of the share count to chosen investors, Jabal said. Adhering to the process meant settling on a higher price than some top asset managers were willing to pay, effectively choosing process fidelity and data over clubby Wall Street relationships or convention. As Jabal soon learned, the process wasn't perfect. Some of the largest investors in Unity's offering have already unloaded the shares, she said. She keeps tabs on the movement of shares through securities filings and other means. Employees should come first Unity's third objective was to give employees a chance to sell shares during the first day of trading. The company considered letting them sell in the IPO, but then realized they would miss out if the company successfully managed a modest first-day gain, Jabal said. The company also didn't want to keep employees from selling if the stock rose over several weeks and then fell to a lower price by the time employees could sell. The decision meant the company was less reliant on banking clients like hedge funds or ultra-high-net-worth people, who often get IPO allocations in return for selling the shares for a quick profit on the first day of trading, for liquidity. "There was maybe like an hour before the stock opened where everyone was convinced we were wrong and there wasn't going to be enough liquidity," Jabal said. "And it was fine." Armed with those features, Jabal and her bankers watched the stock rise 31% on the first day. Since the shares started selling for $52 in the September 17 IPO, the stock had risen 182% through December 16. So how did Airbnb and DoorDash change the Unity template? Airbnb and DoorDash set out to pursue a similar process, though one on which they would each put their own spin. In an interview before the Airbnb and DoorDash IPOs, Unity's Jabal said she had spoken with execs at the two companies, and Roblox, about how she handled the process. "I talked to all those guys," Jabal said. "I'm glad to hear that they're taking a page from our book." To start, both companies pursued a similar technology option as Unity's, with DoorDash using the same one that Goldman used for Unity and Airbnb using something similar built by Morgan Stanley. The process gave them more information to base their decisions on. How they executed the strategy differed in meaningful ways. One was in their willingness to push up the price. At Airbnb, the initial price range was $44 to $50, but as orders came in, it became clear the price should move higher. The company raised its range to $56 to $60, according to an SEC filing. Ultimately, company executives settled on $68 a share, or more than 40% above the midpoint of the initial range. There was some discussion about going higher still, but executives decided against pushing the IPO price into the $70 range, according to the people familiar with the deal. The stock opened at $146. To some, this was simply a sign that the company's underwriters set the initial range too low and anchored expectations, or guided investors to a particular price. 'The good news and the bad news' At DoorDash, bankers initially set the range between $75 and $85 a share before realizing that demand for the stock was strong. The company increased the range to between $85 and $95 a share, according to a filing. The company and its bankers settled on a price of $102, or 28% above the midpoint of the initial range. They did so knowing that they may be setting the stage for a pop, according to two people familiar with the deal. Raising the price would have meant leaving behind a couple key investors who were unwilling to offer higher prices, one of the people said. The second person said the decision to keep a few investors by not raising the valuation didn't play a meaningful role in the ultimate price. "The good news and the bad news is this new platform gives you a real demand curve of who you lose," the first person said. The stock opened at $182. Here are some more reasons why the stocks still popped There are at least three other reasons the deals may have popped. One is a structural pricing dynamic that's common for hot IPOs. Investors know they won't be able to fill their entire position in the IPO and have to go into the secondary market to top it off. They underprice their IPO offer, figuring they'll have to pay higher prices later, which leads to an average purchase price that matches their model. A related concern is that of a supply-demand imbalance that often plays out in attractive IPOs. Startups don't sell a big percentage of their companies these days, constraining supply and creating a frenzy among investors for the shares that are available. Airbnb, for example, sold just 9% of the company, though it did allow employees to sell on the first day. DoorDash sold 10% and will allow employees to sell after its first earnings report. The frenzy can be especially wild when the companies attract a devoted consumer following. Fidelity's retail clients, for example, placed almost 84,000 buy orders for Airbnb on the first day of trading compared with just over 10,000 sell orders, according to a snapshot of Fidelity data captured by CNBC's Leslie Picker. Even so, startups and their bankers don't want to set the price based on retail orders because that can mean playing into hype, or lead to stock declines once retail investors, who typically haven't done the homework to understand the underlying fundamentals, get out and institutional investors find they can't justify such an elevated share price. The prospects for Airbnb and DoorDash are also harder to understand because of the pandemic. DoorDash has seen revenue soar as people dine in restaurants less and turn to delivery services more, while Airbnb's revenue tanked as lockdowns took hold, only to pick up as people turned to more local and longer-term stays to escape cities. Retail demand can be hard to predict, a fact that Airbnb CEO Brian Chesky noted last week. "We were very, very rigorous," Chesky told Axios' Dan Primack. "There is a very big interest from retail, everyday people, who want to buy the stock. You never can quite tell how that's going to play out." Indeed, by this week, the company's stock prices were beginning to come back down. Airbnb was 103% above its IPO price, while DoorDash was up 55%, through December 16. And to be sure, not every recent IPO has skyrocketed — shares in ContextLogic, the parent company of the online retailer Wish, opened trading below its IPO price on Wednesday and tumbled more than 16%. Roblox will have a chance to strike the right balance when it finally sells shares. In a memo sent to staff last Friday, CEO and founder David Baszucki told his "Robloxians" that the company's discussions with advisors would delay the IPO until early next year. One way it may change the process, according to a person familiar with the deliberations, involves ditching a plan to allow employees to sell shares in the IPO. The practice harms employees if a company's stock price pops on the first day, robbing employees of gains they could have enjoyed. "We've seen companies take innovative approaches to creating a more market-based relationship between investors and companies," Baszucki said in the memo. "Based on everything we have learned to date, we feel there is an opportunity to improve our specific process for employees, shareholders and future investors both big and small."