What made Google's IPO unique

Google's IPO was conducted by a process referred to in the press as a "Dutch auction" or "Dutch IPO" - economists would refer to this as a Uniform Price Auction. It's actually quite simple:

• The seller (Google) announces how many shares they are looking to sell.
• Any buyer can submit bids in terms of both quantity and price they are willing to pay (such as: I want 500 shares at a price of no more than 100$).
• All bids are ranked from highest price to lowest. Shares are sold to the highest price bids up to the bid which clears the market of shares.
• The price buyers actually pay is the price of the bid that exactly cleared the market.

Here's a NYT infographic that explains the process:

Google was not the first or last company to use this mechanism (other American examples include Overstock.com, Interactive Brokers and around 20 others [1] - Wikipedia is wrong here about "unique" or "built by Morgan Stanley and Credit Suisse") but it is very rarely used in US markets. Investment banks usually strongly discourage taking this path.

In a traditional IPO (which is done in >99% of IPOs) an investment bank handles the pricing. The bank represents the seller, and pricing is set in an opaque process involving the bank and its major clients, usually institutional investors. The investment bank usually receives 7-8% of the proceeds in the IPO, and is thus supposed to have incentives to get the seller a high price. However, the buyers are long-term customers of the bank, so the incentives also go the other way.

What ends up happening is that in a traditional IPO the IPO price is significantly lower than the opening price on the first day of trading. The average difference is 10-20%. This means that the seller (including owners, employees and original investors) usually gets a much lower price than it "should" have, at the benefit of the institutional investors. In LinkedIn's recent IPO, the underpricing was on the order of 100%! There's a huge economic literature trying to explain this phenomenon, but no single explanation seems to nail the problem (citation [3] below does a reasonable job of summarizing the different explanations).

On paper, the Dutch auction process is supposed to bring better results for the seller (including employees, owners, original investors in the company) since the bidding process is supposed to target the top demand in the market. In practice, there's mixed evidence regarding its success, with some claiming that it actually yields worse results than regular IPOs.

In Google's case, there was still a jump of 17% between the IPO price to the opening price on the first day of trading, suggesting that the IPO still underpriced. In fact, 82% of the IPOs in 2004 experienced a lower price jump (and 60% of other search-related tech companies IPOing that year) and the average IPO-to-open price jump in 2004 was only 8.6% [4]. It's hard to say exactly what caused this "mispricing", but Google executives were blamed for mishandling communication around the IPO, including lowering the declared target price of the IPO by 20-30% just a few days before the auction. There were also "conspiracy" theories saying that the investment banks discouraged their clients from bidding in the auction to create a negative outcome for Google that will deter future companies from using the auction model.

The debate around LinkedIn's IPO several months ago [2,3] reveals that the main puzzles around IPO pricing and the right choices for companies going public is far from settled.

Not my work, originally and wonderfully answered by Yair Livne on Quora: https://www.quora.com/How-was-Googles-IPO-unique/answer/Yair-Livne?srid=tYFC