Wired, June 2009 at 110, describes Google’s auction method, one which could apply to legal departments when they put work out to bid for a fixed fee. The winner of a competitive bid would be selected to do the work for one thousand dollars more (or some other pre-determined increment) than the bid of the firm with the next-lowest offer. So if firm A proposes to handle all your immigration work for two years for $300,000 and firm B proposes to handle it for $320,000, you pick firm A and award the work at $301,000. This method reduces the winner’s curse (See my post of Jan. 14, 2007: winner’s curse.) and is more likely to reach a fair price. This technique is often referred to as a “second-price auction.”
I have criticized auctions for legal work, yet I strongly back competitive bidding processes and the second-price variant. To reconcile my positions, I have to make clear that what I do not like are electronic, online, real-time auctions for legal services of any complexity (See my post of May 21, 2007: auctions with 7 references.). I favor market discipline and information for fixed fee arrangements.
Other kinds of auctions exist, such as Vickrey-Clark-Groves (See my post of Feb. 20, 2007 #1: expressive bidding algorithms for bundles of services; Feb. 1, 2006: “double auctions” that match seller and buyer rankings as well as “unique bid” auctions; and May 23, 2008 #1: combinatorial algorithm auctions.).
Since that metapost in mid-2007, this blog has dropped the gavel on several more auction items (See my post of Jan. 14, 2008: competitive bids and market cost; Feb. 16, 2008: few references to auctions proves their unpopularity; May 2, 2008: Esq.-harmony to match law firms and law departments; Dec. 9, 2008 #4: online auction by Intel for patents in 2006;and April 9, 2009: eLaw Forum and a massive competitive bid.).