H ow much can I ask forit?” That is the ques-tion. A reasonable buyerwould not offer the high-est possible bid in anauction. A reasonable

seller of professional services will not offer thelowest possible bid, either. In this article I willdiscuss client-provider engagement and pricingpolicy from the point of view of game theory. Thearticle is intended both for clients, who are inter-ested in getting the lowest price possible whileretaining the best service providers, and for ven-dors, who are motivated to get a fair price fortheir services and to resist price pressure.

Pricing strategy has always been a majorissue for market players on both sides of anytransaction. In complex modern marketssuch as the professional services market, thedrive to sell high and buy low explodes into awhole universe of sophisticated questions.Can I win the contract by offering a betterprice? What is my discount floor for this par-ticular project or client? How should I carryout a tender between suppliers in an attempt tofind the lowest and most reliable and highest-quality offer?

The search for the rules of winning goesdeep into sophisticated economic applications ofgame theory. Potential customers sometimesquery prospective suppliers only once, andthen other factors are taken into account.However, in most cases the customer sets upthe first round of the auction to narrow thesearch and select fewer suppliers to workwith, then negotiates further with those initiallypre-selected. A prospective customer typically

accepts offers in several rounds. In each oneof them, he or she communicates the lowestbid to potential suppliers and asks themwhether they want to come up with a betteroffer. Such a mechanism of finding the truevalue of the item is called an auction. The auc-tion mechanism is also de facto invoked to asignificant extent even in the absence of a sin-gle active seller, but with the aggregate sellersof the professional service market. In this situ-ation, the market itself communicates “thecurrent lowest bid.”

Mathematician John Nash, whowas awarded a Nobel Prize in1994, showed in 1950 and1953 that suppliers com-peting for one and thesame item can, in theo-ry, try to negotiate witheach other the limitsof acceptable price.He also highlightedan important detail:an opportunity tomake a last-minutebetter bid is always amore attractive alterna-tive for an independentmarket player than losingthe contract. This drives theprice further and further downto the level that constitutes equi-librium, which is below the sustain-ability level. This is due to an amazing commonconsequence proven to be true for all auc-tions: Presuming that market players on aver-age estimate the value of the item and their

bids correctly, the winning bid produces lowerthan feasible or even negative profit.

Specifically, bidders tend to base their bidson unconditional expected value of the item.These bids represent their own estimates ofvalue. We reasonably assume that all biddersare professional service providers and there-fore their estimates on average are correct.You only win, however, if your estimate hap-pens to be the highest of those competing forthe item. Winning against a number of rivals

following similar bidding strategiestherefore implies that the win-

ner’s estimate is an overesti-mate of the item’s value (or

underestimate of a feasi-ble contract bid) condi-tional on the event ofwinning. This mind-boggling mathemati-cal consequence iscalled the “winner’scurse” (you win, youlose, and you curse).

The winner’s curseresults in less-than-

acceptable or even neg-ative profits for winners

in all auctions. Winner’scurse is an intrinsic proper-

ty of auctions of all types. Itexists regardless of whether the

bidders realize this danger or not. It isa consequence of the problem definition itself,rather than the bidders’ approach, tactics orbehavior. At first glance, this conclusion evokesskepticism. It reveals the counter-intuitive fact

M U L T I L I N G U A L C O M P U T I N G & T E C H N O L O G Y • V O L U M E 1 6 I S S U E 6 59

FUNDAMENTALS OFPRICING STRATEGYCountering the winner’s curse ofpricing low and paying dearly

SERGE GLADKOFF

BUSINESS

This mind-bogglingmathematical

consequence is called the“winner’s curse” (you win, you

lose, and you curse).Winner’s curse is an intrinsic

property of auctions of all types.It exists regardless of whether

the bidders realize thisdanger or not.

that bidders repeatedly commit mistakes intheir own field. Laboratory experiments show,however, that even experienced executivesimmediately fall prey to the winner’s curse inauctions in other fields. Numerous practicaldata show that the winner’s curse is a pervasivephenomenon in the real world, reported inmany industries, from oil and real estate con-tracts to publication and distribution rightsand, of course, professional service contractstoo. The winner’s curse is an inherent conse-quence of adverse selection on the side of who-ever owns the item to be sold, being the resultof the seller’s sole decision rule. It is thereforecompletely out of the buyer’s control. Winner’scurse is like a hurricane or other act of nature— to attempt to change it is useless. You canonly learn how to live when the circumstanceslook like it is coming.

Exactly how significant is the penalty of thewinner’s curse for those unaware of it? Studiesshow that overly aggressive bidding translatesinto negative average profits for inexperiencedbidders and losses for winners in more than50% of auctions. Experienced bidders are lesslikely to see losses, but laboratory experimentsstill show that 41.6% of bidders suffer fromwinner’s curse.

So what is the strategy to follow for thesupplier to work above the sustainabilitylevel on one side and for the customer to geta reliable supplier on the other side? Theanswer to this question is undoubtedlyimportant for all parties. Nobody is interestedin driving the price below sustainability levelbecause in this scenario the supplier will bedeprived of the capability to invest in busi-ness development, technology, quality and

reliability improvement and will cease toexist. Customers, on the other side, will faceconsiderable extra costs and risks.

At a Dublin Localization Summit carried outby The Localization Institute, a representativeof a large, reputable and innovative clientshared an amazing story of how the companytried to implement an auction model in anattempt to lower localization costs. Client engi-neers devised a web application where clientmanagers could place localization contractlots and potential suppliers could place bidsfor these lots. Fortunately, the localizationmanagers who “owned” the projects were leftto make the final decision on whether actualsale of the lot was made, and they closely mon-itored the auction. Both regularly approvedsuppliers and new suppliers were allowed toparticipate. During the first auction round,bidders indeed went quickly down below theusual suppliers’ price level. Regular supplierssoon left the auction, leaving only marginaland unknown providers to participate. Clientmanagers discovered that the last round leftonly companies that were totally obscure andunknown to the project owners to competewith one another. Project owners protestedagainst placing the order with them, and man-agers decided to suspend trading and cancelthe lot. Notably, this decision was condemnedas unfair and inappropriate by all auction par-ticipants, including even those who quit thegame earlier.

To cure this problem, the rules werechanged so that the auction would continueonly if at least one of the regularly approvedsuppliers still participated. In this new setup,however, the regular supplier naturally won

the auction because it offered the same pricebut had an advantage of much higher credi-bility with the client. Former outsiders were“used” in this setting only to draw the regu-lar supplier lower than usual. This clearlydefeated the purpose of the auction, whichwas strikingly demonstrated when the lot was“sold” to one of the regular suppliers at alower-than-usual price. When the auctionwas completed, the winner called the clientand complained that the “excitement of gam-bling” had unwillingly drawn him below fea-sibility level and he was not actually able totake the job on board on these terms. Couldthey please renegotiate?

The problems revealed by this case studyfit into the picture from other fields and ourown experience. The scope of participants isthe first most sensitive issue. If you open anauction for everybody, then how do youensure that the bidders qualify and actuallyare capable of meeting the quality and pro-cess requirements? If you pre-select the bid-ding audience with a l imited scope ofapproved providers, then again you are run-ning into the same problem because all ser-vice providers are actually different; theirprice floors reflect their scale, costs and cap-abilities; and you are not comparing apples toapples. The process repeats itself. Whoeveroffers lower cost most probably offers differ-ent service. Even if we totally disregard indi-vidual differences and assume that whoeverwins among approved providers is capable ofdelivering, the most probable outcome wouldbe approved providers taking bids to thelower level and then supplying lower qualityat this lower price.

Other complications arise, such as globalsuppliers usually being located in many differ-ent time zones, which makes it difficult anduncomfortable for some of them to participatein the auction interactively. Yet even more thanthat — because an auction implies instantreal-time decision-making process in a dan-gerous break-even zone for the company, onlyhigh-level executives can possibly be autho-rized to represent their companies as qualifiedparticipants. This fact brings significant over-head on any company in this bidding contractplacement model.

Several important lessons can be learned bothfrom the game theory and this scenario, as well asfrom other fields. First and foremost is thatalthough auctions are attractive as an idea, theyare very difficult to implement as a workingmodel, especially in the field of professional ser-vices. There are issues of the scope of suppliersallowed to participate; of complex, changing andhard-to-formalize multifactor selection criteria;and of unwanted cooperative behavior.

M U L T I L I N G U A L C O M P U T I N G & T E C H N O L O G Y • V O L U M E 1 6 I S S U E 660

BUSINESS

UNWANTED COOPERATION

Undesirable cooperative behavior hasbeen observed even in commodity auc-

tions and may appear whenever a communityof participants is formed. It is in the nature ofhumans to cooperate and make alliances, andit is in the nature of humans to use all meansto achieve their goals. Unexpected cooperativebehavior has been observed on large com-modity auction marketplaces such as eBay.

On eBay Motors, for example, car dealersinitially used open auctions where buyerscould chat and exchange information on theitem and seller with one another. What hap-pened, however, is that buyers started to de-velop unfair cooperative behavior to exertextra money from dealers on extras and tothreaten and actually carry out defamationattacks — even to blackmail and intimidatesellers. Experienced eBay sellers now generallyuse only closed auctions on high-price-tagitems, protecting themselves from buyers’learning and cooperative behavior which isoften employed to bend the rules of the gamein the buyers’ favor.

Does this mean that auctions should not bean instrument in professional service indus-tries? Yes and no. Competitive sealed-bid auctions

are commonly used in the commercial con-struction industry in the United States. But aconsiderable amount of private sector work,particularly for larger, more unusual jobs, isawarded through “negotiated” contracts ratherthan through auctions. This proves that withcomplex services and large projects buyers doappreciate the complexity of the problem andcarry out a multi-factor analysis taking intoaccount other considerations and complexindirect costs and values.

As mentioned before, the first key differ-ence between services and commodity tradingis that a service transaction is not completeduntil the project is over and services are deliv-ered and accepted, while commodity auctiondeals are done as soon as payment transactionand physical transfer of the purchase arecompleted. This extensive project durationbasically means that the deal (and thereforerenegotiation opportunity) remains de factoopen for quite some time. Another importantdifference is that there is an ongoing relation-ship between client and service provider.These two differences provide at least threemechanisms of escaping the winner’s curse,either by claiming that “arithmetic errors”have been made, by asking for “help” from the

customer or else by submitting “changeorders.” All three mechanisms allow the win-ner to escape winner’s curse when bidding toolow and immediately regretting his or her bid.These mechanisms are well known in the con-struction industry and are well described anddeveloped, so the outcome in the mentionedcase study is not an evidence of model deficiency,but rather a natural correction mechanism.

The relationship existing between buyer andsupplier in the professional services industry isnot a marketing pitch. It truly represents a quitetangible economic value. This value, specific to aparticular client, may vary and is difficult to esti-mate and quantify, but it can be evaluated. Onecan distinguish three channels where dollarvalue emerges.

First, the customer wants a professional ser-vice supplier to be experienced and have a trackrecord with the customer’s industry or companybecause it saves time and money on training andendgame quality improvement.

Second, the customer wants the supplier tobe reliable and to deliver on time and with goodquality so as to minimize the risk of rework,non-delivery and delays.

Third, it is also comfortable for a client inprofessional services to trust the supplier, and

M U L T I L I N G U A L C O M P U T I N G & T E C H N O L O G Y • V O L U M E 1 6 I S S U E 6 61

BUSINESS

TM MARKETPLACEA CONCEPT WHOSE TIME HAS COME

OFFERING UNIQUE OPPORTUNITIES

FOR TM BUYERS AND SELLERS

208 265 9465US TOLL FREE 888 533 7886

INFO@TMMARKETPLACE.COMWWW.TMMARKETPLACE.COM

THE TRANSLATIONMEMORY BROKERS

uncomfortable to wait to see whether the projectwill succeed or fail.

It would be rather shortsighted for the cus-tomer to believe that the winner’s curse is amathematical incarnation of the fact that interestsof the buyer and seller of professional servicesare not the same and that the buyer is not inter-ested at all in any excessive indirect investment insupplier’s business. Such a statement from thecustomer side would mean that the customerdoes not understand the economic value of therelationship between supplier and contractorand therefore is not fully realizing all the risksand costs that are invoked when selecting thesupplier on price only.

POINTS FOR CUSTOMERSAND FOR SUPPLIERS

The conclusions we draw from all this con-cern both buyers and sellers. The cus-

tomers should take into account several points.First, correct implementation of the auction

contract placement mechanism requires signifi-cant research, extensive field experimenting andnumerous safety checks and adjustment mecha-nisms. If developed correctly, this is going to be avery expensive “toy” which requires careful fine-tuning. If taken lightly, the auction idea willimmediately discredit itself due to a number ofindustry-specific and human factors which willinvalidate the results. In the absence of a fullydeveloped and qualified engine, it’s best to stopthe auction at some point where providers areshort-listed and place the contract manually,weighting all the factors.

Second, if you do not have yet well devel-oped software or even the process to take inand evaluate a competitive bid, take intoaccount the following “rule of thumb”: the low-est price offer is definitely not the best one inservice industries. Globalization went farenough to destroy the tenfold price difference indifferent regions, and now professional servicesare roughly of the same hourly price in Bombayand Hong-Kong, Moscow and Jakarta. You canbuy multilingual desktop publishing services atanything between $3 and $10 per page, forexample, but the lower you accept below $3 perpage, the more chance that you are running intorisks that you do not realize. A lower price oftenmeans that the supplier overestimated his orher capabilities, is probably working below sus-tainability level and therefore has no reliabilityreserve or guarantee. This means that the qual-ity will be lower than you expect and the workwill take longer than planned. There is a riskthat the work will not be delivered at all. It isnot in the nature of the average professionalservices supplier to admit being unable to fulfillthe contract with due quality in this timeframeand within this particular budget. In most cases,

the supplier will simply deliver at his or herbest, leaving you to deal with this “deliverable.”It is then your own fault alone that you haveunderestimated the true cost of the requiredquality and actually ordered something less thanyou expected.

Third, in any case, the price should not bethe only criterion for selecting the supplier. Youmust remember that different suppliers actuallyprovide very different services. They differ in allaspects, from communication (or lack thereof)to experience, knowledge and scalability, toquality of deliverables and many other aspectsthat are difficult to quantify. This simply meansthat a more expensive, overqualified suppliercan prove to be actually more profitable towork with because he or she is more experi-enced, reacts better to unexpected situations oroffers some unexpected benefit that you maysuddenly need. Extra costs are involved inworking with cheaper suppliers — the cost ofadditional monitoring, the risk of rework costand the cost of product release delays.

Fourth, by selecting an offer above the lowestyou are not overpaying. You actually increase thereliability of project completion, decrease yourindirect costs and manage the risks.

Last, you have to know your supplier well.Invest into research of the supplier’s actualcapabilities. If you like the supplier but thinkthat the price is a little bit high, work with himor her further to negotiate the price to the levelthat is acceptable for you.

Suppliers, on their side, have to rememberthe following points.

First, generally, the answer to overcomingthe winner’s curse is to determine the value ofthe prospective contract conditional on theevent of winning, rather than to try to deter-mine its value unconditionally. This means thatthe experienced supplier avoids winner’scurse, watching vigilantly for the bid pricebeing above sustainability level as internallydefined to be acceptable for his or her busi-ness. This slightly decreases the probability ofwinning, but in return offers reliable protec-tion against financial losses due to the ill-setprices. This experienced supplier knows thathe or she is not willing to win the contract if itmeans working below sustainability level.

Second, experienced bidders learn a set ofsituation-specific rules of thumb that help avoidoverbidding in the real field setup. As notedbefore, however, these rules are situation- andindustry-specific to the extent that experiencedbidders in one field fall prey to winner’s cursein auctions in other fields. This is well demon-strated in laboratory experiments.

Third, by all means try to avoid the “bare,”commodity-like auction where the price is theonly criteria. A company with a sensible pricing

policy has little chance of winning on priceonly, and trying to win by price dumping is aself-destructive practice. You may well findyourself going further and further down only todiscover suddenly that you have descendedbelow sustainability level before you knew howyou did it. I would even say that the client whoregards that the price is the only criterion isnot worth working with and fighting for.

Fourth, it is extremely important to acknow-ledge your costs correctly and under no cir-cumstances accept contracts that are belowprofitability level. Whatever the customer says,there is no guarantee whatsoever that the con-tract will be extended, and therefore theattempt to “buy in” to a client relationship byoffering the price below sustainability is anextremely risky investment. And it is next toimpossible to increase rates and return to anacceptable level. Even if you manage to pre-serve the relationship, chances are that youwill have to work with that client at the samelower prices.

Last, whenever possible, try to explain to theprospective customer how your services aredifferent from the others. A truly strong andlasting relationship between customer andprovider and a stable equilibrium market isbased on those unique features that makeproviders different from one another. It is thevariety of services that helps to create a stableand mutually profitable equilibrium of partiespursuing different interests of their own. It isof critical importance to position yourself inthe market correctly, actively demonstrating tocustomers and competitors the unique fea-tures that you have and that actually make itlegitimate for you to charge what you chargeand continue to be engaged in your field in thesituation of fierce competition.

We cannot possibly cover all possible set-ups in all their details, as well as their modelsand numerous details and conclusions. Wehave only attempted to come up with very gen-eral principles applicable to noncooperativesetups in professional services markets. Ibelieve that this field will develop further andthat new software will be created soon todemystify and automate the multifactor deci-sion-making process of proper contract plac-ing for complex professional services, takinginto account specific details. It is time to pre-pare yourself, regardless of which side you

may find yourself on inan auction.

Serge Gladkoff ispresident of LogrusInternational. He canbe reached at sgladkoff@logrus.ru

M U L T I L I N G U A L C O M P U T I N G & T E C H N O L O G Y • V O L U M E 1 6 I S S U E 662

BUSINESS