| Up to Dec.2012, the market scale of online advertising in China has brokenthrough75billion and growth rate reached46.8%, respectively.With the rapiddevelopment of the internet, advertisers and online media’s demand to monetize thenetwork traffic has becoming more urgent, the traditional operation mode of onlineadvertising can’t adapt to large scale online advertising transactions. As the bridgebetween advertisers and online media, ads network has changed the operation mode oftraditional online advertising. However, the ads network itself also has many pressingcritical problems to be resolved. Based on the game theory and contract theory, theurgent and critical problems of online advertising which based on ads network has beenresearched. It provides decision-making support for the online advertising operation, themain contents include the following four aspects:Firstly,22Bertrand price game model was introduced and integration ofnetwork media resources by ads network was analyzed. The result shows that optimizedand integrated advertising market structure is beneficial to every partner of the wholeonline advertising industry chain. Advertisers can save cost, ads network can achieveeconomies of scale and scope economy and network media can monetize its traffic,retaining relatively stable income. For choosing appropriate network media, theevaluation index system of network media based on the view of publishing onlineadvertising was built. The maximum deviation method based on deviation degree andpossibility degree of ulti-attribute decision-making were applied, integrated the precisedata and interval data and constructed an assessment model to solve the problem ofhybrid multi-attribute decision making.Secondly, a basic game theory model for click fraud monitoring was established,the model was extended to the two-echelon monitoring game model that the AdsNetwork would be punished by advertisers, and then further the two-echelon model wasextended into a regulation model when considering the competition among thepublishers. Analysis of Model shows that click fraud can be effectively prevented bymany ways including strengthening the supervision and control of network media, implementing punishment on Ad network by advertisers, reducing informationasymmetry, choosing the honest network media to publish advertising and building thecompetitive mechanism among network media. According to the perspective of doublemoral hazard, the proportional coefficient of fines and ratio of click fraud on networkmedia were introduced and the relationship between the revenue sharing coefficient andthe proportional coefficient of fines and ratio of click fraud on network media werediscussed. Analyzing on the equilibrium of the model shows that the optimal revenuesharing coefficient is negatively correlated with the ratio of click fraud, increasing thecost of click fraud is helpful to prevent click fraud.Thirdly, based on the perspective of risk-averse, the two-stage revenue pricingmodel between the ads network and online media was built and Nash bargainingmethod was applied to determine the optimal two-stage pricing problem. The feasibleconditions of the implementation of the two-stage contract were identified and therelationship between the negotiation power and the first stage pricing was analyzed. Thestudy shows that two-stage pricing contract brings ads network and publishers moreflexible options. The two-stage pricing not only makes ads network (advertisers) avoidhigher advertising costs, but also effectively stops click fraud, avoids the risks ofasymmetry information and publishers will gain extra revenues. The three pricing modelthat CPM and CPC, CPM and CPA, CPC and CPA were constructed by usingexponential utility function. And then a new pricing conversion rule among CPM, CPCand CPA pricing contract was put forward. The conditions under which the partnerswould agree on the three different advertising pricing agreement were analyzed, and theinfluence of click-through rate, action rate and risk attitudes on three pricing modelwere analyzed, too. The results showed the optimal pricing model for the ads network(advertisers) and network media under different conditions.Finally, by introducing Cobb-Douglas sales function and applying optimizationtheory and method, non-cooperative investment model, cooperative investment modeland dynamic two-stage investment model of single-period and two-period were buitbased on the portal online advertising and the long tail online advertising. The optimalinvestment revenue conditions were gained by analyzing model of single-period onlineadvertising investment and two-period online advertising investment. Learning effectsfactor and the discount factor were introduced into two-period investment model, the research results indicated that the bigger advertising learning effects factor resulted thebigger influence on later investment and revenue. Relative to the previous investment,increasing advertising learning effects factor, the later investment is smaller while therevenue is more. The two-period revenue is positive with discount factor, increasing thediscount factor, the more two-period revenue, when the values of learning effects factortends to be zero, two-period investment model will degenerate to a single-periodmodel.The results of some critical problems on online advertising operation have apositive effect in further improving online advertising theory and further promotingonline advertising market development. |
