Abstract: Software as a Service (SaaS) delivers a bundle ofapplications and services through the Web. Its on-demand feature allows usersto enjoy full scalability and to handle possible demand fluctuations at norisk. In recent years, SaaS has become an appealing alternative to purchasing,installing, and maintaining modifiable off-the-shelf (MOTS) software packages.We present a game-theoretical model to study the competitive dynamics betweenthe SaaS provider, who charges a variable per-transaction fee, and thetraditional MOTS provider. We characterize the equilibrium conditions underwhich the two coexist in a competitive market and those under which eachprovider will fail and exit the market. Decreasing the lack-of-fit (or thecross-application data integration) costs of SaaS results in four structuralregimes in the market. These are MOTS Dominance! Segmented Market! CompetitiveMarket! SaaS Dominance. Based on our findings, we recommend distinctcompetitive strategies for each provider. We suggest that the SaaS providershould invest in reducing both its lack-of-fit costs and its per-transactionprice so that it can offer increasing economies of scale. The MOTS provider, bycontrast, should not resort to a price-cutting strategy; rather, it shouldenhance software functionality and features to deliver superior value. Wefurther examine this problem from the software life-cycle perspective, withmultiple stages over which users can depreciate the fixed costs of installingand customizing their MOTS solutions on site. We then present an analysis thatcharacterizes the competitive outcomes when future technological developmentscould change the relative levels of the lack-of-fit costs. Specifically, weexplain why the SaaS provider will always use a forward-looking pricingstrategy: When lack-of-fit costs are expected to decrease (increase) in thefuture, the SaaS provider should reduce (increase) its current price. This isin contrast with the MOTS provider, who will use the forward-looking pricingstrategy only when lack-of-fit costs are expected to increase. Surprisingly,when such costs are expected to decrease, the MOTS provider should ignore thisexpectation and use the same pricing strategy as in the benchmark withinvariant lack-of-fit costs.
