Research Spotlight

Pulling the plug on technical support? The commitment problem for software manufacturers

Continue technical support or end it? This was the dilemma Professor Debabrata Dey, Marion B. Ingersoll Professor of Information Systems at the Foster School of Business at the University of Washington, analysed in the paper he presented in the Information Systems seminar series at the Indian School of Business on August 31, 2017.

Continue technical support or end it? Not so long ago, this was the dilemma Microsoft faced with the XP version of its Windows operating system. First released in 2001, Windows XP saw its sales terminate end-June 2008, nearly a year-and-a-half after the launch of its next version, Windows Vista. Although Microsoft officially stopped providing mainstream technical support in 2009, in March 2014 (merely a month before the impending deadline for end of extended support), 30% of all windows machines worldwide had not upgraded to newer versions of Windows.

Microsoft found itself in a Catch-22 situation. On the one hand, unless it yanked off security patches and support for Windows XP, users would have little incentive to switch to the improved (and higher priced) Windows Vista 7 or 8. However, pulling the plug on support imperilled multitudes of Windows XP users , exposing them to security exploits, and in turn also posing a danger to Microsoft’s own reputation.

This dilemma lies at the heart of the ‘commitment problem’ analysed by Professor Debabrata Dey, Marion B. Ingersoll Professor of Information Systems at the Foster School of Business at the University of Washington, and his co-author Professor Dey - whose work lies in the broader arena of information systems economics, with a special interest in information security, information piracy, and associated issues - presented his research in the Information Systems seminar series at the Indian School of Business on August 31, 2017.

Generalising from the Windows XP dilemma, Professor Dey placed the context of his paper in the larger sets of questions surrounding ‘the security concerns of consumer, realised or unrealised’. “When a new version of a software comes into the market, what happens to the old version, which has an expiry date for support? If the vendor does not stop support as announced earlier, then its ability to profit from the new version goes down. If the vendor charges a high price for the new version, then a lot of people don’t upgrade, and closer to the expiry time you are left with a consumer base, which remains vulnerable to security attacks. So there will be a lot of consumer uproar, consumer disenchantment, and the vendor becomes compelled to extend the support. So in other words, we’re looking at a problem where the vendor is squeezed between the devil of not profiting enough from the software and the deep sea of disenchanted consumers because they are not getting the support they expect. This is the commitment problem — the inability of the vendor to commit to a firm expiry date hampers its ability to profit sufficiently from the new product.”

In the game theoretic model of the commitment problem, under certain conditions, Professor Dey stated that “the vendor will say one thing, but at a later stage, it will have the option as well as the incentive to do exactly the opposite. That’s why the commitment problem exists.”

The paper is able to identify a limit on the price set by the manufacturer such that “the consumer is convinced about its commitment to discontinue support. Interestingly around this price, the manufacturer’s profit actually increases with its security patch development cost.” Using rebates or coupons, the authors are able to devise a mechanism by which the manufacturers can “(credibly) signal that support will end.”

The abstract of the paper is as below:

Drawing a Line in the Sand: Commitment Problem in Ending Software Support
We examine the commitment problem faced by a software vendor in ending critical support, in the presence of network security risks. When releasing a new version of a product, in order to drive up its demand, the vendor must cease supporting the old version. However, the vendor’s ability to leverage the increased demand can be limited because of a commitment problem. For, when the demand increases and the vendor accordingly sets a higher price, many consumers might opt not to upgrade, creating a situation where stopping security-related support simply becomes too risky. To avoid this risk and any subsequent losses in reputation, the vendor can renege on its earlier decision to stop support. We show that this commitment problem hurts the vendor’s profitability and find that the no-commitment equilibrium profit can surprisingly increase with the cost to extend support. Accordingly, we propose a commitment mechanism. Further, the consumer surplus may actually increase if the vendor desists from crossing the proverbial line in the sand and discontinues support as planned.

Keywords: Software support, pricing, security, patching, game theory, commitment problem, analytical modeling, economics of IS.