As enterprises increasingly turn to managed services models, pricing-by-application is gaining attention as an emerging pricing model in the application development and maintenance (ADM) discipline.
The framework has two defining attributes – a price tag is assigned to every application based on an estimate of the cost and complexity of supporting the application, and pricing is completely decoupled from FTE counts.
As such, the model departs from traditional pricing, which links an application portfolio to a pool of FTEs, with contractual measures for adding and removing applications. This traditional model – in alignment with the dominant governance and management philosophy in contemporary ADM sourcing – relies on managing service provider resources closely, and knowing at all times the number resources working on the project. Under a managed services approach, monitoring output rather than monitoring effort becomes more acceptable. This very gradual evolution enables the growing acceptanceof this new model of pricing application maintenance projects.
Pricing maintenance by application can be done by segmenting applications into tiers with a fixed price for each tier, or by pricing each application individually. In both cases, price depends on the cost and complexity of supporting the application (s) in question, and is independent of the number of FTEs. Application maintenance cost depends on the level of support required, incident volume, volume of backlog defects, number of users, number of interfaces, and number of transactions, among other factors (but interestingly, not on the size of the application, as measured by function points or other methods).
Some of the drivers of application maintenance costs are variables that the enterprise can control. The demand profile of certain applications is seasonal and may vary across months and quarters. Price of maintenance depends on users and transactions. Therefore, the pricing tier of the application in question can vary across a calendar year. In addition, across a portfolio, different applications require different levels of support (based on support window and SLA tier) and pricing can reflect that difference. All of these factors have an impact on optimizing application maintenance spending.
In practice, such engagements typically start with conventional fixed-fee pricing models, specifically a portfolio of applications linked to a fixed resource pool. Over a period of 12 to 18 months, the service provider and the client collect enough data to construct a model to discover the cost of supporting each application, and how costs vary with support tiers, users and other controllable variables. At the end of the defined period, the enterprise decides whether to migrate to a new pricing regime.
Per-application pricing enables clients to finely modulate application maintenance spending, allowing them to buy exactly the level of service required. Also, the framework incents the service provider to introduce automation and other efficiencies such as resource pooling. In an already mature sourcing environment with a large offshore component, this framework can reduce costs further. For service providers, a pricing scheme decoupled from FTE numbers creates an opportunity to grow margins, and aligns with the industry’s current focus on discovering non-linear sources of revenues.
Per-application maintenance models are still rare, accounting for well under ten percent of totals. However, ISG expects adoption to grow as both clients and service providers have solid incentives to at least consider the new framework.