How Application Classification Framework impacts Infrastructure Support services

Enterprise IT managers have actively been moving to outcomes based commercial constructs such as Managed Services for operational functions such as, Application Maintenance and Support. This model has been successfully industrialized for Infrastructure Support (IS) services, wherein the entire IS service work gets efficiently decomposed into well-defined Resource-Unit (RU) which can benchmarked, priced, bundled and governed by outcome-based SLAs. Application Maintenance and Support (AMS) has lagged in its industrialization journey partly due to lack of consensus on how to decompose various services into units of work.

AMS Managed Services deals are usually based on fixed prices for application bundles or projected ticked volume or a combination of the two. These deals leave most of the control and visibility of efficiencies in the hands of the vendors who are often the primary beneficiaries of the said efficiencies. Some leading organizations have started experimenting with AMS deals based on per application pricing driven from a classification framework covering variables, such as workload, criticality, etc. These frameworks not only create a unit-driven pricing mechanism for AMS but also facilitate proactive application portfolio management. The lynchpin of this approach is a robust application classification framework, often the least understood aspect of managed services model due to the variety of frameworks being used by vendors, often to their own advantage.

Many organizations are frustrated by the ‘Black-Box’ nature of many Managed Services deals and seek to gain a better understanding of the services they receive and the levers employed to deliver efficiencies. Our framework attempts to provide the necessary transparency while staying true to the spirit of supplier owned managed services delivery model.

Application Classification Framework

Our experience indicates that a move to a standard application classification framework can yield several benefits:

• Consistent and objective definition of effort needed for application support, as it is driven off operational data
• Allows for easy change management in case of Adds and Deletes in the portfolio
• Facilitates a structured approach to driving efficiency in the portfolio where the vendor must demonstrate increasing application stability and reducing workload
• Minimizes wild swings in workload estimates as it is not driven by ticket volume only

An adequate model must reflect the complexity of the application landscape while not creating a large administrative burden. It is also critical to define the model prior to engaging with any service providers. The key principles of a classification framework include:

1. Objectively and transparency – be data driven (ticketed and non-ticketed workload, business impact, etc.)

2. Technology requirements and support risk profiling

3. An ability to predict support cost for new applications

We propose a four-stage approach for developing an Application Classification Framework (see Figure 1). The framework enables a structured Change Management process that allows additions and deletion of applications from the portfolio.

1. Application Profiling: The process starts with creating profiles for applications in the inventory based on set criteria that includes Activity/Workload, Technical Complexity and Support Criticality. Each profiling criteria comprises numerous sub-parameters to ensure each application is profiled to an adequately granular level. The next step is to gather historical data for each of the sub-parameters. The time-period for which the historical data is gathered for profiling is critical to ensure that seasonal spikes in support activity (at quarter or year end, for example, or during periods of very high usage), get adequately covered.

2. Attribute Weights and Scoring: Once historical data has been gathered for the sub-parameters under each of the profiling criteria, the IT and business needs to collaboratively assign weights and scores for each sub-parameter.

Weights are applied to each of the sub-parameters, an example of which is shown in Table 2. Weights should reflect the organization’s relative tolerance towards specific sub-parameters.

The sub-parameters are then Scored on a 5-point Likert scale. The data collected for each sub-parameter in step 1 need to be normalized on 5-point scale. (see Table 3 for sample scale definition for Number of incidents sub-parameter)

3. Application Profiling Scores (APS): Application profile score is computed as an aggregation of attribute scores and weights. The applications are then classified into bands, based on these scores. (see Table 4)

The sample structure shows ‘A’ class applications are of the highest importance to the business and hence will need to have the strongest SLAs, penalties, and monitoring.

4. Pricing Bands based on Application Classes: Post classification, the pricing bands get defined for each Application Class. The price band needs to be defined in consultation with the vendor – the highest scoring class commands the highest price per application regardless of the functional area it supports. Lower classes reflect lower workload and hence lower prices.

At the end of each billing period, the supplier bills the client based on the number of apps in each Class through a “Price-X-Quantity” exercise. This process also facilities change management as new applications added to the portfolio are added to the quantities based on tehri class and the ones removed are likewise taken out of the count.

Impact on Application Portfolio Management

The application classification framework is not merely a tool to manage contracts with vendors; it is in fact a key component of the application portfolio management framework. For it to have the maximum impact, the application classes must be tied to an application rationalization efforts or the typical ‘Invest’, ‘Migrate’ and ‘Eliminate’ framework. 

Vendors should be required to shift apps from the higher classes to lower ones through reduction in workload by automation, shift left or other efforts leading to improved portfolio stability.

To summarize, large AMS organizations can receive a one-time transactional savings from signing a Managed Services deal, however, sustained year-over-year improvements are likely to remain elusive in the absence of a structured AMS framework – one that aligns the incentives for the buyers and suppliers while providing transparency to all parties. The proposed framework helps standardize and industrialize a service that has hitherto been challenging to manage for many organizations.

New Balance Mens Shoes