Banks are feeling the pressure to implement new features on their transaction banking systems. Faced with competition from peers and all forms of third-party processors, the pressure on fee income is being felt across banks. Build vs. buy…to customize or not to…outsource, co-source or in-source…quality vs. stability; this list can go on. The bottom-line is that these are turbulent times for the technology initiatives of most banks.
A typical implementation of a complex transaction banking solution usually is painful. The average program lasts 18-24 months, takes up majority of the technology program resources and costs millions of dollars. Even with all this money and time, most banks ‘Go-Live’ with quality, stability, integration, and user experience issues. Just Quality Assurance spends can be more than US $3 million+ with at least 3 iterations to the original budget. All of this is not due to a specific product, but is the general nature of the beast. It is evident across products, projects, and banks.
Here are some implementation risks that we have observed and validated with customers:
- An average transaction banking system implementation can have anywhere between 6 to 10 vendor code drops depending on the size, complexity, quality, and stability.
- For each vendor code drop, internal development teams will usually match with a code drop of their own.
- Customers usually plan data migration from one system to another in waves. It’s only in the first wave that you realize there are mismatches, and you need to go back to the drawing board.
- Most product vendors bring a set standard of quality and stability of their applications to the table. Banks tend to over engineer with too many customizations, underestimate the complexity of their programs and are not always ready for projects of this size and complexity.
- A typical transaction banking system implementation will have a minimum of 3 full rounds of regression testing across the implementation. Smaller selective rounds are also run.
- Single Points of Failure and managing scope in these projects are some of the primary causes for many cost and schedule overruns.
- Quality and Stability of these applications are most vulnerable in customizations, integrations, environment, and data variables.
- A typical QA program for implementation of a transaction banking system lasts 18 months and costs approximately US $3 to US $4 Million. This is total cost of ownership that includes all activities such as user migration, data migration, test management etc.
The cost of finding ‘Go-Live’ impacting issues later in cycles not only increases the cost of fixing them but also could seriously derail your entire implementation program. The same goes for critical missed issues prior to ‘Go-Live’. One major issue could cause you to lose revenue or reputation, get pulled up by the regulator or just result in a very poor user experience.
A possible solution is an early detection, prioritization and resolution of issues that reduce your Total Cost of Implementation and subsequent release projects. This would involve establishing ‘Go-Live’ criteria upfront and using a combination of analytics, domain experience and application knowledge to carry out an objective assessment of the implementation to determine its readiness to ‘Go-Live’.