Today's banking and financial services institutions must operate using an asset/liability management process that will be critical to their long-term success. An effective ALM software model is essential in carrying out the responsibilities required of ALM personnel. This article will discuss in detail the process of selecting and implementing a software model, with emphasis on vendor characteristics, product demonstrations, and quality assurance.
Vendor Selection. Vendor integrity is an extremely important factor in the ALM software selection and implementation process. Buying and implementing a new ALM package is a major financial, resource, and management investment. Investing the time and expense necessary to accomplish the evaluation will pay for itself over and over again in the future. Accordingly, important software vendor characteristics that should be considered as part of the ALM software selection process include the following:
* Industry expertise
* Product functionality
* Vendor responsiveness and availability
* Model integration skills
* Vendor financial stability
* Familiarity with your bank's systems and functional needs
Vendor Product Demonstrations. Those vendors who are selected to submit bids should demonstrate their products' strongest features to your selection committee. These demonstrations should incorporate your bank's data, thereby enabling the vendor sales staff to concentrate on your specific parameters. The purpose of viewing the product demo is to see the capabilities of the model, giving particular emphasis to your bank's specific needs.
A second product demonstration is important and should be held at your bank's location. The vendor should bring technical support staff to run the model, while the bank should provide the data that you intend to model. This exercise will provide an opportunity to ask a trained system user or consultant to perform tasks with the model that are not necessarily part of the standard demonstration. This will prove to be an ideal opportunity to see what the model is actually capable of doing. In this second demonstration, you will be able to evaluate the difficulty of using the software, reveal the product's limitations, and determine the suitability of the model in meeting your specific ALM needs. A reputable vendor will be more than pleased to offer such a demonstration, as it will provide them an opportunity to reveal the lesser-explored benefits of their software.
Vendor References. Checking vendor-supplied references is a natural part of the software selection process. Ask the vendor to supply you with a reference whose usage profile might closely match yours. This may sound like a very time-consuming effort but you will get to see the model operating in a real bank. Ask your host to demonstrate the system and, if possible, actually watch some parts of their production in action. Ask to meet their team privately to discuss their perception of the product's capabilities as well as the vendor's support and post-sale responsiveness.
Quality Assurance. An ALM software model deals with extremely complex financial algorithms. As a result, an entire facet of the development process should be dedicated to quality assurance (QA). The only real way for a developer to prove that a complex simulation is accurate is to build its equivalent independently, run data through the system, and then reconcile the results. If the results of such tests do not agree, at least one of the answers is wrong. This kind of testing assures users that the vendor is dedicated to offering a quality product that actually works. As a potential user, you will want to fully understand the developer's QA process because you want to know that the model is both stable and accurate.
Selection Criteria. Listed below is a summary of several essential criteria to be considered prior to purchasing your ALM software.
* Simulation analysis. A critical need in A/L modeling is the ability to perform strategic alternative analyses. This is a mission-critical characteristic necessary in today's turbulent financial environment.
* Available operating systems. It is imperative to inquire if a networked version of the model is what is being purchased or upgraded to, as some vendor models require their own server equipment and costly software/ installation procedures.
* Customization and system interface. User-defined asset and liability categories are essential to individual institutions in adjusting their reporting requirements. The ability for the system to interface with spreadsheet programs also allows for customized reporting.
* Historical data storage. The ability to store historical data within the primary system for eventual comparison reporting and historical trend analysis is essential to many users.
* Report generation. The generation of presentations and reports continues to be an important feature in A/L software.
* Internet. A key point to consider is the relevance of the Internet when selecting ALM software models. Determine if the software is available via the Internet; if applications can be run over the Internet; if specifics such as prepayment forecasts, historical rate and spread information can be accessed from database information providers; and if the prospective software vendors offer access to information through their Web sites. The Web can save time through mobile computing, and surfing the Internet for products alleviates the time otherwise spent on the vendor screening and interviewing process.
Decision Making. Deciding which ALM model to buy is always a challenge, even if you have prior experience. Along with the criteria previously mentioned, consider paying for training before you buy. When you attend a full training course in advance of purchase, you will get the chance to examine the software as it really is, better understand how you would model your balance sheet, and share the experience with others who have already purchased the software.
The opportunity to actually work with the model directly is one of the key elements of truly understanding what you are buying. With training, it is entirely possible to make a studied evaluation of the model. All the demonstrations in the world will not equal one hands-on training experience. Accordingly, pre-purchase training is often worth its weight in gold.
Additionally, back testing, the process of historical replication, affords you the opportunity to test the model against known values. The goal is not to have the model exactly predict the historical future, but rather to reconcile your beliefs about how you think the model works with your actual historical data. With back testing, it becomes possible to make an informed decision about the adequacy of the simulation's thoroughness.
The amount of time invested in a successful ALM software selection process will be significant. However, this important decision will be a critical part of your bank's future success. Consequently, the decision process must be objectively approached with thoughtful and rigorous analysis.
Source Citation
"ALM software--making the correct purchase." Bank Asset/Liability Management Oct. 2010: 4+. General OneFile. Web. 27 Nov. 2010.
Document URL
http://find.galegroup.com/gps/infomark.do?&contentSet=IAC-Documents&type=retrieve&tabID=T003&prodId=IPS&docId=A239348252&source=gale&srcprod=ITOF&userGroupName=18551_mcpls&version=1.0
Gale Document Number:A239348252
No comments:
Post a Comment