IBSI Announces Subscription Flexibility !

IBS Journal now starts at £9 per issue, and custom Reports at 40% off
READ MORE

The thorny issue of the Return on Investment (ROI) on a system replacement project has flummoxed the industry for as long as IBS has been reporting. A recent survey by IBS showed up the continued failings. What could be done to bring more science to this area?

What are the benefits of replacing a core banking system? The industry has pondered this conundrum for as long as IBS has been reporting. However, the musing has seldom resulted in anything conclusive. Individual suppliers attempt to do some measurements from time to time (Misys is doing so at present) and occasionally they make claims around, for instance, the profitability of their customers in comparison with the rest of the market. At an individual bank level, such analysis is even less common. Given the risk, time and cost involved in a core banking system replacement, the lack of metrics is alarming and would cause bafflement from people in other metrics-driven sectors, such as retail.

Would measuring ROI have been a useful exercise?

Would measuring ROI have been a useful exercise?

IBS’s pondering brought a survey of banks that had replaced their core banking systems, with a good response (over 50 detailed replies). The problem was that only seven of the respondents had done anything like a detailed Return on Investment (ROI) study of their projects. So we learnt a lot about why they didn’t do such projects, but little about what happened when they did. Moreover, of those that expressed an opinion, almost two-thirds of respondents who failed to do an ROI study felt it would have been a useful exercise.

The most common reason for inaction was a lack of senior management demand, followed by a lack of pre-project metrics and an inability to accurately measure some or all of the benefits. And there, in one question, we probably have the crux of the matter.

Of the reasons not to do such a study, perhaps the lack of senior management buy-in is the most surprising. Again, in comparison with other vertical sectors, this is a notable failing: would a large retailer, for instance, be able to carry out an IT project of this scale without well-formulated metrics? While project failures can be career-threatening for senior managers in banks, as in other industries, there is not the transparency and scrutiny, perhaps, to encourage a more rigorous, scientific approach to projects.

The other two main reasons for not doing an ROI study are more logistical ones. If you don’t measure the current situation then you have no benchmark against which to measure the benefits of a system replacement. Few banks have a good understanding of their processes and the associated costs. Even in an area that is more ‘industrial’ than others (such as processing letters of credit – often still fairly manual but a more or less standard product which could lend itself to measurement), there is seldom an understanding of the costs, either purely internal or in relation to the rest of the industry. This can be a hindrance to process reengineering per se, regardless of whether it includes a system replacement.

The inability to measure some or all of the benefits means that, basically, most people just feel it is too difficult. It might be feasible to measure pure efficiency (e.g. how many staff does it need to process an FX deal or has there been a reduction in software and hardware costs?) but it becomes far more nebulous when you look at the competitive benefits. An increase in customer acquisition or retention will be from a range of factors, including economic and commercial (the attractiveness of the offerings, the financial institution’s appetite for risk etc), so to what extent can changes in business and profitability be attributed to a system replacement? And how do you measure, for instance, improved management information or improved compliance?

Even those benefits that might be attributed to the project itself are not necessarily to do with the technology, per se. As stated by one north American respondent to the survey, which had done an ROI study, ‘ROI is not only directly dependent on the system but also on the success of the people change management’.

As well as the difficulties, a fair number of banks that did not measure ROI cited lack of resources. Indeed, it is likely that considerably more banks set out on their system replacement projects planning to do ROI studies than actually do so. They quickly become embroiled in the reality of the selection and implementation projects. Everything comes to focus on going live within timescales and on budget, and the internal view of the success or failure of the project tends to end up purely on this. After go-live, there is bound to be a busy period of bedding in the system, ironing out technical and process issues, adding patches, additional training and testing, additional milestones to pass, such as end of month, and then additional phases in the roll-out (new modules, more branches etc). There is never a moment of calm to allow for reflection, with merely a sense of relief and fatigue before moving on to the next challenge.

It could also be, of course, that the lack of measurement and openness stems from a lack of confidence about the results. It could be deemed politically and diplomatically better to leave things vague if the results are expected to disappoint.

One other problem is likely to be the lack of industry-wide metrics and expertise. None of the banks that did an ROI study turned to external help, using instead a mix of internal business and IT staff. The most important criteria were felt to be reduced support and maintenance costs (which is arguably the easiest thing to measure), improved time to market for new products, and reduced manual processing and errors, in that order. And all would continue to measure ROI on future projects, so presumably they found it a worthwhile exercise.

What could be done to improve the situation? The industry could do with consensus around what should be measured, how it should be measured and how the information might be shared to constitute a meaningful benchmark. A lot of the time, the topic only becomes of interest when a bank has made the decision to replace a system, rather than fuelling that decision in the first place. However, the analysis should be of interest to banks regardless of whether they are intending to replace their systems, allowing them to see their efficiency in relation to their peers, allowing them to build a business case for change (process reengineering, as well as investment in technology), and monitoring and measuring the outcomes, with the lessons carried over to future projects.

Figure 1: Cantonal banks ranked by corporate group cost/income ratio

Figure 1: Cantonal banks ranked by corporate group cost/income ratio

We’ve looked before at the pros and cons of using cost/income ratios to measure efficiency, taking a subset of banks and analysing the facts, figures and arguments. Measuring like for like is always a major obstacle to benchmarks, which is why we chose the Swiss cantonal banks, as they have fairly similar business models.

One thing an individual banking system supplier could do is measure the efficiency of its own users (the aforementioned Misys talked about doing this some years ago, purely in a standard area such as FX and MM for its Midas users, but nothing came of it). There is still a tendency across the industry for suppliers to sit back and wait for RFPs, letting the banks come to their own conclusions about needing to replace their systems and only then responding, rather than doing much to sow the seeds of change in the first place.

Cleary, IBS interviews many, many banks that are embarking on projects, have completed them, or are mid-journey. Barely a day goes by when someone in IBS Towers isn’t speaking to a bank, somewhere around the globe. We ask fairly stock questions about the business drivers and the benefits, we ask about whether the latter were measured, but on most occasions the answers are high-level and, even if the vision is ably communicated, there is an inability to drill into detail. Commonly, there is an intention to measure the benefits at some time in the future but, nine times out of ten, this good intention does not seem to be fulfilled.

It isn’t good business practice and it isn’t helpful for the industry as a whole nor for the individual financial institutions, but we still seem to be in the Dark Ages when it comes to ROI on major system investments and, despite the tough environment and reduced budgets for many of the last few years, which might have focused minds on such things, there seems little sign of the sector finally addressing this.

On the plus side…We asked about any unexpected benefits from system replacement projects. Among the comments were:‘Being connected to an international peer group’
‘Standardisation of development methodology’
‘Possibility to outsource running the baseline to vendor’
‘Using their EDF toolkit, it allowed us to develop and add quite a few additional data screens and fields on our own without having to resort to the software house. This has greatly reduced development costs by the supplier’
‘Ease of integration to other group systems’

 

To what extent did you conduct process reengineering BEFORE core system replacement?

To what extent did you conduct process reengineering BEFORE core system replacement?

To what extent did you conduct process reengineering DURING core system replacement?

To what extent did you conduct process reengineering DURING core system replacement?

Conducting process reengineering AFTER core system replacement

Conducting process reengineering AFTER core system replacement

Was the new core banking system delivered on budget?

Was the new core banking system delivered on budget?

Was the new core banking system delivered on time?

Was the new core banking system delivered on time?

Important criteria

Important criteria

Reasons for not conducting ROI

Reasons for not conducting ROI

Important criteria

Important criteria

 

by Darshana Adanwale
imp-loader
preloader