The panel session on enterprise data management at Swift’s Sibos conference in Boston earlier this month was a little light on detail to satisfy the true EDM enthusiast, but as a reference data 101 it was nonetheless a lively and interactive debate, ably facilitated by the EDM Council’s Mike Atkin. A series of audience digivotes also yielded some interesting stats (see charts), although it should be borne in mind that the listeners were an eclectic mix of payments and securities specialists with a primarily back office bent. Notable by their absence was a representative from Swift on the panel, perhaps surprising given the co-operative now has a reference data strategy based on BICs and a supposed effort around standing settlement instructions (Reference Data Review, September 2007) – though announcements about the progress of either were not forthcoming during the event as a whole.
As far as insights went, the territory covered was pretty familiar. All the panellists – David Goldberg, executive director and global head of data management at UBS AG Investment Bank, Ken Price, CEO of Avox and Thiyagarajah Rajah, chief information officer at CLSA Ltd (Hong Kong) – made reference to the impact of greater regulatory scrutiny as a key driver for investments in data management. Price suggested that an evolution is currently under way, however. “The regulators have done us a great favour in getting us to dig so deeply into reference data,” he said. “Having consistent quality and an up to date reference data platform enables firms to capitalise on new market opportunities. If the reference data is there it is easier, for example, to implement a new trading platform.” Rajah agreed: “From a reference data point of view a major driver is regulation. But what keeps management ticking is the business issues.” He also cited good old cost as a continuing driver for data management improvement efforts. “When you have cost as an issue, there is no question that you need to centralise,” he said. And Goldberg threw STP – that much tarnished moniker – into the mix as well. “STP is driven by good counterparty data,” he suggested. “If you don’t have it, costs are driven up.”
An audience vote on the types of reference data considered most suitable for centralisation ranked customer, counterparty and entity data as number one – followed by security master data, with SSIs coming up third. This in part could be reflective of the audience make-up – there are likely to have been a number of people in the room who are intimately acquainted with Swift BICs, for example – but it also reinforces a definite trend of EDM efforts looking beyond just instrument data. The results of the vote on the importance of unique identification of legal entities for EDM are also worthy of note – 65 per cent said it was essential, and 30 per cent rated it important – with only three per cent saying it was irrelevant (and a curious two per cent answering “Don’t know”). If Swift’s plans for extending the BIC into a universal identifier are picking up pace as many suspect they are, then this result was doubtless music to the co-operative’s ears. That said – as Atkin told Reference Data Review after the session – you don’t need an industry standard to implement unique and consistent entity identification internally.
As is so often the way, much of the most interesting segment of the discussion focused on the challenges to EDM implementation, in particular challenges related to data ownership, governance and internal politics. When asked “How much of an issue is data ownership in your firm?” 41 per cent of the audience said it was a “major issue that negatively impacts”. Thirty per cent said it was an issue but was not creating a negative impact yet. Price suggested a distinction needs to be made between ownership and accountability. “Ownership is bad, because the owners of the data want to hold on to it, and its universality is then immediately under threat,” he said. Rajah agreed that “as soon as someone owns the data, it’s trouble”.
The politics around enterprise data management are “very difficult”, Goldberg said. “When data is a shared activity, the politics around allocation of resources et cetera need to be carefully managed. This comes back to data governance, and having the right leadership in place. If data management is technology-led or compliance-led, it becomes a technology programme or a compliance programme. We’ve aligned into the CFO’s office – looking across the enterprise.”
Getting buy-in from the C-suite is critical, said Rajah. “Without it, it won’t work,” he warned. “Without organisational and structural change, EDM is not going to happen. Data is certainly discussed around the C-table. The major issue is getting the agenda right and all the players around the table.”
Returning to the also familiar theme of business case building – specifically, that it’s challenging to do for EDM – the panellists were united in their belief that securing quick wins is the way forward. “Doing a complete business case for an enterprise wide data management strategy would probably cost more than implementing the strategy,” Price said. “You have got to keep your eyes open for the opportunities to make quick wins.” There are “so many wins you can have”, agreed Rajah, “and you need to get those hits. It then becomes easier to drive programmes forward.” Tangible value has to be generated in year one of a programme, Goldberg contended. “The worst situation is to have a very large, structured programme that delivers nothing in the first few years of its existence.”
A tricky conundrum is that to prove a business case and the ongoing demonstration of value from an EDM project, you need to put metrics in place – but, as Price said, there’s a chicken-and-egg scenario in play, because you need to justify investment in the measurement capabilities in the first place. Depressingly, 51 per cent of the audience confessed that they do not currently attempt to measure performance of EDM. Of those that do, the biggest chunk said they use internal KPI assessment. As Goldberg pointed out, the hardest part of measurement is also the most important – the root cause analysis of failures. “What’s difficult and expensive is drilling down into the whys,” he said.
With a nod to his hosts, Price gamely called on the audience to rally around the metrics cause. “As part of the Swift community you have an opportunity to consolidate metrics and create industry benchmarks,” he told listeners. It remains to be seen whether the Swift community takes this role on: perhaps we’ll find out next year in Vienna.