When the run started on Northern Rock in mid-September 2007, it revealed two important things. The first was that banks had become over-exposed to high-risk lending. The second was that nobody had a clear idea about the net number of depositors and the money they were owed.
Nothing has yet been done to fix the first issue. On the second, the Government immediately increased the amount of deposit guaranteed under the Financial Services Compensation Scheme to £50,000. The Financial Services Authority has spent this year in consultation over how to give the FSCS the right tools to pay out in the event of a future bank collapse.
The consultation paper FSA 09/16, “Verification of the single customer view and changes to deposit compensation”, sets out a plan of action due to come into force next year. By the end of July 2010, any deposit-taking institution will need to have an implementtion plan for a SCV, with the actual integrated customer data warehouse in place by end of December.
As well as giving systems integrators and consultancies a potential £1 billion gift, the plan also poses some real challenges to banks. A key issue is whether they have the skills, resources and experience to bring off customer data integration on this scale. In most companies, not just banks, the only function to have built a SCV is marketing. So is its approach robust enough to give it a central role in the FSA-mandated project?
Tom Scampion, partner in the information and technology risk practice at Deloitte, notes, “when we have been asked to do this before, we find it is sitting in the risk or compliance function. Since the board has to sign-off the FSA requirement, that gives the project a high profile so it goes into that channel.”
Even without a regulatory push, many financial institutions were starting to look into SCV to give a better view of their risk exposure. “If someone is a reckless borrower with maxed out credit cards and a heavy mortgage, risk wants to know about that before making any other lending decisions,” says Scampion. By giving marketing access to this view as well, it can avoid propositioning those high-risk individuals. However, these two functions do have different needs in terms of the latency of information that can be tolerated and the speed of update necessary. An example Scampion gives is where a key has been assigned to households and two people move into one address, for example as a result of marriage. “Now you have got two keys – which one do you keep without losing the history of the relationship?” he asks.
This is where two very different ideas about how to approach SCV come into conflict. To support FSCS, the FSA proposals clearly intend “industrial-strength matching” of customer records to ensure that individuals get the money they are owed, but are also not over-paid. This is a potential issue in the event of a bank failing since the unmatched residue of records creates a liability that the asset base may not cover.
When the banks sit down to decide how to meet the new requirements and choose the team to lead their project, they may be spoilt for choice. How they exercise that choice will not be easy.
“My experience, which is mainly based in financial services, is that the problem is not that companies do not have a single customer view – it is that they have several,” says Richard Smithies, head of database marketing and analytics at LaSer UK (formerly Creation Financial Services). “Often there are multiple versions of single customer views that exist in an organisation.”
On one side, an operational view of the customer will have been buiilt to assemble all accounts held, usually for annual statementing. “In this example, the way accounts are matched to customers is normally based on customer matching rules that are extremely tight,” says Smithies. Accounts have to be definitively associated with each customer since getting this wrong will trigger complaints and potential liability.
On the other side, a marketing customer database will have been built, often with more liberal matching rules. “If data matching suggests that two customers are highly likely to be the same, but not absolutely 100 per cent certain, allowing the records to match may be desirable, from a marketing perspective. One of the main benefits, for example, is less duplication in marketing campaigns, resulting in reduced campaign costs and improved response rates,” he says.
Boards are likely to assume that the functions which built the operational SCV, such as IT, have the most valuable experience. Yet Smithies believes marketers have more practical experience of overcoming the challenges in such projects, especially data quality. “While inviting marketers to lead IT projects might be seen by the IT department as akin to putting the fox in charge of the hen-house, there are a number of common sense reasons why this can work,” says Smithies.
Usually, when a regulator adds a new requirement to the slate of things that regulated businesses have to do, they bear the cost with little complaint. In the current economic crisis, however, deposit taking institutions may want to look to leverage some business value out of the SCV.
Where that is the prevailing attitude, then it seems likely that the data integration project should not be viewed in purely technical terms. Those who have been through the process argue for a much broader perspective.
Bertrand Labays, general manager customer data and analytics at TUI UK, says: “The key point is that it has to be driven by business, as opposed to IT. When you are drawing up the requirements and functional specification of a system, it is important that it is led by business people.”
He first started pushing for an integrated view of customers four years ago when “no-one was interested in customer data. I was barking around to get sign-off and eventually got the approval of the marketing director and some board members.” Even with that backing, it took two years to get the funding and access to data required for a pilot “customer single view” to provide a proof of concept.
Senior buy-in also proved crucial when the “sparse IT resource” of the company was dragging its feet. “I was in a meeting with our commercial director about the project and he had to phone the IT director to push him to give me access to the data,” recalls Labays.
When the Thomson customer single view was finally built, it created the business case for further investment and expansion of the project. Without unifying data in this way, the company’s data model was obsolete and did not support customer-centric marketing and service, which the business had recognised was critical for competitive advantage.
Labays recalls that, “at the time, nobody was interested because there was no pressure on sales. CRM was in a cupboard on the fifth floor.” Like many companies during the boom decade, growth was so easy to come by that few felt the need to adopt best practice.
In changed economic conditions, this mindset has also changed. The second phase of TUI’s customer single view, called Pluto, is pushing information about holidaymakers out to the front line so that resort-based representatives know who they are talking to and their value to the business.
“I was able to demonstrate to the business that sharing customer data with the frontline would add value. I didn’t ask for money, just a small scale project to develop learnings about how to use that data, because customer service had no access to any customer information,” notes Labays
While it is often the persistence of individuals like Labays that can drive through a successful SCV, for the banks, there is little choice but to get theirs up and running. They also do not have the luxury of time, given the constrained timescale dictated by the FSA.
However, in many instances they are not starting completely from scratch. “That would be to underestimate the extent to which banks in Europe have implemented operational single customer views – most banks do have one,” argues Simon Doherty, European financial services consultant at Teradata.
Pressure from other regulations, such as Basel II and Anti-Money Laundering laws, have been the drivers of this compliance-oriented investment. Smaller institutions and those in emerging markets may be behind the curve, he points out. “There are also some banks that for one reason or another haven’t leveraged their SCV, either because their CRM programmes are less well advanced or in some cases because they are still fragmented,” says Doherty.
There is another reason for failing to drive value out of this asset. “Some bank managements have a culture that is anti-information,” he says. Determination to show leadership and rely on gut feel is precisely what led RBS into trouble, for example. Even after the credit crunch, there will be institutions that resist a more scientific approach to doing business.
Richard Lees, chairman of The Database Group, gained experience of building enterprise-scale SCV when working at Standard Bank in South Africa, before setting up his bureau business. “They needed to understand customer potential to make 20 to 30-year decisions on properly leases and whether to open or close branches. It was driven by property,” he recalls.
One of the challenges he identifies in building SCV that supports business functions as well as compliance is around the language and definitions used. “What marketing views as a current customer is not the same as how risk might view them,” he notes. Creating a data dictionary to align all of these views is one of the difficulties that will be faced by any organisation that wants to leverage a central data resource.
While not unsurmountable, this need for multiple views does lead to complex data models. Information critical to operational functions, such as the outstanding balance on a mortgage and the current rate at which interest is being calculated, may be less relevant to marketing than when the mortgage was opened and how long it has left to run, for example.
This can be one of the obstacles preventing completion of SCV projects. “I’ve worked with on insurance company that brought us in four years ago to build a SCV and they still haven’t managed to get there,” says Matthew Spencer, head of vertical markets and Authenticate at Experian QAS.
He argues that data quality will be a critical dimension in the success of any project. In order to merge account data successfully, the matching points, such as name and address, need to be verified, validated and enhanced. This is a process that some financial institutions can resist since they see the original record as sacrosanct, even if it was created before any data quality programme had been put in place.
As well as ingrained cultures, the other barrier to SCV can be what Mike Hudgell, business development director at Evaxyx, calls “business as usual. The challenge a lot of the time is to show the owners of lines of business that there is a benefit to them. They rarely understand the value to them of sharing their data with other units.”
If the FSA proposals have their origins in the political need to transform the banking environment, then their consequences will be highly political for the banks involved. As anybody working on a SCV project knows, getting disparate functions and business units to agree is never easy. A single view of the customer may be achievable, but only if you can first create a single voice of the business.