|
International Issue: September 2006
Feature
How dirty is your data?
Inaccurate network inventories and faulty customer ID information will wreak havoc on operators’ broadband access plans
by Ken Wieland
The call centre agent and the network engineer, inadvertently, are helping telcos waste millions of dollars a year. Through human error — such as the inputting of wrong customer ID information at the call centre, or the failure of the engineer to log properly any network alterations — they are contributing to one of the biggest and most expensive problems that telcos face today: dirty data.
In July 2006, the Yankee Group — reporting the results of a survey commissioned by Evolved Networks — said that inaccurate network inventory records on local access assets alone would add about €280 m per year to the bottom line of an operator with 30 million access lines, which is roughly the size of BT and France Telecom. (Evolved Networks, a privately-funded company spun out from BT, says it has network data quality solutions to address this problem).
And most organisations — not just telcos — aren’t doing a particularly good job at cleaning up their dirty data problem. According to a survey conducted earlier this year by Trillium Software, a US-based provider of ‘total data quality’ solutions, only 14.5 per cent of respondents — the survey sample comprised 216 attendees, from ‘large companies’, at data quality and data management industry events around the world — were tackling data quality issues at a strategic, enterprise-wide level.
By contrast, most of Trillium’s survey respondents — 51 per cent — said they had a tactical approach to the problem. That is, tackling data quality issues — as and when they arise — on particular applications without having an overall data quality control strategy in place (Figures 1 & 2).
Ed Wrazen, who heads up Trillium’s marketing activities across the EMEA region, argues that bad quality data and ineffectual short-term remedies can ‘cripple’ organisations. “Many companies from different industry sectors are facing similar, major challenges,” he says, “such as the implementation of CRM [customer relationship management], SCM [supply chain management] and BI [business intelligence] initiatives. While there are a number of factors that determine the success of these initiatives, including the adoption and acceptance by the company’s workforce, the main reason why many don’t succeed is because they are riddled with data quality problems.”
The problems that Wrazen refers to include customer ID information, residing in different databases, that don’t match. “It’s often the case that organisations don’t discover this problem until they are about to go live, say, with their new CRM implementation. There is then a mad scramble to fix the problem but when that happens, the project usually fails to meet user expectations. According to Gartner [December 2005], some 50 per cent of CRM projects fail to meet their targets, whether it be budget or timetable.”
Due to greater regulatory pressure to improve corporate governance — led by the Sarbanes-Oxley Act — Wrazen notes that organisations are increasingly being forced to think more strategically about data quality control management. “This should not be seen as an IT department problem,” he says. “It’s something that needs to be initiated from senior management across the whole enterprise.”
Nigel Turner, head of information and knowledge management control at BT, believes the scale of the problem is only going to get worse if no action is taken. “The reason for that,” he explains, “is that organisations, on average, are typically creating 50 per cent more data than they have in the previous year. You are not dealing with a fixed problem but one that is accelerating.”
What can be done?
BT, arguably, is one of the world’s leaders in developing a strategic approach to cleanse data. Addressing the problem as far back as 1997, BT set up a formal data quality programme in 1998, which ran to 2004. During that time, Turner says that BT made cost savings of around £600 m (US$1.14 bn).
After BT finished the programme in 2004 (the processes introduced became business as usual) Turner and his team now act as consultants — operating within BT Global Services — to advise other organisations in different industry sectors. Trillium is one of BT’s software partners, and Turner is finding growing demand from the utilities, financial services and government sectors for his team’s data quality expertise.
So how did BT save £600 m? “The first thing to recognise is that this isn’t just about cleansing data or simply an IT problem that can be fixed by software,” says Turner. “A cultural change is required within organisations on how employees view the importance of information. You’ve also got to convince senior management in terms they understand best — the financial benefits.”
When Turner first looked at the problem back in 1997, the first thing he did was scour around for existing best practice in other organisations. But apart from a few IT companies, such as HP, Turner quickly came to the conclusion that there wasn’t a lot of best practice out there. “We built our own data quality improvement methodology and used best-of-breed toolsets, one of those being Trillium. We were particularly interested in their name and address [NAD] management software.” By getting the NAD database more accurate and up-to-date, Turner talks about the ‘ripple effect’ that this can have in other parts of the business. Billing is more accurate and the network engineer is less likely to go to the wrong address, which wastes time and money.
There is also the customer satisfaction angle. “Who wants to do business with companies that get your name wrong?” asks Turner.
For levels of ROI (return on investment), Turner said that these ranged from 4:1 to 40:1 across the various projects that BT undertook during the data quality programme. “The biggest ROI returns we managed were in network inventory management [using the Trillium Software Discovery tool],” he says. “Prior to using Trillium, if a customer ordered a private circuit and then changed his mind, we had no processes in place to put that line back in stock. That meant that we would commission new stock when we didn’t have to. Savings on capital outlay were very big for us.”
But Turner emphasises that the data quality programme was as much to do with setting up the correct processes as software. “A lot of it comes down to training the workforce on the importance of having accurate data and to give people different sets of objectives to reflect that,” he says. “For example, the measure of a call centre agent’s efficiency used to be how many calls they could handle in an hour. Now we’re looking at how accurately customer ID information can be captured. We have software that can check that, so incentives can be given to those who are the most successful. This is not just a back-office issue but also a front-office issue. We have to make it as difficult as possible to input the wrong information.”
Within BT, the next data quality priority is to establish an ‘enterprise information’ platform. The intention is for that to become the definitive source of all key customer information. “Through this database, we will get the so-called 360 degree view of the customer, as it will pull in all the product, service and billing information into a single place rather than data being dissipated across numerous systems.”
Problems in the local loop
An immediate problem for operators to address is inaccurate local access network inventories. According to the Yankee Group survey in July this year, this inaccuracy has the potential to wreak havoc on operators’ broadband business case plans. The research and analyst firm calculates that up to 50 per cent of existing local network records in Europe and the US — that is, the copper loop — are just plain wrong.
“Without having accurate information at the local loop level, operators will jeopardise the success of their higher-speed services, such as IPTV,” says Matthew Edwards, CMO of Evolved Networks. “For the next five years or so, these services will need to be delivered using DSL technology so the copper network is not going to go away any time soon. But if the data records of that network are incorrect then it’s going to cost money and increase the likelihood of churn.”
The costs related to inaccurate date records include a higher number of truck rolls to fix network problems as they are more difficult to locate without up-to-date records; inefficient capex outlays (it is more difficult to identify which part of the network requires investment); and higher customer dissatisfaction (customers may be promised a broadband service they are not eligible for due to the operator not knowing the quality or length of the local loop).
Evolved Networks’ approach to this problem — and one which it claims to be unique in the marketplace — is to crossreference both logical inventory data (information around the circuits that feed the billing systems) and physical inventory data (which describe the geographical layout of ducts and cables).
“The logical data is believed to be around 99 per cent accurate, so what we do is check that what is being billed for has an infrastructure record and vice versa,” says Edwards. “In this way, we can increase the physical data accuracy of the local loops from present levels [50 per cent] up to 90 per cent. And this is done automatically and continually to make sure that the records are always up-todate. There is no manual inputting of information, which would increase the chances of error.”
If we take what Edwards says at face value, such an increase in network inventory accuracy could have far-reaching financial implications. According to Yankee, for every one per cent increase in accuracy of these records, the operator can generate a saving of US$0.30 per customer line per year. Assuming that the 40 per cent increase in accuracy can be achieved, that means a saving of US$12 per subscriber per year. And if you have 30 million lines, a nice sum of US$360 m is added to the bottom line.
“The broadband services that operators are now offering are low margin, so they have to do everything they can to lower their operational costs and satisfy their shareholders,” adds Edwards. Evolved Networks, which launched its commercial product last year, counts BT, Eircom, TeleDanmark and Jazztel (an altnet in Spain) among its customers.
Edwards doesn’t reveal any specific costs of deploying the Evolved Systems’ solution other than to say around US$4.00 per line for an operator the size of BT. “When you consider that it would cost around £1.6 bn [US$3 bn] for BT just to do an audit on its network, let alone try to fix it, then we represent good value,” he says. “We can give a payback on investment within 12 months and guarantee, as a bare minimum, that we will save one per cent of an operator’s annual turnover.”
To paraphrase an old English saying, where there’s dirt there’s money.
|
