A question of quality

A new audit scheme from the DMA aims to clean up data across the direct marketing industry writes David Murphy.

Next to the people, data, they say, is a company's most valuable asset, and any company with a well-maintained, clean customer database will know just how valuable it is.

But all companies routinely need to look beyond their traditional customers to prospect for new ones, and when they do; they usually rely on a list, bought direct from a list owner, or via a list broker or manager. But just how good are these lists?

List quality is a long-standing issue for this industry. Out of date, inaccurate data leads to diminished response rates which in turn leads to unhappy clients laying the blame at the feet of the list seller. Then there's the issue of data collection. Is it an opted-in list through a genuinely open and transparent collection process? How recently was the data collected?

In the business to business world, data quality is an even bigger issue. Conduit has just published its fifth annual audit on data decay revealing that data decays at a rate of 27 per cent with businesses wasing £220 million each year on inaccurate mailings. Over six per cent of business have moved in the last year and 19 per cent of managing directors have moved positions.

Into this less-than-perfect world comes The Direct Marketing Association (DMA), which is attempting to tackle some of these issues with the launch of the Business List Audit scheme in October. Business list owners who wish to be accredited for the scheme will fill in an online questionnaire on the Business List Audit website (www.dmabusinesslistaudit.org.uk). The questionnaire asks for details of how they process, clean, verify and update their data. Those lists which adhere to strict standards of quality and best practice will be granted accreditation and the right to use the DMA Business List Audit logo.

In addition, the DMA will employ a compliance manager, who will be empowered to remove or suspend companies' accreditation if they fail to justify the information they provide in order to obtain accreditation.

Finally, the DMA will be introducing a Customer Satisfaction Survey to gauge what list users think of data quality and hygiene. Consultations with advertisers, list owners and list bureaux will be used to establish a performance benchmark, after which the surveys will form part of the compliance process, administered by the DMA's new Best Practice and Compliance business unit. Companies whose business lists fail to meet the established performance benchmarks will have their accreditation removed and suspended.

DMA director of postal affairs and industry development David Robottom says the organisation expects to achieve 14 per cent market penetration (around 40 companies) in year one, 30 per cent by the end of year two, and 50 per cent by the end of year three.

"We've had an excellent reaction to this scheme so far," says Robottom. "Previous list audits in the b2b marketplace have not been successful, but the use of a Compliance Manager means that the DMA's Business List Audit will have real teeth."

With the Business List Audit ready to roll-out, Robottom says the DMA is now starting work on a b2c version, which should be ready in a year's time.

Given the obvious problems with bought-in lists, it would seem that it can't come soon enough. Yet John Regan, managing director of data company Cognisance believes the situation is not as bad as it may seem.

"People not involved in the business might be surprised to hear that this is how things are, but it is better than nothing" he says. "A bought-in list will never be as good as customer data, but it will give an uplift. It does actually work."

It many indeed work but according to Peter Kempsey, managing director of data solutions company Intellidata, list owners could be doing a lot more to ensure that their data works even harder. B2C list owners rarely, if ever, clean their own lists, says Kempsey, relying instead on the client who buys the list to put it through a data bureau for cleansing and suppression.

This means that although the client pays for a complete list of x number of records, he is likely to end up with much less once it has been verified against the main suppression files. Kempsey cites a recent example of six bought-in lists totalling 6 million records that Intellidata was asked to clean on behalf of a client. Of those six million, 27,482 were Mailing Preference Service (MPS) registered and 79,605 deceased. Less than 2% of the total perhaps, but still well over 100,000 names.

"Very few list owners even use the MPS" he says. "They rely on the bureaux to do all the cleaning. I think lots of people have tried to put the onus on the list owners in the past but have got nowhere with it. Even major mailers seem to be resigned to the fact that data tends to be what you get."

According to Regan one of the problems lies with the list buyers. "Buyers may have a degree of expertise in targeting, but they don't always have expertise in the technical aspects of data" he says. "In many cases they don't know which questions to ask to ensure that they are being given value for money."

Regan advises buyers to ask questions that ascertain if previous use is being assessed at individual or household level; and on whether the recency of the data is judged on when the data was loaded onto the database or when the information was actually collected. In some instances, says Regan, there could be as much as six months difference between the two dates. Also, are there default answers, which are assumed when a question, such as age, has not been filled in?

"Lots of people don't like to fill in their age on surveys and on lots of lists, a default answer will be entered where no answer has been given" says Regan. "If the default answer for no age is 99 and if a client wants a list including everyone aged over 50, then all the people who did not enter their age will be included on that selection."

Clients can avoid some of the problems by asking for variables to be supplied with the data, so in the age scenario, the client would see not just the names of the "over-50s" but their ages too, and could identify that there was a high number of apparent 99 year olds.

If b2c list quality could be better, things seem equally bad in the b2b field. B2b data company Conduit has a core list of the top 150,000 limited and public companies, which it combines with a list sourced from Thomson Directories to create a total of two million records, which can be "diced and sliced" in a variety of ways to create a Fleet database, IT and Telecoms database etc.

Conduit managing director Mike Grey says its core list of 150,000 companies is telephone validated at least every 12 months, but he says this is not typical list owner behaviour. "It's expensive" says Grey. "If you outsource it, it can cost anywhere up to pound a call, so most list owners don't bother. Our approach has always been to try to establish a reputation at the quality end of the market, but a lot of people just throw the mud against the wall and see how much sticks."

Simon Foster, managing director of media buyer PHD Confidential is openly suspicious of the quality of b2b lists. "We see it as our job to be sceptical about media owners" he says. "Often, the list is just a by-product of something else, just a means to make money. Once we have established a supplier or a list that we know is good quality, then maybe we can develop it in a more positive way but in general, we have to be sceptical."

The DMA's new list audit, it seems, couldn't come soon enough.

Cleaning a really useful list

List broker HLB has been managing the Really Useful Theatres mailing list for three months. The list comprises around 100,000 records, people who have ordered tickets for productions at Really Useful Theatres venues, either via a hard copy order form or via the telephone.

Before it began marketing the list, HLB's first quality check was to ensure that the correct Data Protection question was being asked on the order forms and the telephone booking scripts.

Next, the list was sent to a bureau for PAF validation to check the accuracy of all addresses and ensure that all had a postcode attached to them.

Next, the list was deduped at address, surname and individual level. Where the same address appears twice with different names (Mr and Mrs for example), renters are offered the option of mailing one or both individuals at that address.

The list was then run against a number of suppression files, namely the MPS file, the Mortalities Register, and a file called GU Swear, which flags records which appear to contain an expletive or where the initial and surname together form an expletive.

HLB business development manager Dominic Murdoch says that these checks will be run again every three months to ensure the list is up-to-date and accurate.

"The list has to be screened for MPS and mortality every quarter under DMA rules, but we would always recommend to clients that they screen against everything else too" says Murdoch.

So far, says Murdoch, the list, with a bias towards older, more affluent consumers, is proving very popular with renters ranging from Internet Service Providers to charities.

If you have an opinion on this or any other issue raised on Brand Republic, join the debate in the Forum here.