±±¾©Èü³µpk10

It's not hard to imagine direct marketers being enthusiastic about the brilliance of their latest marketing strategy or the creative genius behind the execution. But when did a direct marketer last become enthusiastic about the impeccable cleanliness of the company's database?

Do they sit around the table with their agency and data experts arguing which file should sit at the top of the suppression hierarchy; whether a postal return represents a genuine goneaway or the pros and cons of fuzzy matching?

Benefits of clean data

The answer is probably not. Data hygiene doesn't tend to inspire passionate discourse, unless you're a suppression provider defending your patch in the face of a rival offering, but it does fall squarely into the category of 'boring, but important'. Boring because it's process driven and can be time consuming, important because its impact on the brand, response rates and ROI is considerable.

Given the well documented benefits of good-quality data, not to mention the DMA's drive to improve industry standards in line with its Government targets, it's surprising that so many campaigns still include dirty data.

Research by EuroDirect found that nearly a third of unsolicited direct mail in the UK contains errors in the recipient's name or address details. Charities were the worst offenders, with 77 per cent of mailings containing inaccuracies, followed by the FMCG and automotive sectors with 65 per cent.

The message still isn't getting through, but the industry needn't hang its head in shame. First, the UK tops the table for suppression uptake, according to figures from Experian. Second, keeping data up-to-date, clean and suppressed is no quick and easy task.

Chris Leahy is assistant manager at Lloyds TSB's wholesale data management team, responsible for managing, cleaning and modelling data on businesses. "Business data is very challenging," he says. "Companies have multiple addresses, multiple respondents and different internal departments."

Front-end accuracy

Keeping abreast of vital information such as office moves and staff changes is an ongoing task. The first and most important step is to ensure that the data is clean and accurate at the point of capture. Even Leahy says that the data-entry checking that takes place at the front end is "not as robust as we would like it to be." He adds: "It's something we're looking to improve."

The good news is that there is no shortage of products to take companies through every step of the process from correct data capture, through to the various suppression products.

There have been many new entrants in this market in the last year including disConnect from Equifax, a file of goneaways, deceased and complainers, and Acxiom's Purity File which has more than 13 million records of people who have been verified as having left their addresses.

In April a new suppression file will be available from The REaD Group, following its acquisition of My Right To Be Private, a pay-as-you-go suppression service. This will contain information on consumers who don't want to receive certain direct mail.

Help or hindrance?

Is this abundance of suppression files a help or a hindrance?

Data-cleaning business UK Changes has already identified five key goneaway suppression products in its benchmarking exercise (see box, page 8). Chris Duncan, managing director of database services company Alchemetrics, says there is a downside to the different offerings. "Once you start flooding the market with different suppression products there can be massive overlaps. If someone dies, for example, this may be picked up in different places and is duplicated across files."

Companies could end up paying twice for the same person to be flagged up as a goneaway or a deceased. Cost-effective suppression can quickly stray into more complex areas, which is one of the reasons why many firms turn to dedicated data bureaus, which establish what file best suits their clients' suppression needs.

For example, it may be more cost effective for companies to pay an annual licence fee for a suppression file, rather than paying a royalty of about 20p to 40p based on every match. This is often the case for large volume mailers, which are likely to generate many hits during the year. "In effect, this turns suppression into a fixed cost, rather than a variable one. Companies can budget for it as a fixed monthly charge, which is something we recommend," says Duncan.

Having paid for use of the file for a fixed period, it makes sense to use this first in the hierarchy of different products, filling in the gaps with other files on a per-hit basis.

It's all about achieving the highest quality of data at the most cost-effective price, which is one reason why so many companies try before they buy. Most bureaus offer a free data audit process. Having analysed the database and established the cost and impact of a thorough data-cleaning programme, clients can decide if they want to proceed.

Online solutions can provide an even quicker way for firms to assess their data. For example, Experian's Intact data-cleaning solution and UK Changes' service Direct Connect Online, both have free data audits.

Companies can upload their data and within minutes establish how clean it is and the cost of suppressing and cleaning records.

There's really no excuse for firms not to investigate the cost of cleaning their data.

A quick calculation, which compares the cost of this exercise with wasted mail pack costs, will invariably favour data cleaning.

COMMENT

GLYN DAVIES, managing director of computing services, Lloyd James

"Confidence in the reliability and source of each suppression file is critical to hierarchy building. A few years ago, there were just a few suppression products on the market, but now there are many. The best files are compiled from verifiable sources."

STEVE COOK, managing director, Market Location

"Ideally, sales people within B2B companies should be trained in database management, so their customer records are kept up to date. This is better than simply relying on mailing returns."

CHRIS CUFFE, managing director, helpIT systems

"The key is to invest time in understanding and selecting the most effective and appropriate solution for your circumstances and to trial against a subset of your data."

COLM O'HARA, database manager, EuroDirect

"For value-for-money data-cleaning, the first step must be to ensure the database is de-duped. If the same person features twice on a database, then they will be flagged twice and you will pay twice. So ensure the database is as clean and compressed as possible in the first place."

TOP TIPS FOR DATA SUPPRESSION

1. Be aware of permission issues - it is essential that you only target those who have given you permission to do so.

2. Data audit - take time to analyse the accuracy of your data, including quality of address records from the outset.

3. Do your research - it has become increasingly difficult to keep abreast of developments in the range of suppression products available. Enure you use the right product for your needs or take advice from a bureau.

4. In some instances it may not be appropriate to suppress. Like targeting, you need to identify the areas of your customer prospect base that are most important to you and treat them accordingly.

5. Make sure you know what you want to suppress against - these requirements will change depending on the campaign or marketing needs.

6. Think about matching routines - there can be many variations on how the name and address data is held for the same person, which can cause doubt over whether the records can be matched.

7. What kind of suppression data is contained on the suppression file you want to use? Different suppression files will hold differing levels of information. Some files may contain only surname information and some contain both individual and surname level data.

8. Be clear on costs and licensing - be sure you know what products you need and how often you will use them. Suppression products are priced based on a single use or on a permanent use basis through an annual licence agreement.

9. Don't forget business data - the UK has a dynamic business environment and business records change very quickly.

10. Once is not enough. You should clean your data at least every month or every time you run a campaign.

Source: Mark Roy, chief executive of The REaD Group

THE BIG ISSUE - DATA CLEANING HIERARCHIES

Creating a hierarchy of files is a well established way of cleaning data in the most efficient and cost effective manner possible. By starting the cleaning process with the files that are the most effective for the data in question - according to cost, coverage, quality or a combination of these - the optimum number of records are flagged early on. Further down the hierarchy, alternative files can be used to plug gaps in the data if necessary.

There are no hard and fast rules about where the various suppression files should sit in the hierarchy, as this depends on the data and the client's objectives. For example, if cost is the over-riding factor, then files which carry a lower royalty cost per match may be used early on.

However, there are some files which should be used to begin with. The first is a company's own in-house suppression files, as these cost nothing to use. Most organisations will have compiled a database of goneaways, through postal returns, and complainers who have asked not to be contacted.

The next file in the hierarchy tends to be the MPS. "Most suppression hierarchies will start with the MPS because the cost is low," says Glyn Davies, managing director of computing services at Lloyd James. "As a data-management company, we pay an annual licence to MPS and then levy a modest processing charge for clients.

If the client is paying an annual licence to use a particular suppression file, it would always make sense to use this next as they will not accrue any additional cost per match.

The pecking order hereafter is up for grabs and data management professionals will have their own views on which files best suit their clients' needs.

Accuracy, quality and market coverage are the three criteria they will tend to look at. Last year, UK Changes' benchmarking exercise scored the various goneaway and deceased files against these criteria. Its findings on six goneaway suppression products are shown below.

Suppression file Accuracy Data Market TOTAL score quality coverage National change of address file 60.0 19 14.0 93.0 Goneaway suppression file 57.0 19 15.5 91.5 National suppression file confirmed 56.0 19 11.0 86.0 Universal suppression service 37.0 19 10.5 66.5 National suppression file complete 26.8 19 15.5 61.3 Absolute movers 20.0 19 5.0 44.0 Source: UK Changes CASE STUDY Brief: Clean customer data for Ramada Jarvis Hotels' Scottish Escapes campaign Target Audience: People interested in visiting Scotland Supplier: Alchemetrics

Hotel chain Ramada Jarvis faced one main issue when it came to cleaning its customer data in preparation for its Scottish Escapes campaign. Ramada Jarvis was looking to attract customers to its offer of a two-night break at one of its eight hotels in Scotland.

The database consisted of data taken from individual hotel reservation systems. Customer details were taken from handwritten registration cards and some of the records dated back to 2002 or earlier, which meant the quality of data was not always very high.

The first step for the company's data partner Alchemetrics was to conduct a data-quality audit to provide Ramada Jarvis with a snapshot of the quality of the data.

Alchemetrics' internally developed applications cleaned the records where bad characters appeared in inappropriate fields and highlighted the presence of business records.

The next step was to run the data against Royal Mail's Postal Address File to enhance records to Royal Mail standards. It was then screened against the MPS, National Change of Address File, Bereavement Register, Mortascreen and GAS files, in ascending order of royalty costs.

At each stage of this process Ramada Jarvis received data-quality reports, highlighting specific areas of improvement that could be incorporated into the data-capture process.

The Scottish Escapes mailing achieved a considerable reduction in returns compared with previous campaigns at less than one per cent. A response rate of eight per cent was achieved contributing to a 40 per cent increase in revenue stream.

"It's essential for us to maintain focus and understanding of the implications of poor data quality," says Sharon Kalsi, leisure marketing manager at Ramada Jarvis. "We have seen a huge improvement in the accuracy of the data we use for direct marketing and an extremely low amount of returns from mailings."