Ask the experts - B2B match-making is heaven

Because you had a universal database in B2C called the Electoral Roll, you could be confident of achieving a representative sample of matches. In B2B, however, there is no universal dataset.

This is a problem because B2B data is complex. Businesses are multi-sited and multi-functional. Take Tesco. The superstore has a hundred different trading names and 2,500 locations in the UK, as well as sub-brands such as One Stop Shop and Dillons Newsagents. It owns farms and could have as many as 300 different suppliers for one product area.

If a client can't get the required number of matches between his or her database and a composite, off-the-shelf business database, the segmentation will be flawed and he or she could be losing revenue.

All B2B databases are biased by their method of data collection. So, for instance, if you use a telephone directory as your base, you will only get sites that your target company wants you to call.

The solution? You need access to a multi-sourced database against which to match your database, exceptional matching capabilities in your database software and an understanding of the structures of business and the abbreviations commonly used.

CHRIS DICKENS REPLIES:

Suppliers have been aggregating consumer lists such as lifestyle surveys and the Electoral Roll for years, but there is no equivalent in B2B. A utility or phone company wanting to increase its penetration among businesses will need a big database that doesn't exist.

A single, generic solution is not the answer. Client companies should co-operate to pool data with non-competing partners. This is the way forward but you need to choose your partners carefully - it can be a client-led, rather than broker-led, initiative, with the help of their agencies.

MARTIN DOYLE REPLIES:

I'm afraid B2B record matching is more of an art than a science. There is no silver bullet.

Matching has many levels, starting with matching the profile of the perfect customer using variables such as turnover, employees, and so on. When you've bought your perfect data records of prospects in companies such as Xerox and Tesco, because they fit your perfect profile, you need to ask whether they match the 'Zerox' and 'Texco' records you have stored in your marketing databases. Or will you be introducing duplicates?

But it's the rules you apply to matching records that matter. For example, what do you do when two records match, yet one says opted-in, the other says opted-out? Which do you take: the newer one, the older one, the safer one, or just ignore it? This is a simple example of a complex issue for B2B marketers.

POWER POINTS

- B2B matching is more of an art than a science

- B2B databases are biased by their method of data collection

THE PANELLISTS

The practitioner

- Iain Lovett is executive chairman of B2B specialist Blue Sheep the agency

- Chris Dickens is head of strategy at Geronimo, The software developer

- Martin Doyle is chief executive officer of DQ Global.