A MULTISOURCED prospect pool takes list-based business-to-business targeting to the next stage of sophistication and effectiveness. Rather than treating the United Kingdom market as one with limitless potential, it helps identify the finite number of prospects likely to buy from your company and supports efforts to reach them.
Just as in business-to-consumer marketing, a pool allows your firm to create the clearest, richest picture of its market, and to save money on data purchases by negotiating terms in advance. As a structured repository for data collected in house via customer contact and research, or for data brought in from third parties, response analysis or modeling initiatives, the pool provides a solid base for ongoing targeting.
“Not every company will buy from you, no matter what you do, so analysis of existing customers helps us to define which companies you can sell to,” says Simon Lawrence, joint managing director of U.K.-based B-to-B direct marketing firm Information Arts. “Further insight work and segmentation will help [a company] understand which groups of customers buy when, and how much they buy. This insight can be used to massive advantage, and all that's learned can be retained.”
A useful way to estimate future return on investment is to work with a supplier to build a sample from your company's customer base and examine it to see what potential exists in the market. The amount of work needed also may lead to the classic in-house/outsource debate. Many suppliers or specialized consultancies can host larger marketing databases — for example, D&B with its Market Insight software, or Information Arts with its Webview system. Larger companies most likely can devote resources to building a prospect database, but the benefits apply equally to smaller operators in more specialized markets.
“There's no lower limit on volume for prospect pools,” says Verran Townsend, research and development director at call center LBM in Belfast, Northern Ireland. “You may only have hundreds of prospects, but you should have the most comprehensive and accurate coverage of your target market.”
The initial requirements are the same as in list-based targeting: a merged, clean, single view of current customers and contacts from which to derive a picture of your company's best accounts. Project work will focus on identifying the types of firms that should make up the prospect universe, at first employing standard business demographics, then adding a host of other possible indicators.
It's essential to get the structure of any new database right. Which fields will be needed and, if targeting larger outfits, how will the relationship between contact, job title, department, company and parent-firm structure be handled? When applied to customer and prospect databases, corporate data from suppliers like D&B will help estimate the actual overall levels of business your firm gets from a particular customer, and not just a company's different business units. This may reveal the need to deal with a central purchasing department rather than several contacts at the same company.
Working in tandem, the customer database and the prospect pool probably will be merged on the same platform, with the existing customer base a subset of the whole customer universe. They both must be populated consistently with any extra variables so that insight from one can feed the other directly.
“They are talked about as separate things but it's usually the same platform and system,” says Nicky Keyworth, director of sales and marketing solutions at D&B. “That way, you can track the whole journey from prospect to lead to customer.”
As information from multiple feeds is formulated and put in place, the supplier and client have to decide on the final database's structure well in advance of delving deep into its construction. Changing fundamentals when a project is under way could bring about costly revisions.
Working out which external or internal data sets, records or fields are the most accurate, and which source will take precedence over the other, is all part of this work. The lack of a date stamp on many B-to-B lists can be frustrating here. The database also will have to cope with future expansion into new markets and be flexible enough to handle all possible data types your company might want to append.
As only 300,000 or so firms file their accounts at Companies House (the official government register of U.K. businesses), there are, of course, gaping holes in coverage of certain key variables among the unregistered. So it may be necessary to create models to fill any gaps where such information is missing from both customer and prospect records — another advantage of the merged database — or conduct market research when this isn't practical.
Some of the major U.K. list companies have tried to fill the gaps in standard demographics using a mix of telephone research, modeling or input from other data sources. Experian's National Business Database is a good example, where volume sources such as Yellow Pages and Thomson Directories are combined with modeling work to provide U.K.-wide coverage of business variables such as net profit and number of employees.
But to obtain the clearest picture of the whole U.K. B-to-B market, some kind of merging work will have to be done to bring together the largest files. Well-known data suppliers like Experian and Prospect Swetenhams are active in the United Kingdom but data-independent agencies such as Information Arts and Marketing Improvement also are prominent.
As well as supplying marketing consultancy and modeling services — often in conjunction with third parties — independent agencies can act on behalf of clients to strike long-term licensing deals for large amounts of data. This is similar to what's happening in the consumer sector with suppliers like U.K. customer intelligence specialist ClarityBlue, though volume discounts will apply only to the largest B-to-B purchasers.
Examination of customer behavior or value can go well beyond simple descriptive profiling, employing multivariate modeling and classic stats techniques such as regression analysis. This helps to identify the combinations of variables that make up the picture of your best customers and prospects, and the relative importance of each one, which may differ depending on the sector, the proposition and so forth. Within each segment or sub-segment, large companies might build product-specific predictive models to help hone their proposition. In fact, a variety of hierarchical segmentations can apply.
For example, the top-level segmentation might outline the buying behavior of different parts of the market, describing channel choice, proposition and creative. Modeling those companies in each segment with the highest propensity to respond will allow targets to be prioritized. Within this model, a second value-based segmentation could help decide how much to invest in contact and again reveal which prospects should be approached first.
“By breaking down the prospect universe by different variables, you can find out how to market most effectively using a combination of two or more segmentations,” says Information Arts' Lawrence.
JAMES LAWSON is the editor of Database Marketing magazine (www.dmarket.co.uk) in Nottingham, England.
FedEx Delivers on Its Database
Before starting work on a customer database and a merged European data pool, FedEx was throwing away information on more than 85% of its telemarketing prospects, and some 25% of the addresses on its customer file were undeliverable.
Working with B-to-B DM company Information Arts, FedEx pounded its customer data into shape and populated a new standalone database with a “master” and “slave” data structure that could take account of corporate linkages. It also introduced strict in-house data-entry processes, and sought and gained buy-in to the database from important future users such as the field sales teams that quickly realized its potential benefits.
Multiple external data sets were used to enhance the base data with standard fields such as SIC, turnover and company size, while two years of aggregated customer transactions showing value and volume by time, service and route provided the evidence of different customers' buying behavior in a variety of sectors and locations. A mix of Experian and D&B data was used to flag a business as an importer or exporter and helped identify that firm's level of international shipping activity.
“Our old view was ‘Keep mailing and we'll keep getting customers,’ ” says Jeremy Elder, FedEx's marketing manager for Northern Europe. “Now, with our current service offer, we have a limited market to go for. We know how many are out there, their value and roughly which destinations they should be shipping to. We can look at the market demand model and say, ‘We only have 30% of that customer.’”
Prospect identification is now so good that FedEx no longer pre-qualifies using external telemarketing, and this in turn saves quite a bit of money. Other savings have resulted from using its central database for all European marketing. Station-location planning and drive-route optimization have helped in this area as well.
“It's a simple model but complex, too, because of the amount of data involved. We've made the base as wide as possible and have invested in continual data improvement and enhancement,” Elder says. “We should be referring to it for any strategic decision.”
— J.L.
Powergen Energizes Data Efforts
Powergen Retail, part of the E.ON Group, sells gas and electricity to one in four small to medium-size United Kingdom businesses. It recently started working with Experian to revamp its 500,000-record customer database. Initial work involved matching to the National Business Database to clean up contact details and append extra demographic fields, including the Commercial Mosaic postcode-based classification codes. Using profile information from this exercise, a separate prospect database was created which is matched monthly to the customer file to dupe out existing customers.
“We use the number of employees and number of sites as a [substitute] for energy consumption,” says Powergen's CRM manager Mark Perrett. “We also use SIC and the individual classification codes from Thomson and Yell. We don't go after everybody. Above a certain energy consumption, customers are handled by our account teams.” As well as modeling out from existing information, Powergen uses its call center to fill gaps in the company's prospect and customer databases.
“We enhance our database where possible, asking for consumption figures, their existing supplier, whether or not they're on contract and the expiration date, and the name of the decision-maker,” Perrett adds. Using the Ensemble campaign-management package to handle customer and prospect data, expiration dates are used to set the timing for future calls.
“Finding the right person to contact is still one of the major challenges,” Perrett says. “If you have the prospect file in house you can tweak it more easily and get the matching right. That's hard to do with B-to-B data, and we have quite complex matching processes. Matching across multiple sites for the same legal entity is a challenge, too. It can be confusing.”
Powergen also has built propensity models to supply information on channel choice and proposition. The phone is by far the most effective channel.
“It's hard to get to the right person with direct mail,” Perrett says. “Timing is very difficult to do, and few people will buy from you during the first call. We also look for retention indicators, but that's difficult to do with contracts. Getting prospect data that sales advisers can trust is a big challenge. I think the accuracy of our field modeling is about as good as we can get it now.”
— J.L.
