Merging And Matching

The story of UK consumer data supply has traditionally been one of innovation. Whether building and maintaining niche lists, introducing complex stats to build modelled Electoral Roll-based files or offering third-party-supplied names with transactional data attached, this market rarely stands still.

Head to head?
The latest entrants to consumer data sourcing have continued to seek out new sources. The Trading Floor has made insurance application data from the web available for marketing use, while companies like Transactis or Abacus with its Publishing File have brought data that would traditionally be restricted to closed co-ops out to a wider market.

The availability of transactional information on real buying behaviour plus proprietary UK-wide merger databases are the two biggest innovations of the last few years. Now these files are maturing and, with the grater volume and selectivity, are ready to take on lifestyle.

“The pools created by people like Alchemy and Transactis are emerging as alternatives to lifestyle databases,” says Michael Smith, senior account manager at Prospect Swetenhams. “You get plenty of selectability plus the transactional data shows what people really buy. With credit data and many charities contributing, there’s some really interesting stuff on there. It’s used across the board and works well in most sectors.”

Many UK marketers still stay loyal to one data supplier, often choosing to simply licence a large chunk of lifestyle data. Matching the variables to customers, running profiles and then constructing ongoing campaigns is much more straightforward when you only have to deal with one type of data. So are these pools or “open co-ops” really a valid alternative to lifestyle data?

“The majority of proprietary pools still do not remotely have the depth or the selectability or lifestyle,” says Richard Webster, group communications director at DLG. “The source of contributed transactional data also tends to be anonymised so it’s hard to test one source against another. With lifestyle, you can test on 400 separate variables. And there’s also sponsored client questions that simply aren’t possible with a third-party co-op or pool.”

Traditionally, there has been skew on both lifestyle and transactional data. The former because surveys have historically been more attractive to older women, the latter because they are composed of mail order buyers who also tend to be older women: great if you are targeting though the post, otherwise perhaps not so desirable. There’s also a question mark over lifestyle’s reliability though this can easily be addressing by testing.

“You do get a really skewed sample of the population on lifestyle,” says Zoe Vine, head of data services at The Trading Floor. “They tend to be female and over 40, or students. Also, people filling in insurance applications tend to tell the truth.”

“Because of multi-channel collection, the female skew simply isn’t true anymore,” argues Webster. “We take a third from the net and gather the majority of the rest over the phone.”

In fact, with multisourcing and merging now accepted as the best way to get optimal prospect coverage and depth, it’s becoming apparent that lifestyle and transactional data are far more effective working in harmony than posing as alternatives to reach other, whether the merging work is done by the data end user or offered as a proprietary service.

The success of this approach was borne out in the classic 2005 study by consultancy Tank where a multi-sourced file built from two Electoral Rolls (ER) and a third source outperformed individual Rolls on both coverage and predicted response. Specifically, coverage was up 25 per cent with the merged file and 6.63 per cent of its top decile were predicted to respond, the highest figure of all files tested.

Chasing transactional
Collecting data via transactional means in some form and adding it to their base file is becoming very desirable for many data owners. It allows targeting based on actual behaviour rather than what has been claimed in a survey and the continuing customer relationship also serves to periodically validate the data. It’s also a lot cheaper to collect and collate than traditional or even online surveys.

“It’s becoming less and less viable to manufacture data with surveys,” says Chris Morris, managing director of Transactis.

With much of the UK’s data becoming ubiquitous due to multiple licensing deals, merging and modelling work is becoming more of a differentiator in campaign performance, and being able to offer transactional data is very desirable. But it’s not all that easy to get hold of – there are only so many high-volume UK mailers. As in the suppression market, stringing deals to source these files from home shopping or travel companies is conducted in great secrecy in case the competition get wind of it and try and corner that feed for themselves.

For example, Transactis doesn’t pass on transactional indicators to third parties, partly due to the fact that much of the company’s data is only permissioned for analysis and enhancement rather than for list usage. “We keep these for ourselves,” says Morris.

DLG has already moved some way toward this third-party approach by signing up Friends Reunited as a contributor. Though this is not transactional data, customer details are continuously reverified as registered users enter the site.

We’re coming from the other side,” says Webster. “We’re looking to draw other data onto lifestyle as part of our consumer data hub. We’d like to p0artner or acquire a significant owner of transactional data.”

The drive to build multisourced data products is far from new. Companies like Experian have long since used a mixture of their own and external data to build geodemographic segmentations while other suppliers happily employ data from third parties to fill in the gaps in their national coverage. For example, Transactis supplies AcXiom with data to help fill the gaps in the ER base file it used for its modelled Infobase Lifestyle Universe product.

For its part, EuroDirect’s Data Exchange has to be the granddaddy of the open consumer prospect pools and the company could fairly claim to have been well ahead of the game when it launched it almost a decade ago. Claimed to be the UK’s largest prospect pool, it is built from a mixture of every type of consumer data on the market: compiled response lists, credit, lifestyle, contributed transactional information, public register and more.

“We have transactional data that is typically sourced from mail order,” says EuroDirect’s database solutions director Tim Pottinger. He says that the transactional data is more typically used as an indicator rather than being used at a more detailed level to drive offers.

“We use that information at individual level to model out to the rest of the population,” he says. “If someone is mail-order responsive then we’ll use that as an indicator of channel preference. It’s our intentional to gather more transactional data and put it into Data Exchange at a lower level with greater details.”

Beyond the data owners, data-independent companies have been building merged UK files from license third-party data for some years now, with products like CACI’s Ocean and Alchemy’s AMS pool bringing many of the benefits of multisourcing to straightforward “per-thousand” rentals. But to realise the most from multisourcing, some form of data set bespoked to your own requirements is the way to go, bringing together the most relevant files that may not be available in “pre-merged” form.

For the largest mailers, building their own pool in this way has long been the most controllable and efficient way to run prospecting. Over time, more and more marketers have started to take this route, attracted by the results it offers and also by the savings to be made by efficient long-term usage. Pay-as-you-go deals, introduced by Experian and now taken up by a number of other suppliers, have helped ease the upfront costs, but financing a bespoke database build is never cheap.

Accessible pools
Now more suppliers are coming up with a lower-cost route into merged prospect pools by offering their clients access to their master pre-merged database and then adding bespoke data on top rather than starting each project from the ground-up each time. Ai was one of the first to go down this route.

“We have a large merged database as standard,” says managing director Jon Cano-Lopez. “We can build client-specific views based on which supplier data they want to buy. There’s not much transactional data in there as standard, but we will add anything a client wants us to add that is available.”

The company has recently added credit data and also offers Pipeline, a web-based pay-as-you-go list service that pulls records from the same central database.

“This gives access to a multi-sourced pool for smaller users,” says Cano-Lopez, who notes that each client’s view of th4e data in Pipeline has to be set up in Advance depending on how many sources they want access to.

More data owners are now starting to do something very similar to the data-independent Ai, building out from their base data sets by adding in third-party data from other owners and offering them as a richer base for a bespoke pool (and spinning out subsets of the data as list rental sources too). Here, multi-sourcing improves selectability, reduces skew in the original data and crucially makes much large volumes available to sell to their clients.

Transactis has used CACI’s Ocean as the base file for its consumer data from its earliest days and adds in Equifax credit data for pre-scoring prospects, plus Council Tax Band information to help with household and neighbourhood- level targeting.

It labels its Vision service as an “off-the-shelf solution that includes every key data component used in today’s targeting arena”.

“Prospect pools aren’t just a solution for the big mailers anymore,” says Morris. “Prohibitive de-duplication costs from bureaux and poor access to their data are holding companies back. The pre-built element of our solution lowers the cost and we’d add any data set the client wants to mix.”

The Trading Floor has hung its data on a third-party data skeleton since day one (CACI’s Ocean again). With around 8.5 million of its own records, it is now adding more data to the mix to improve volume and details. “We can have fantastic information for one transaction, but might need more behavioural or life stage detail,” says Vine. “We partnered to bolster these aspects. Now we’re adding EuroDirect’s Cameo Financial overlays and merge in more insurance date renewals from external sources.”

With its pre-merged file as a base, the company has recently launched a database management division offering bespoke pool building and online access to hosted prospect databases. However, simply gaining volumes is not what it’s about.

“Anyone can build a UK reference file,” says Vine. “We’re more interested in accuracy. It’s not so much about volumes, it’s the value of the records that are in there.”

The increasing number of merged files and pooling services can make it very hard to tell the difference between the competing suppliers. More and more databases are on offer with most records being the same data license from the same original owners. And where you are dealing with the data owner, there will inevitably be pressure on sales staff to sell their own data rather than that of a third party simply because of the grater margins.

As to how any pool is built and run, any supplier mentioned here should be happy to do it all for you including hosting, or to work with your incumbent data processor if they are doing some or more of the work. With the newest data pools offering cost reductions over bespoke builds due to their pre-merging work, how the lowest “total cost of ownership” matching up with performance during testing looks like the equation to solve. More and more vendors are offering packaged deals rather than volume-based pricing, and the cost-per-acquisition route to pricing we mentioned by all commentators as an option.

Merging the most
So rather than the latest data sources taking over from lifestyle as top dog, the direction is the other way: merging lifestyle data with everything else and seeking out ways to tap transactional data sources that haven’t already been ring-fenced by competing suppliers. A series of “super-merged” UK universes have emerged that, when combined with some specialist files, have performance comparable to the best bespoke prospect pool.

We can see these rich pools becoming the base for a new range of products and services. For example, EuroDirect has long since used Data Exchange as a base for its classifications and now has an alerting service based on new incoming data. If an individual’s income or address changes, they are flagged up for potential contact.

“There’s no getting away from testing in direct marketing, but the combination of transactional, Electoral Roll, credit and lifestyle data has to be the way to go,” says Morris. “Now it’s coming down to two key sources: lifestyle and transactional.

(Database Marketing - Dec 2007)