The online data game

Marketing Direct, May 2001

Gathering data online is getting trickier as consumers wise up to the value of the information they divulge as well as the ease with which they can opt-out of contact Suzy Bashford reports.

Most of the world (excluding direct marketers, of course) is just cortoning on to the fact that data is valuable. And so a mad rush has ensued to collect any pieces of information about consumers in cyberspace as if they were nuggets of gold.

Points schemes, lotteries, questionnaires and competitions have all hit the net in droves in pursuit of data. Some offer straightforward incentives while others are able to create sophisticated profiles of their registered users down to what cereal they eat or what credit card offer they would be likely to respond to.

Bearing this in mind, you'd expect targeting of customers online to be top notch. You'd hope that when you opened your inbox exciting and relevant offers would be awaiting your perusal - hand picked by your favourite brands.

More often than not, when people access their email account it is full to the brim with irrelevant and instantly deletable commercial messages from companies. Spammers are likely to be flogging their wares; usually porn, fake degrees and credit deals.

If the average inbox reflected a typical British consumer then our country would be crawling with sex-crazed, uneducated people with no cash. With all this data being gathered, what happened to the good old-fashioned DM skills of segmentation and targeting?

This month's Think Tank panel sheds a great deal of light on the situation. The most surprising explanation was that clients are turning down offers of highly-segmented customer information and opting to play a numbers game.

Donald Hamilton is sales director at Bananalotto, an online lottery which pledges to give away £1 million a day if customers pick the right numbers and fill in a few details as they do so. Bananalotto now has over 548,000 registered users as well as background information to make any online advertising relevant to each user. But, Hamilton says: "95 per cent of clients we offer segmented profiles to, don't want them."

This is backed up by Wanda Goldwag, executive chair of competing site The Daily Draw (part of PDV) and founder of gay and lesbian portal Queer-company. "We are working with advertisers who think targeting is picking men or women," she says. This is echoed again by David Brosse, marketing director at MyPoints Europe, an online rewards programme that also collects data in return for points. He says: "Advertisers come to us saying 'we don't have a profile'."

These comments bring a whole new dimension to the opt-in/opt-out row that is currently raging between the DMA and the recently formed Email Marketing Association (EMA). The DMA is lobbying for an opt-out ruling because it believes opt-in discriminates against direct marketing specialists who, theoretically, only send targeted and relevant emails. According to Brosse, this is not the case. "The DMA is trying to differentiate between targeted direct marketing emails and spam and, looking at advertisers behaviour, there isn't much to differentiate the two."

Have marketers learnt nothing from traditional marketing about the value of targeting? According to Dominic Owens, head of marketing at office services provider Groupetrade.com, this behaviour partly comes down to the cost savings that the internet offers for sending out messages compared to the expense of direct mail. "It is a wonderful irony that the industry has created a basis whereby you have all this data on consumers but, because it is free to send the email out, this is actually going to spoil the whole thing," he says.

Many companies are hell-bent on destroying the fledgling email industry, content to send out blanket emails rather than using data to tailor offers to individuals. "People that do volume emails know from direct mail that they'll make money. And there is no doubt that a company who has no concern whatsoever with privacy and collects data and sells it randomly, can also make this model work financially. There are very few sanctions against them," says Goldwag.

And it seems to be this pressure to 'make money' that is spurring many business on. The recent economic downturn has compounded the problem. "Dotcoms not only have to think about branding, driving traffic to the site and fulfilling orders but they have to survive financially," says Nick Turner, marketing director at online data capture specialists Mandoforms.

Commercial pressure

This commercial pressure appears to be hindering the sophisticated growth of the online data market. "Venture capitalists say 'sweat that data or you're dead'. It's not a theoretical problem about not using databases properly, it's a commercial problem that companies are being told to sell everything they can and they're desperate," adds Owens.

There is no doubt that many of these 'desperate' dotcoms will soon go out of business. But, instead of jumping on the bandwagon and slagging the dotcom entrepreneurs, the panel are refreshingly sympathetic. Blame for their demise partly lies with the business people who gave them funding. "You would never start a traditional company with the business plans that were being used for websites," says Joy Taylor, head of e-business at Lowe Live.

The panel resists taking the moral high ground when it comes to selling data in desperation but all agree there is a huge problem with its valuation. Many companies now selling data did not set out with this as a revenue-generating objective, so it is not their area of expertise. The knock-on effect is poor data quality. "I got offered a list of 70 million names for $200 the other day. And I thought to myself, this just can't be quality, can it?" says Goldwag. Her view is that the online list selling market should be killed before it can grow but this opinion is not shared all around the table.

Data without buying

Instead of buying data to find customers, many traditional brands are side-stepping the list buying market and going online with their own data. This is not straightforward either, especially for large, unwieldy organisations such as banks and airlines. Taylor and Goldwag both used to work at British Airways. "We had 24 databases not including holidays and Airmiles. A consumer in any given week could communicate several times with the brand but the company would have no idea it was talking to the same person," says Taylor. Nodding on agreement, Goldwag adds: "And they replicated this online."

BA has since undertaken a project called Ocean Wave to incorporate all the databases into a single warehouse, according to Taylor. But there are still many big brands grappling with the problem of disparate databases and this does not bode well for embarking on a 'relationship' with a consumer online. The panel agrees that over-communication online risks alienating the customer for good.

So if most traditional companies are struggling with legacy systems - is anyone successfully using data gathered online? Predictably, one of the few names the panel came up with was Amazon. However, this compliment was gradually retracted as the panel agreed that the online bookseller is actually doing basic direct marketing with no frills.

Goldwag points out that she was initially impressed with Amazon's recommendations to her based on her previous purchases - until she was repeatedly made irrelevant offers relating to a present she had bought for a friend (a crucial titbit of information Amazon had neglected to gather).

In many cases, companies do not have the right systems in place to carry out sophisticated data mining - Amazon included, according to Hamilton. "Amazon has a standard system. They haven't got guys sitting there with a regression and chaid table saying 'oooh, can you see the correlation there?"

Multi-channel contact

Generally, large companies do not seem capable of communicating on an individual level with customers at the moment. This is where smaller, niche players have an advantage. For Owens' company "it's not all about the numbers". After a transaction, customers are contacted by phone for feedback and this data, he says, is more valuable than any gathered online.

For larger companies, however, gathering data online is the only way of understanding customers better. But capturing information is riddled with problems. To start with, many consumers nowadays have multiple email addresses - some opening accounts specifically for commercial messages.

Taylor is one of these consumers. "I purposely have five different emails. And, on average, one in ten sites contact me with something that may be relevant," she says. If companies are directed to this purpose-built email account, the challenge is to 'trade-up' into the email address that is most valued. For Taylor, this is her corporate address.

So how can companies make their messages stand out in the crowded inbox and progress to 'corporate' status? Taylor thinks it's a case of good 'section management' - which she describes as allowing the consumer to self-segment by choosing a series of options. But it is then vital that this data is used to create a relevant communication in the future.

All too often, questions are asked that do not seem to have the slightest effect on the communication. Taylor gives the example of a girl who requested information to be sent to her about holidays in Scotland. Not surprisingly, she became irritated when her inbox filled-up with offers of holidays in the Caribbean, Spain and practically everywhere except Scotland. So, like many others, she hit the 'unsubscribe' button.

Turner endorses this view of 'section management', especially when consumers are asked to fill in longer questionnaires. Mandoforms specialises in creating 'intelligent' forms for clients such as Barclays, so consumers are only asked relevant questions. "Imagine a mortgage application form which is 157 questions. If you're single, divorced or widowed or whatever you only really need to answer 40. You must lead the consumer down a track which is relevant to them and hide all the rest, " he says.

The panel accepts that the theory is slightly different for prize draws and lotteries because consumers, as Goldwag says, "understand there is a bargain going on". Because of this they are likely to give companies more leeway in terms of relevance of messages. But how much?

Many schemes argue that they can create a relationship between an advertiser and a consumer. Brosse says MyPoints can do this by emailing its customer base alerting them to advertiser offers.

However, some of the panel think this is going too far. "The customer may have said it wanted a relationship with company X but what they get is marketing from company Y, albeit in a communication from X," says Goldwag.

Whatever approach companies choose, it will only be successful if the consumer is completely clear about how their data will be used. 'Trust' repeatedly crops up as an essential element of good online data capture. But don't think this can be instilled by whacking a logo from an accreditation body on your website. According to Brosse's research, this has no effect on how much information a consumer gives.

Respect the consumer

Talking to them in a way they feel comfortable with will help. An example is the address they submit. Some online questionnaires use the Postal Address File (PAF) to validate addresses and will not accept one that does not exactly match its database. "My PAF details are wrong. I live in a house and it says the basement is a different flat. So some companies won't process my request," says Owens.

Telling a consumer where they live is not advisable and for this reason Experian (according to Brosse who was its former head of new media products) uses PAF only to validate post codes. Then, if a consumer feels happier living 'near Edinburgh' rather than Dalkeith, communications will respect this.

The panel is full of ideas on how to fine-tune data capture online but the reality is that suppliers and agencies are streets ahead of clients. At the moment, clients are focused on getting a message into an inbox by hook or by crook.