Key to model behaviour

Improving the online customer experience is top of the agenda for many marketers – except, perhaps, for the 40 per cent who are already using behavioural targeting.

Speaking at a recent Alterian event, Rebecca Jennings, senior analyst at Forrester, said that this figure marks the (small) proportion of marketers who have adopted techniques to optimise their online marketing. And with the Internet Advertising Bureau putting digital marketing spend at over £2bn for 2006, there is clearly a need to ensure the money is delivering the maximum return.

The issue is how to achieve that. Web analytics is at a crossroads that could determine whether marketing spend continues to go into more than just paid-for search. To continue to lay claim to marketing budgets, the panoply of online ad inventory needs to be able to prove what it can do.
That is why one in five marketers plan to invest in better Web analysis and optimisation tools in the next 12 months, with a similar num­­ber planning to do so more than a year hence, according to Forrester. Technology and data must be better aligned to marketers’ needs to sustain the current levels of activity.

What is driving this investment? “Behavioural targeting is justified by 18 per cent on the basis of increasing site traffic and by 45 per cent through linking it to transactions,” said Jennings. “Some 20 per cent say it is too early to tell. And 49 per cent want measurement of the tools before they adopt them.”

Behavioural targeting is still in its infancy and there are relatively few success stories to speak of. For some marketers, there is still confusion about what it means.

Lucy Stafford, media director at Tri-Direct, says: “The type of behavioural knowledge you need is dictated by the indication of someone’s propensity to buy your product. This could be via a previous interaction with an ad or piece of content that flags up need.”

She believes historical insight into propensity towards a product is one of the most powerful findings. In offline marketing, this is usually built up within client organisations over many years and by commercial data owners through a variety of measures.

“But it can take several weeks to establish a consumer’s pattern of behaviour. It may de­pend on the urgency and flexibility of the campaign how practical that information would be,” says Stafford.

Between online publishers, ad networks and media buyers, there is debate on how that pattern of behaviour should be tracked. Most systems rely on cookies or capturing an IP address. Both have flaws. “Neither is too relevant or important. So long as you can identify the mass of your audience behaviourally to bring about an improvement in results, it is sufficient,” says Stafford.
Publishers face a crucial decision as to how far to develop behavioural targeting. They can already report on paths taken by unique visitors and segment their proposition according to popularity, length of stay and where visitors go next. Advertisers want a broader insight, however. What a consumer does on an individual site only allows them to optimise their expenditure on that site. As most campaigns spend across a range of opportunities, there is much value in a common currency that tracks behaviour more widely over the Internet.

Donald Hamilton, managing director of behavioural targeting provider Wunderloop, says: “Marketing at any level is about finding an audience that is most likely to buy your product. That doesn’t mean it has to be in the ‘right’ context.” Behavioural targeting has the advantage of being able to place ads in front of potential buyers when they are on indirectly-related sites.
This is especially valuable for portals and broad interest sites, such as online newspap­ers. While they may have specific content that advertisers can buy, they also have a wide variety of landing pages where an ad might equally trigger interest in a product or service.

“There is no point targeting pages if you don’t know who’s going to be there. You have to know who the audience is or you will only get a response rate of 0.001 per cent. In any other area of marketing, that would not be acceptable,” says Hamilton.

It might have been assumed that clickstream data and Web analysis tools would have generated better targeting facilities early on. Yet, as Hamilton says: “In the digital space, there are lots of people running around who don’t understand what marketing is about.”

While advertisers and online networks struggle over how to improve targeting, software developers are taking steps that could introduce a new era of individual-level behavioural targeting. Through a process of consolidation, Web analytics and offline business intelligence tools are converging.

Online optimisation software house Omniture acquired behavioural targeting provider Touch Clarity in March and is integrating the two propositions. “There is increasing interest not just in us, but in the whole Web analytics space,” says Neil Weston, senior vice-president and general manager at Omniture.

He believes existing Web analysis systems have offered weak data mining and simplistic reporting. “You can’t assess anything with them,” says Weston. He believes the fusion of business intelligence and Web analytics will deliver the “single version of the truth” about online activity that clients want.

Omniture already integrates data from 30 partners and aims to build that to over 100. Since each email broadcaster, ad serving network and Web publisher operates on a different data model, optimising multi-stream campaigns has previously been a nightmare.
“Touch Clarity is oriented to individual-level marketing. Where some people are segmenting users and offering demographics to marketers, behavioural targeting builds a knowledge base of every individual transaction the user has. It stores those and develops a profile of their characteristics,” says Weston.

Where this has been used to optimise online advertising, campaign performance has im­proved by 20 to 100 per cent, claims Weston. “Touch Clarity fills in one of the components of online optimisation,” he says.

Another acquisition has been by Visual Sciences of WebSideStory. The combination of data visualisation tools with Web analytics is intended to provide a new benchmark, not least in the ability to process the terabytes of data that clickstreams can rapidly generate.

“We’re providing a real time analytics platform that can scale up to the Web and is also usable for other channels. That is where we see the market going,” says Simon Gibbaud, business development manager at Visual Sciences.?“Multi-channel organisations are looking at the Web as strategic – most of their customers are now coming through the Web. They are using it as part of their core business. That places a burden on Web analysts,” he says.?Firms want answers to key questions, such as which page is most popular. But they want to drill into why that might be the case. “Web analytics has been very good at answering what happened. ‘Why’ is more difficult and questions are often unplanned,” says Gibbaud.

The volume of data generated by websites makes processing complex queries a challenge. Applications have used conventional relational databases to deliver table-based reports. Visual Sciences adopts a radically different approach, which yields more dynamic outputs of the sort marketers expect to see from offline data. “The interface is designed for business users, not hardcore analysts. It allows train of thought analysis,” says Gibbaud.

None of this technology is of value unless it is actively deployed by marketers. There needs to be a cultural shift that recognises online advertising as integral to the business.

Broadsystem chief executive Caroline Worboys says: “Customer data capture has never been so crucial. The insights drawn from customer data can shape an entire communications strategy and are key to its effectiveness.”

She adds: “Pushing the most relevant advertising and content to Web users is a giant leap forward and will serve to benefit our industry. But rather than pushing content, it is becoming increasingly important to allow customers to set their own agenda and pull the content they want. Not only is this true consumer choice, but also reaps even better returns.”?She cites The Times and Sunday Times weekly email newsletters. Content is based on recipients’ preferences.
Subscribers create their own profile via the Times website, ensuring the newsletter is as relevant and interesting as possible. Content is dynamic and evolves around how the customer behaves and interacts with the newspaper, both on and offline.

“This makes it possible for every email to be unique and personally tailored to meet the needs and wants of users. From the customer data gathered, it is possible to create a customer journey map and analyse the user’s behaviour to inform communication programmes, such as driving sales, engendering loyalty, or promoting a certain service and future acquisition activity,” says Worboys.

Few advertisers have integrated their digital activity in this way. To sustain the levels of investment in online advertising, many more will have to. In doing so, they will be demanding individual-level behavioural targeting.?Worboys summarises: “IP addresses and identified unique users are most certainly ­useful. However, there is nothing quite like individual data. It is this move towards individualisation that will deliver greater cut-through and results.”