Smart software helps to target brand pitch

By Helen Vause

As the advertising industry reaches the end of the century, finding ways to "do it better" has been a hot topic both locally and globally.

Creativity, accountability, getting to grips with the internet, technology and data-gathering are just some of the issues that have the industry's best brains in overdrive.

Doing the rounds in Auckland this week is the value of "neural networking," which involves getting greater marketing intelligence from data by programming a computer with software that allows it to "learn" from experience.

The particular neural network under the spotlight is the latest tool from McCann Erickson, which is another step from this international agency as it reappraises media and media strategy.

An overhauled media operation, Universal McCann, was launched recently to put far greater emphasis on media, new media tools, and to re-establish the media department as a key in-house function, contrary to agency practice worldwide.

McCann's new system is the baby of Sydney-based Spencer Tilley, who claims marketers are in danger of drowning in data and becoming intelligence-poor.

His new weaponry is sophisticated software designed to reassess the meaning of that data to make it more useful.

It is a system, says Tilley, that enables him and his team of specialists to consistently out-forecast brand managers working with leading international brands.

Marketers, he says, have been coming to wrong conclusions based on the way they assess factors that make the tills ring. He claims an astonishing 97 per cent accuracy rate in forecasting brand performance, or what advertisers get for their dollar.

Neural networking is not new but the application of it has come a long way since it was developed by a logistics whiz and a biochemist in 1947. It is a system that organises data by recognising patterns of information.

While proponents are not suggesting a computer can think yet, they say neural data gathering brings up new patterns in the marketplace and provides richer consumer data that can lead to new strategies and more intelligent decisionmaking.

For clients it generally gives greater insight into brands and a better feel of what their advertising is doing.

The next-generation analytical tool is a software system that recognises patterns to learn from recent performance history.

The neurons of the brain are simulated in software and joined together to produce powerful recognition capabilities.

lt figures out these patterns from data that would typically include pricing, advertising, promotion, trade deals done, seasonality, weather, awareness, distribution and stocks.

Tilley says that when all these factors are looked at together it throws new light on deciding where and when to invest advertising dollars for optimal return.

Often, he says, marketers have switched through various media plans but do not know which worked best because they cannot understand the overall effects of tinkering with various components.

The neural network "comes up with the sort of information where you can isolate and quantify the variables affecting sales, such as awareness, seasonality, the effects of core brand activity on other brands, and learn what factors are hindering sales response," Tilley says.

"It can give valuable information on strategies such as pricing deals, enabling advertisers to maximise the efficiency of these deals and work out when and by what magnitude to conduct a price deal or how to intelligently phase promotional activity to maximise sales returns. Often this means a better outcome for less spend.

"It is information that leads to a different approach with media spend by giving the ability to calculate the optimal spend in the right place at the right time - to fish where the fish are," he says.

"For advertisers who cannot afford saturation cover, this can lead to very significant savings on ad spend to achieve the same or better sales."

The neural network has also been used very effectively to create persuasive performance case studies for brands that have otherwise been unable to gain listing in major supermarkets, says Tilley.

Subscribers to the new advertising models resulting from neural networking include Unilever, Reckitt & Colman, Boots and L'Oreal.

Tilley claims the brands that have been driven by the neural system have achieved much better performance.

He says a British ice-cream manufacturer increased revenue by millions of dollars after reorganising promotions and media activity on the basis of a fresh insight into the impact of seasonality.

As neural networking gains acceptance as a hot forecasting tool, Tilley says it is also leading to a shift in thinking in some marketing behaviour. Frustrated marketers far removed from the northern hemisphere headquarters of many international brands would be pleased to hear that neural networking is challenging the wisdom behind the global marketing guidelines that often rigidly accompany brands in wildly diverse markets.

"Brands behave very differently in different markets. With the ability to pinpoint the influencing factors it is much easier to persuade head office that local facts can be much more successfully combined with local intelligence for the best results. It is nonsense to believe otherwise."

* This is the last advertising and marketing column this year. The next will be published on January 20.