The man once labelled the nation's most reviled marketer has given attendees at his business seminars the opposite advice to the practice recommended by legislators of the Spam Act.

Perth businessman Wayne Mansfield has been the target of many Australians' frustrations with spam since late 2002 when he tried to sue an antispam activist for allegedly blacklisting his company's IP address. (He told attendees at a recent seminar that according to one media report: "I am Australia's most reviled marketer".)

Mansfield has maintained the activist's accusations of spamming were mistaken, but conceded that before the privacy legislation of recent years, his marketing company mailed to "whatever [e-mail addesses] we could find".

Those practices were abandoned three years ago, according to Mansfield. Since then he has used e-mail only (no Web site) to promote his business seminars. Attendees last year numbered 20,000, he said, as well as a raft of tertiary education customers for his eMailMagic (marketing) seminar.

"We provide enough information to build an effective e-mail marketing initiative in accordance with the law," he said.

"The people from the universities that have come are trying to attract more students but comply with the rules."

Mansfield, however, has given attendees advice different to that offered by the Australian Communications Authority (ACA) on what legal experts say is a grey area of the legislation.

Compliance with the Spam Act should become clearer for e-mail marketers in the next few months, once the Australian Direct Marketing Association issues its eMarketing Code.

Mansfield is not a member of the association. In order to keep mailing to e-mail addresses collected before the Spam Act, he has advised clients to send an opt-out e-mail as a means of gaining inferred consent to send future e-mails.

He said he had advice from a major law firm that for addresses collected before the Act, an addressee's failure to unsubscribe from a mailing list, having received a number of e-mails with an unsubscribe function, could constitute conduct which implies inferred consent.

According to the ACA, inferred consent requires an active step by the recipient.

The regulator has recommended an opt-in practice since the Act came into effect, arguing a user's refusal to respond does not infer consent.

"It is our view that consent to receive future messages cannot be inferred from the mere failure of the recipient of an unsolicited commercial electronic message to unsubscribe," ACA executive manager John Haydon said in a recent notice to marketers.

Although supportive of the legislation and much of the ACA's guidance, Mansfield said his business, and others, could lose years of work of careful database building if opt-in was enforced.

"We only market to business. And if someone's sending you something that could help your business, you generally accept it," he said.

"The problem is we've got these gurus of doom in the media who are advising people not to unsubscribe to e-mails," Mansfield said, arguing the opt-in regime placed an unfair burden on marketers.

Mansfield cited a couple of attendees at one of his recent seminars who had been on his mailing list but unresponsive for two years before registering to attend [a seminar].

The issue was further complicated by how long an inferred business relationship might last, according to legal experts.

"The ACA position is very reasonable. The legislation is so imprecise and vague - that's what's causing the problem," said Clayton Utz partner Peter Knight.

"It puts good people in an impossible situation. Do you trash five years of work?"

"Most people are not getting the two-step process. They think 'I have existing relations, I can send'. You also have to have consent.

"Marketers have argued 'what about prospective or past customers?' These things have to be wrong," said Knight, adding that a sale a few years ago may not justify advertising a product/service to the same person today.

However, Duncan Giles, special counsel at Freehills, said neither opt-in or opt-out was the issue.

"The issue is trying to get evidence of consent.

"Opt-out is one way of evidence of consent, but it's not as good as opt-in."