On television, advertisers pay based on the ratings of a particular commercial. On the Internet, they pay according to a measurement like a click on an ad. And in magazines, advertisers choose where to run ads based on the results of door-to-door surveys done twice a year.
    While other media add technology, magazines rely on a technique favored by Civil War-era salesmen. And they believe that is hurting them. Their share of the money spent on advertising fell by 0.6 percentage points last year, to 11.5 percent, and is expected to decline faster than any other medium except newspapers, according to media agency ZenithOptimedia.
    "There's definitely a measurement concern," said Brenda White, senior vice president and publishing activation director at the media-buying agency Starcom USA, part of the Starcom MediaVest Group division of the Publicis Groupe.
    Accustomed to results from Internet marketing, advertisers are pushing magazine publishers and ratings services to refresh their approaches. And the magazine world is responding.
    Two of the main research houses, Mediamark Research and Intelligence, known as MRI, and Affinity, which runs the Vista service, are updating their offerings. MRI, which already measures a magazine's readership, plans to introduce studies soon that examine how many people saw a certain ad. Vista already measures magazine ads, and will add a readership survey so it can estimate its ad viewing results for the entire American population.
    "It's not enough anymore to say, 'Hey, we hit the right people, give us a pat on the back,'" said Steven Kotok, the publisher of The Week magazine. "The Internet's raised the bar."
    Currently, how much an advertiser pays is tied to a magazine's rate base, the number of guaranteed subscribers and newsstand sales. But to figure out which magazines to advertise in, advertisers rely on MRI's research. It measures how many readers a magazine has, including people who did not buy it but read a friend's copy or flipped through it at the doctor's office. It also profiles the readers of all the magazines, including their income levels, attitudes and toothpaste-buying habits.
    MRI divides the country into representative neighborhoods and sends researchers to the zones to conduct a 45-minute interview, with 26,000 people a year, asking them to remember which magazines they have read in the last six months.
     The researchers leave behind a 104-page survey about what sort of television shows people watch, what kind of products they use, and what social or behavioral traits describe them. MRI then tries to adjust its results so they are representative of the whole United States.
    The $34 billion in magazine advertising is allocated based largely on demographics and MRI readership numbers -- the estimate of how many people read, not just bought, a certain magazine. "They are the currency for the magazine publishing industry," said Betsy Frank, chief research and insights officer for Time Inc.
    Although everyone relies on MRI, publishers, advertisers and even MRI itself say they recognize problems with the current approach.
    Scott McDonald, senior vice president for market research at Conde Nast, said that although MRI's sampling methods were "scrupulous," he had a few misgivings over the service. MRI asked people to remember which magazines they had read over a few months, and people have unreliable memories, he said. Another problem, he said, is that while readership numbers for popular magazines tend to be steady, small magazines bounce around quite a bit based on a handful of respondents.
     And MRI numbers do not always reflect other measurements. In its most recent statement, MRI measured Elle Decor's readership as increasing by almost 50 percent from a year earlier. But in its most recent filing with the Audit Bureau of Circulations, Elle Decor reported its subscriptions and newsstand sales had increased less than 1 percent over a year. And while MRI said Tennis magazine's readership dropped almost by a third, its subscriptions and newsstand sales rose slightly.
    With MRI, "we've all had situations where one year you're up 20 percent, and one year you're down 20, but nothing changed on your mix, nothing changed on your file," said Keith Fox, the president of BusinessWeek. "I think it's pretty fair to say that it's a blunt instrument."
    MRI executives defended its methodology, but conceded certain problems with the door-to-door approach.
    "While there are very big advantages to doing this face-to-face interview, there are some very clear disadvantages," said Kathi Love, MRI's chief executive. "Not only are we not as timely as the publishers like, we don't get ad-specific measurement," she said.
    In late 2007, to address the timeliness question, MRI introduced measurement of individual magazine issues beginning a few weeks after the issue came out. But this was based on an online panel, a methodology some statistical experts oppose because it underrepresents groups that are not active on the Internet. Love agreed that there were weaknesses with online panels but said MRI weighted the data to try to reflect a magazine's national readership profile.
    Affinity has ramped up its Vista research service, which started in 2005. Vista asks online panelists which magazine ads they saw, and whether they bought anything as a result. Because it relies on an online panel, it raised the same methodology questions as the MRI issue-specific measurement. And Vista's results were limited to its sample. It could not apply those results to the whole United States.
    Now, MRI and Vista are becoming more competitive. Next year, Vista plans to introduce the American Magazine Study, a Web-based survey that will let Vista add readership data to its ad ratings and estimate for the whole population.
    And MRI plans to make public its own ad-measurement service in mid-June. Like Vista's, it will be based on an online panel, and MRI plans to weight the ad readership figures to the national figures, so advertisers can estimate how many Americans saw and reacted to one of their magazine ads. "I'm really hoping these metrics help magazines get their fair share," Love said. Some publishing executives said this could have a significant effect. "Time Inc. feels that this is a game changer," Frank said. "This will validate magazines as the medium that is the most accountable."
    That is a tough claim to substantiate, up against online media. And some publishing executives said that the MRI issue-by-issue data was not correlated with other numbers like newsstand sales.
     One publishing executive said he worried that the issue-by-issue data varied because of statistical quirks, not any actual changes -- but apparent dips in how many people saw an ad could lead advertisers to immediately remove ads.
    Even so, White of Starcom said, it beats the old way. "This is progress," she said. "We know it's not perfection." (NYT)