Monday, July 27, 2009

overpromise of metrics

There's a new app that purports to analyze your headline. Just feed in your line (20 words or less) and the analyzer digests it, swirls it around and spits out numbers that show how well it will resonate with consumers.

If you make ads for a living, or know people who do, you probably agree that this idea is ludicrous. Too much about writing a headline is immeasurable. How it plays off the visual, for starters. "Lemon" wouldn't have worked nearly so well if it had appeared for Sunkist, over a shot of a lemon.

The headline analyzer is, to me, a perfect example of the overpromise of metrics.

Metrics has become a big part of the ad business lately because so much of the business is shifting online. The overall budget for digital is predicted to double in five years while that for  traditional (you guessed it) is predicted to flatten. Part of the reason for this shift is metrics. "You can track it!" can sound awfully persuasive to brand managers wondering (as the saying goes) which half of their ad budget they're throwing away.

Ad Contrarian crankily observes that advertising now consists of two very different disciplines: (1) making ads and (2) making justification for ads. Ironically, the latter may ultimately prove more remunerative. A torrent of analytic apps are now available and many more are in beta, measuring not only rudimentary CTR and conversion rates but dwell time, trending, sentiment, chatter, shareability, influence. Sure, those analytics tell us something. But what they tell us depends largely on what we want to hear. Numbers can be made to say almost anything, as anyone who's been audited by the IRS knows. 

That's not to say that metrics aren't valuable. For one thing, they're often still key in selling clients on the idea of doing digital advertising, just as response rates once convinced them of the worth of doing direct. What analytics turns up can be endlessly fascinating. But, curious, isn't it, how often stats prove irrelevant, failing to influence marketing behavior. Managers can be quick to cite data when it supports their thinking; just as quick, if it doesn't, to explain it away. 

Which, imo, is the way it should be. Advertising is, and always will be, more art than science. Because its success depends on human persuasion, which has required skills of creativity since Eve sold the apple. Data can help point up what you should say, but how you should say it (and where and when) is an art. If advertising were a science, every widget-maker with a headline analyzer would be a global success.


Teenie said...

One of our clients drew up a score sheet for focus groups, and insisted the winning concept would have to score above 7 to proceed. Same client them tested the finished shoot and couldn't understand why the scores were lower by a 1-point-something.

As if there was a winning formula for concepts! Yeesh--if there was we'd all be rich. Or unemployed.

Susan Ellis said...

Here in Canada, there's a wonderful radio show on CBC each week, called The Age of Persuasion with Terry O'Reilly..
It's all about advertising and persuasion -clients, consumers, copywriters...great show

Tom said...

I just scored 100% EMV on the headline analyzer... I put in.

"Beautiful beautiful beautiful in every respect"

Something tells me the thing needs fine tuning.

Marilou said...

Very wonderful i like the best

M.M. McDermott said...

"Fart like you mean it" scored a solid 40% with an emotional classification split between intellectual/spiritual.

"We bring good things to life" scored 33%.

I don't know what you're all talking about - this magi-o-matic headline analyzator is unadulterated genius.

Ad Broad, oldest working writer in advertising said...

@Teenie Advertising by CYA metrics. Bah!

@Susan Ellis Thanks for tuning me in.

@Tom Hilarious!! One more "beautiful" and it prob would have broken the damn thing...

@Marilou :)

@M.M. McDermott Guess GE stock wouldn't be in the tank if they'd gone with the first line, instead ;)

Mak said...

I work with websites/databases, and let me just say I love developing metrics tools. It's art to me.

I've also come to realize most of the marketing reports are not very realistic indicators of trends. For example, for certain products we get random spikes in traffic for a month, but those do not translate to sales or even inquiries for them.

I think it just gives marketing people stuff to do :P

Steffan Postaer said...

The "Headline Analyzer" is far better satire than tool, like something you'd see on Conan.

California Girl said...

Of course I'm going to have to try this thing for kicks. Not that I'm a great headline writer but, after 30 years of radio copywriting and now I'm blogging, it's certainly a practiced art.

We're all going to be replaced by machines one day anyway, right?

Ad Broad, oldest working writer in advertising said...

@Makolyte Yes, my point exactly. Metrics can be fun...but at the end of the day, not sure how much influence data actually wields.

@Steffan Postaer When headline analyzer was first sent to me, I assumed it *was* satire, couldn't believe it earnestly existed.

@California Girl And speaking of being replaced by machines, have you seen

California Girl said...

July 31

Ad: No, I hadn't seen the article so thank you. Scary stuff.

I referenced machines because I've been haunted by Stephen Vincent Benet's "Nightmare Number Three" ever since college.

This excerpt from

Nightmare Number Three envisioned a bloody, black humor revolt by all of mankind’s machinery, with the nameless narrator huddled in mad fear, speculating on how things could have gone so wrong (”letting six million people live in a town./ I guess it was that. I guess they got tired of us./ And the whole smell of human hands.”).

And now it's coming true...

Mak said...

RE: Machines taking over the world

Haha! Let's just hope the machine overlords are Robin Williams 'Robots'-ish and not Terminator-ish!

Mak said...

Oops! *Robin Williams 'Bicentennial Man'ish

wrong movie!