Category Archives: ad measurement

Harry Potter’s intersecting media

So someone unearthed a scratchpaper grid Joanne Rowling used to plot the “J. K. Rowling” Harry Potter books. The columns of intersecting characters and themes remind us that all advertising components are interrelated. One fallacy direct-response marketers may fall into is to think they can isolate all communication elements — search, print, radio, TV, online display, mobile, social media, PR — from each other to track individual channel response. It’s a nice theory, but all ripples on the lake of consumer interest send waves elsewhere. Each media character has unique measurable value, but is also influenced by others. Just ask Harry.

Surviving the death of the click

Yikes. At first glance it looks like most consumers don’t give a damn about banner ads. ComScore is reporting that the number of U.S. internet users who click on display ads is down 50 percent. Back in July 2007, only 32% of all users clicked on internet banners in a given month; by March 2009, that number had plunged to only 16%, and 8% of U.S. users drove most (85%) of all clicks.

So other than knotting a noose, how should you, a marketer in charge of internet advertising, react?

1. Don’t be scared by concentrated results. Recognize that all response behavior follows power laws, the Pareto idea that a fraction of any resource drives the greatest return. You know, like the 10% of folks at your weekend party who get wildly drunk and make out on the deck while the rest stand in the kitchen sharing tips on accounting. Sure, in a given month only 1 out of 6 people may click on an ad, but that doesn’t mean the rest of the population never does. The concentration of clickers is a normal power cluster, just like those found in your stock portfolio.

2. Measure ahead of the click. As comScore suggests, clicks by themselves are a lousy metric for monitoring total online ad performance. Ad impressions build with reach and frequency and over time lead to action, usually in other channels. We’ve seen the inverse for clients who have launched campaigns on television and driven brand-specific searches through Google. A recent study of 8,824 respondents in the U.S., Brazil, Germany, Japan and the U.K. found TV had the most influence on audience purchasing decisions, even though it is typically one of the hardest media to track. Impressions here push action over there. Banner ad serving tools can track consumers who see your impression and then come in later via search engines. Social media monitoring tools can tell you the chatter among influencers. Find ways to measure the waves, not just the splash.

3. Watch impression costs. If you do begin evaluating internet ads on impressions and not just click performance (or downstream conversions, etc.), you must be careful with costs. Online ad inventory has exploded in recent years, led by social media pages whose users refresh frequently, artificially bloating inventory (think of a Facebook user rechecking her page every 60 seconds, each page with multiple ad slots). Careful online media planning can reduce costs against a given demo target by 90%, often by weighting the buy toward ad networks (collections of hundreds of sites) vs. individual marquee sites. A CPM of $1.50 instead of $30 against the exact same target should be a no-brainer.

4. Optimize, even if hard data is limited.
Clicks still matter, in terms of monitoring variances in performance between online media outlets and then optimizing the ad mix. Sure, only a fraction of your prospects might click on your ad, but if costs per click have a 20-to-1 skew across your online media plan, that’s an indication that some of the media venues are working well and others are bombing. Clicks and the associated downstream conversions, sales, and return on advertising can be monitored through each discrete online component to gradually improve the campaign performance — often freeing up 30% or 40% of your ad budget for redeployment.

Sure, not everyone walks through your online banner doorway; but evaluating the ones who do to rebalance the mix will provide lots of impressions where they count — on the CEO who watches your sales figures.

Image: Theogeo

Newspaper ad sales fall a whopping 29%

Yesterday the Newspaper Association of America released the worst news yet for U.S. publishers: Q2 newspaper ad sales were off $2.8 billion from the year prior, and ad revenue for the first six months of 2009 crashed 29 percent. Analyst Ken Doctor of Outsell believes only half this decline can be blamed on recession, and the reduced spending level is part of a permanent reset as marketers move budgets to the web.

It’s worth pausing to consider the cause-and-effect behind this trend: marketers spend ad budgets chasing results; fewer dollars flowing to newsprint mean results there are scarce; scarcity of results means consumers are reading less in newspapers, and even when they do they are less likely to respond to the ad. We still recommend newspapers in media plans for some demos, but the importance of measuring results with hard data feeds is now greater than ever. The old days of popping ads into newspapers with no tracking are now a wrap.

Image: Pensiero

Canadian Club loves your dad. But can you measure love?

Alan Wolk, one of our favorite pundits, notes ad campaigns are often held to unrealistic standards. “For years success and failure have been finite notions in the ad business,” he writes. “An ad either worked or it didn’t. A campaign was a resounding success or a dismal failure…”

But is it logical to judge advertising so quickly?

Take whiskey. Canadian Club has been running a wonderful series of print ads that move beyond the product to the complex relationship that its consumer target, men in their 30s and 40s, have with their fathers. The gist is your dad had lots of whiskey-fueled adventures before you were born, so lad, pick up the booze your dad used. We don’t drink whiskey (well, unless we’re in Nashville), but this message has made us rethink it. We miss our dad. Canadian Club has created a clever emotional frisson.

Now imagine measuring the success of this campaign. You could track responses (difficult), variances in media performance (more difficult), consumer awareness (is the brand now more recognizable?), and of course product sales. You could even believe the wild claims of the agencies who create such stuff, as they build their typically inflated case studies to submit to an awards competition. But in a land of millions of product choices, chances are Canadian Club’s message will just lightly break through, as good as it is.

And that’s OK. In many ways advertising is now a required cost of business, a messaging platform to keep up with competitors. The days of positioning, where you could grab a unique rung in consumers’ heads, are dying as product choices overwhelm our mental inventory. The question now isn’t whether consumers rush to respond; it’s what would happen instead if you went invisible and consumers failed to consider you at all.

We’re not ordering whiskey tonight. We won’t jack up Canadian Club sales tomorrow. But next time we’re in Nashville, we may give dad’s drink a try.

$100k homes use more DVRs. Don’t touch that remote!

Speaking of video shifts, Mediaweek reports that more than half of U.S. homes taking in $100,000+ in income subscribe to time-shifting DVR devices: those black boxes that allow you to record a TV show, play it back later, and potentially skip over the commercials. The Mediaweek headline is a little misleading — it is not true that “more than half of $100k homes time shift” since we suspect many consumers have a hard time with the damn remote controls. But the trend among the most affluent consumers is definitely worth watching.

New York Times sells the cover

Call it a sign of the Times.

Today The New York Times began running full-page horizontal display ads on page A1. USA Today started the trend years ago when it stuck small display boxes at the top of inside sections for advertisers such as Northwest Airlines. Most editors have resisted this, especially on the front page, the most hallowed ground for top editorial stories. But with NYT revenue down 13.9 percent in November from the year prior, readers bailing and advertisers retrenching in the recession, it’s natural that NYT would consider selling more space.

This doesn’t always work so well. Front-page ad placement comes at a premium, and when our agency has tracked the actual responses from front-page banner ads for clients, we found that consumer responses often don’t keep up. Large papers typically require 13-, 26- or 52-week commitments for Page 1 display ads, making such visibility a bit of a risky proposition.

Any advertiser considering such placement would be well advised to install a measurement system to track responses from each individual ad. It’s the only way to evaluate whether what feels good — being next to the major news — ends up as a good customer report.

Roll over, Rick: Razorfish patents way to track viral success

Sure, Rick Astley’s Never Gonna Give You Up has turned into an online phenomena, viewed more than 12 million times. But how did it get there?

Razorfish has launched a new patented slice of code that can track the exact paths things take when they go viral online. Think of it as a GPS tracking system for your message, monitoring movement online from generation 1 to 2 to 35 and mapping the key social influencers within the web most helpful in distribution. For example, the recent JC Penney doghouse ad sent a jewelry promo via video to millions of guys who usually buy stupid presents for their wives or girlfriends. It was funny. We got it from our father-in-law and laughed. But how does JC Penney know how important we are in the chain or how many friends we sent it to ourselves?

The Razorfish Incrementing Action Tag solves this by tracing a family tree for every person beneath you, just like your future children, who downloads and passes along your communication gene. The tag uses an anonymous cookie on the computer and does not capture user personal information; but marketers can pinpoint clusters of behavior inside the viral dissemination — you know, the cool kid in Des Moines who at contact level 4 suddenly made your message reach 10 million people.

Based on this knowledge, you can then fire your marketing director, hire the cool kid from Iowa, and use him to advertise in the future for free. Or something like that. Anyway, Razorfish, nicely done.

Why newspaper banner ads took the first hit from recession

Here’s some math behind failing newspapers. Bloomberg reports online ad sales at newspaper web sites fell 3 percent in the third quarter. This is significant because overall online ad growth is supposed to slow but not reverse in the current recession. Why are newspapers now sucking wind online in what should remain a growth medium?

The problem is how newspapers price online banners. Would you rather pay $1.40 or $14 to get a consumer to visit your web site? Newspaper sites are often the most costly.

Newspapers (as well as other “marque” or highly ranked web sites) often fail to deliver results online because they charge on a CPM (cost per thousand impression) basis vs. CPC (cost per click). This is critically important for media planners to understand because a $20 CPM at an average 0.14% click-through rate works out to a $14.29 cost per click. Which, as we noted above, is 10 times higher than the CPCs at ad networks.

Ad networks, or groups of hundreds of sites, can provide the same or better consumer targeting and often charge an extremely low cost per click of $1.40 or less. The irony is that many ad networks can place your ads on the remnant inventory of main marque sites for one-tenth the cost.

Sure, marque sites have value for branding or reaching specific audience targets. But ad nets and retargeting now offer the same at a fraction of the cost. In general, costly CPM buys, no matter how “strong” the main web site brand, are no way to blow your budget in a down economy.

Photo: Network Osaka.

Why most media kits stink

So what are you getting for all those big ad dollars, anyway?

All the advertising you see on TV, radio, papers or online is carefully planned as a war game, and one key step in that conception is to review media outlets by their “media kits.” These 4-10 page documents are the personal résumés of magazines, TV networks or newspapers. As you’d expect, they talk about content, who reads them, and what ads cost.

For decades media kits bragged about “impressions” — how many people they reach. And then a funny thing happened. Internet advertising was born and started talking about results.

The table above is just one illustration from ValueClick, an online ad network that collects 13,500 web sites and runs banner ads across them with behavioral targeting or retargeting. Media planners who work with such ad networks can predict, in detail, not only impressions but clicks through to the site, conversion rates, leads and sales.

This type of information is incredibly useful in media planning, yet old-school media avoids it … and in fact has become defensive about what data it does reveal. Newspapers have begun trying to mask circulation declines with newfangled metrics that combine paper readers with web traffic. Radio networks and industry groups dispute new Portable People Meters that monitor an exact signal from radio stations to define exact audiences. Essence magazine claims more than 7.8 million “readers” with a print circulation just over 1 million, in a mental trick known as “pass-along readership.” Uh-huh.

It’s too bad old media is running and hiding. If instead they helped advertisers forecast real results, they might win a bigger slice of the ad pie.

Our inability to see trends

Business managers usually have two requests: forecast and then measure results. Marketing directors and media planners happily comply. Trouble is, the measurement approach usually looks at tiny intervals of time — a few weeks, a month, a quarter at best. Quant jocks are like ants crawling through the cracks of massive momentum, trying to predict inclines without seeing the real hills ahead.

What macro trends are you missing? Are you really paying attention?