Online advertising in politics

Cookie targeting and digital ads.

Based on the hype we hear, you can run your campaign with no volunteers and a staff of only one: a clever twentysomething armed with a Macbook and the candidate's credit card. You've probably heard the pitch: He can develop full profiles on everyone in your voter database, based on their browsing history and Facebook data, which will enable you the candidate to reach tens of thousands of voters in the most effective manner ever imagined, by displaying your ads and videos on their browser pages and Facebook views.

Or so we're told.

The technology certainly exists (and demonstrates itself every time an ad for tennis shoes appears on your screen after you've shopped for a pair online).

Here's how it works

As we've seen, many websites won't work properly unless you enable "cookies", which are digital signatures containing your internet address (which can approximate your location) and your browsing history (and sometimes the personal information you've entered) on the site. They say it's to "enhance your browsing experience" but we all know it's sold to marketers (on the internet they're known as "ad exchanges") who aggregate the "cookies" left by you and millions of others. These compilations are sold to advertisers who would like to reach people based on their location and preferences; for example, everyone in western Pennsylvania who's shopped for Steelers gear. Based on the billions spent on online advertising, more than a few people think it works.

Advertisers pay between $2.50 and $3 per CPM, which means "cost per mille", or thousand, page views. In other words, every time your ad appears on someone's browser page, you're paying about a quarter of a cent. This cost can escalate rapidly when you consider the hundreds of thousands or "cookie" profiles matched up with your county or district.

When you hear the word "targeted" in a sentence with "online advertising", think of it as a shotgun blast instead of a rifle shot: you'll probably hit the target but with lots more lead. And you'll hit things that were outside your aim.

Like most other forms of advertising, the hype falls short of the reality. That's why we place online advertising in the same category with transit ads, newspaper inserts, billboards and local Cable TV: They can be effective at reinforcing your brand, and in some cases can be helpful as get-out-the vote reminders. So they should be part of your media "mix".

Some things to think about

They're targeted but still not personal. At least when someone sees a lawn sign in the neighborhood, it carries the implicit endorsement of the homeowner.

Fraud is rampant. About 40 percent of clicks and page views are assumed to be generated by "botnets" and "click farms" designed to generate revenue for the ad exchanges and website owners; savvy advertisers simply "price in" the cost of this, and you should too.

Ad-blocking software is on the rise. A recent industry report estimates that use of ad-blocking software increased 48 percent in 2015 over 2014.

The true cost and reach is unknown. If your campaign is well-crafted you will indeed narrow your focus to those who are receptive to your message. But how many? And how many times did they even see see your quarter-cent ad? We read one study that says half of the ads don't even appear because of the thousands of variant screen sizes and configurations.

Online advertising isn't cheap. When the candidate's banner appears on a billboard or a city bus, the cost is fixed. In the digital world, the cost is ongoing: you pay for "page views" (every time it appears on someone's browser page) and "click-throughs" (whenever someone actually clicks on the ad to view another linked site, presumably the campaign website.) These costs can grow. Rapidly. Exponentially.

Internet addresses are inexact. They're similar to cable-TV location "nodes". And they don't resolve the issue of mobile devices and laptops, which move around. Which means the "footprint" of your advertising might miss significant parts of your district, or conversely leave you advertising to tens of thousands nonvoters. If you're told your ad will reach everyone in the county who browses the internet, the statement might be true but it might require reaching half the computers in four or five surrounding counties, as well as thousands of mobile devices that were just passing through when their location was recorded.

All of this makes online advertising difficult to quantify, and difficult to determine if it's the right choice for your campaign. (Our advice: give it a try but stick to what works, at least this year. Let someone else do the experimenting.)