If you’re curious about what I do for a living, or about internet marketing, you might find the following post interesting.
Companies love internet marketing – search engine ads, banner ads, e-mail newsletters, etc – because they can track it with greater clarity than any other form of advertising. Advertisers like Nielsen and ClearChannel may provide estimates of how many users see a TV ad or a billboard. And those estimates might even be accurate to a statistically significant degree. But they can’t say with real certainty how many people follow up with the sponsor because of that ad. They can guess, and they’ve grown really good at guessing. They can post a special phone number or URL that only people who’ve seen that ad would use. But it’s not precise.
Internet marketing stomps on that limitation. We can tell where you were when you saw an ad. We can tell when you clicked on it. If the ad popped up in a search engine (a Google ad, a Yahoo! ad, a Bing ad), we can tell what you were searching for when the ad surfaced. If you click on the ad, visit the sponsoring website and don’t buy, we can tell which other sites you visit afterward. All of this, and probably more, we know. Not about your demographic, or about your personality type – about you, personally, the man or woman doing the browsing.
(Clearing your cookies prevents this, though technology to get around that will soon be widely available)
Very few agencies manage all of a company’s advertising, however. A company might do SEO in-house, contract out for search ads, buy banner ads direct from the publisher, use e-mail marketing templates and trackers from a small agency, and get one overworked intern to handle social media. So you have data streams pouring in from half a dozen different sources. But these data don’t all represent separate people. A lot of these data points might cover the same person.
For instance: you’re thinking about buying a new car. You do a Google search for “used sedans in Boston.” You see an ad in the Sponsored Links column (on the right) for certified pre-owned Hondas at a dealer in Quincy, MA. The “Honda” grabs your attention, subconsciously, because you’ve been seeing a lot of banner ads for a Honda year-end sales event. So you click on the dealership ad and poke around at some of the models they list. You visit several other sites that same day, either by searching or by clicking through sponsored links on trusted sites (like the Boston.com/Cars page).
A few days later, you see a banner ad for the Honda dealership in Quincy you clicked on first. This isn’t a coincidence. You’ve been retargeted – cookied and triggered to serve banners from a site you visited but did not buy or sign up from. The ad mentions one of the models you were looking at (the Accord, let’s say) and says there are some new cars in stock. You click through and look around. Sure enough, there’s an Accord that meets your needs. You fill out the online form to get more information.
A detailed quote is sent to you from the dealership. This puts you on an e-mail list. A week later, the list sends you its weekly e-mail, listing a handful of used cars that have just arrived in the showroom. You click through on one, decide you like it, and call the showroom to arrange an appointment. The customer service rep asks where you heard about the car. From the e-mail list, you say.
For that one transaction – getting you into the showroom – you’ve touched on the following marketing vertices:
- Banner ads (Honda year-end sales event)
- Search ads
- Placement-targeted links (from Boston.com/Cars)
- E-mail marketing
Five different forms of marketing, advertising or lead generation. The ultimate goal here is to get you into the showroom so the car salespeople can go to work on you. Which ad gets the credit?
The e-mail marketing? That was the last thing you touched before calling customer service. Technically you could have gone to the showroom at any time, but you weren’t compelled to until you saw an e-mail you liked. But you never would have signed up for the e-mail if you hadn’t revisited the site and filled out the form. And you never would have revisited the site if you hadn’t been retargeted – meaning you needed to visit once and get cookied. And so on, and so on, and scooby dooby doo.
Does the last touch get credit? The first touch? Do we evenly distribute credit among all marketing efforts? Marketing departments need to know this in order to know who gets more budget next year. And if they don’t get a compelling answer, they’ll guess.
In the above example, the e-mail marketing clickthrough was the only behavior that got you to sign up for an appointment. Through a glass darkly, the first banner ad looks useless – you didn’t even click on that! No revenue generated by the banner ad, so cut display advertising for next year. But without that Honda year-end sales event banner ad, you wouldn’t have clicked on the Google ad. Fewer display impressions means less brand awareness, fewer clicks on search and fewer visitors to the site. Then marketing comes back to you next year, asking why the hell traffic to the Quincy Honda dealership website is down when search volume is higher than ever.
What you (the ad agency) need is a way to integrate data from those five different sources into one single platform. That’s called attribution analysis. There are a few companies that offer this as a service. A lot of companies try to do it in house, integrating their display reports with their search reports and their e-mail reports and mixing it into a bowl. There’s no definitive method – yet. But the demand is growing. Because after 2008 took a dump on everyone’s budgets, every marketing department in America has had to do more with less. They want to spend their budgets where they’ll do the most good. For that, you need a holistic look at all marketing data from all sources – online and offline.
What I’m doing currently (among other things) is researching attribution analysis providers. There’s no one-size-fits-all program: different companies need different attribution systems. For some clients, we might recommend a trusted partner agency; for others, we might do it ourselves. But we need to know more about the state of display, offline marketing and attribution management before we can recommend anything. That’s where I come in.
… I said you might find this post interesting.