Anyone with an internet connection has been served an ad. No matter what you’re doing online, whether it’s shopping, planning a trip or researching a physician, chances are your behaviors are being tracked and ads deemed relevant to you are being published on the sites you visit and the apps you use. Consumers have largely come to accept that they are being tracked, but exactly what information is being gathered by the invisible pixels painting the internet and how it is used is not as commonly understood.
Marketers should be literate in the language of digital tracking and advertising so that they can inform an integrated marketing campaign.
One of the oldest tracking technologies, a cookie is a script added to a URL to track visitor engagement with a webpage. Most consumers are familiar with cookie use for retargeting, wherein their visit to a website or their product browsing is tracked by a cookie, and once they leave that site or browsing session, an ad is served to them related to the site or products they just viewed.
“At the most basic level, we’re using cookies … to identify that someone has either visited a site or seen an ad and [then] identify that same user again when there’s another advertising opportunity,” says Matt Ellinwood, senior director of data products at advertising software platform Tube Mogul.
Using cookies, advertisers can assign anonymous IDs to consumers. “The industry has to infer more about those groups of IDs,” Ellinwood says. These assumptions could be about the gender of the users, their purchase behavior or any number of characteristics to narrow segmentation.
When a visitor has provided a website with personally identifying information such as a name, e-mail address or log-in information, the site can assign a deterministic ID to that person. Actions that the visitor takes on that site can be tracked as first-party data. “That’s really similar to point-of-sale data,” says Ellinwood. “If your customers are coming to your website and are logged in, you can use a cookie to identify them and follow their purchase behavior.”
In the mobile world, cookies are notably less useful to advertisers. That’s because 90% of consumers’ time spent on mobile devices is in-app as opposed to the mobile web, and without a web address, there can be no cookies. Instead, advertisers use device IDs to track and target consumers.
Device IDs differ from cookies in that they don’t have to be placed by an app the way cookies are on a web page; they are recorded from the device on which the user accesses an app—be it an iPhone, Android or tablet—and they are rarely reset, making them reliably consistent across many visits from the same user. “It’s an anonymous ID you’re able to use to identify the same user at different times,” says Ellinwood. “The way you set it and the way you get it is a little different [from a cookie], but the application is almost identical.”
Just like cookies, Ellinwood says, device IDs can be grouped by segments to create audience targets. “Whether it’s a cookie or device ID, the way the industry is going, we don’t really care anymore. We know we want to serve a user a specific ad because we’ve seen them before, and we know they’re male, for example, and we’re able to make that decision based on the ID right then and there.”
Advertisers can also use location-based data points to segment their audience. On desktop, the best way to gather this information is by IP mapping, and on mobile it’s geolocation. Unlike cookies and device IDs, IP mapping and geolocation serve different purposes. Whereas identifying the location of a mobile device can be advantageous to bricks-and-mortar businesses looking to drive traffic to their stores, IP mapping can be studied to surmise characteristics about the consumers who are browsing on an address.
“We’re turning the router inside your home into the most accurate way to target digital ads to you,” says Greg Mosley, director of sales at El Toro, an ad-targeting company that has mapped IP addresses to physical households, which can then be cross-referenced with publicly available data sets. Mosely describes the use of IP mapping to digitally target advertising to consumers like a direct mail campaign. For example, a car dealer could target the home addresses of car lessees whose leases are near expiration or owners of competitors’ cars and cross-reference them with IP addresses that have been verified with home addresses to serve those households digital advertisements.
“Cookies are really your online behavior,” says Dustin O’Dell, senior director of business development at Barometric, a tracking platform. But there are limitations to knowing the behaviors of an anonymous consumer, Mosley notes. “One of the big concerns in the online advertising space is who’s seeing the ads. When you’re targeting in a cookie-based world [the web], you’re targeting an unknown. If I’m selling luxury vacations, cookies say all the people who’ve gone to Forbes.com are highly affluent. That’s not accurate.”
A good marketer won’t rely solely on cookies, but even with the incorporation of device IDs and geolocation technologies, it can still be difficult to recognize consumers across devices. Imagine a consumer sees your ad on a tablet, browses your mobile website on their cellphone and converts on their desktop computer. The data provided by those impressions would suggest your brand interfaced with three different consumers because the tablet and phone will have unique device IDs, and the cookie you may have used on your site recognizes the desktop visit as unique from the mobile visits.
“If you target them as three different people, your targeting may not be as effective as it could be,” says O’Dell, because it removes the possibility of targeting sequentially or upselling your target.
A smart digital marketing strategy won’t rely on a single method of tracking, nor will it take data at face value. What differentiates the spray-and-pray approach from an integrated marketing approach is data science.
Big Data must be translated into smart data. From IP mapping marketers can learn a consumer’s postal address. When a particular device ID, tied to a particular geolocation, shows up every day from 8 p.m. to 6 a.m., it can be assumed that that device lives there, says O’Dell, allowing marketers to deduce that the person using that device lives there as well. When the geolocation of the device matches up with a postal code, any devices in that home environment are representative of the same consumer—or at least a small set of consumers (a family) who are interested in seeing the same ads. And to those devices, a persistent ID is assigned.
“If I’m targeting you at home, whether you’re going to Forbes.com or PerezHilton.com,” Mosely says, “because we’ve identified you as the right target, we’re no longer having to bid on the context of the website.”
Cutting out the guess work or the chance of serving an ad to the wrong audience is every marketer’s aim; achieving that efficiency is a process of experimentation. “That’s the goal: to make assumptions about where you’re going to see better performance for your ads by getting them closer to your target, inspecting performance and then shifting your budget and your strategy according to whether or not those targeting strategies are delivering for you,” Ellinwood says. “What we’re gravitating to is how to serve the best ads to the fewest number of people at the right time so there’s more efficiency and less waste. But that’s really hard.”
Experts agree there is no single best practice or technology for knowing your customers online. Rather, marketers need to evaluate the right mix for their budget and goals. “You balance scale and precision in a lot of cases,” O’Dell says. “I call it the dartboard effect. You can shoot for the bullseye, and those are super premium, top-level audiences, but you may only get 10 people. You should absolutely use those 10, but we understand you need scale, too. The more scale you get, you have to understand the precision may not be there, but if you use a combination of both, that will get you to a place that’s beneficial.”