Programmatic Ad Buying: A Tactical Primer

A few months ago, we covered the four strategic phases of lean display advertising. The first phase consists of setting up your initial tests or experiments, which means choosing your initial campaign tactics to later observe which ones show the most potential for further optimization and scaling. This primer covers most of the campaign tactics available with programmatic ad buying or real-time bidding (RTB) platforms.

For each tactic, my goal is to describe what it is, why it’s important, how it works, and what you should keep in mind — all very high level. Keep in mind that most of these tactics are mutually exclusive, meaning that they generally shouldn’t be combined, unless you really know what you are doing, or unless otherwise noted. Let’s now shine some light on each of these campaign tactics according to their general category: display, mobile and video.

1. Display Tactics

By display, I mean laptops, desktops, and media centers running common Web browsers like Firefox, Chrome, Internet Explorer, and so on — pretty much anything other than mobile devices. You could also group this into the “desktop PC” category, which has been steadily shrinking compared to mobile devices.

Source: KPCB

Nevertheless, it remains the largest segment of RTB supply worldwide, which is why it remains the core of most digital media plans.

There are several approaches to targeting audiences in this environment:

• Site Buys — The oldest tactic for targeting display ad campaigns is to simply choose the publishers you wish your ads to appear on. This is traditional inventory buying, which is still very much possible with programmatic platforms. The only caveat to this tactic is that, as we’ve discussed before, ad serving priority is generally lower via RTB, which results in later session depths. Therefore, if you want to scale up from a well-performing site on RTB, it’s best to approach the publisher directly for a direct deal, or use a “programmatic direct” platform, if available.

• Retargeting — Retargeting is a perfect example of “audience buying,” which is to say: the marriage of vast amounts of inventory with behavioral targeting data. This data is collected by placing a small chunk of code on your website or landing page, similar to how you would add code for an analytics platform. Since this data comes from traffic that you control, it’s often referred to as “first party” data.

With RTB, we know you can evaluate and bid on each individual impression, which not only allows you to effectively buy audiences by matching behavioral cookie data with it, but also allows you to do so at a large scale, thanks to the massive reach of the RTB supply side. When each impression is evaluated by a demand-side platform (DSP), there are unique (but anonymous) identifiers that are passed along, which are then matched with cookie databases that store behavioral audience data. If there is a match, the DSP will bid on the impression on behalf of whichever advertiser was targeting that audience.

Another flavor of retargeting is “search retargeting,” which allows you to target audiences based on keywords they have searched for in the past. In other words, you can target users by intent (the same factor that made search advertising explode) but with display advertising — albeit without the same timeliness. This was never before possible with display advertising, which is why it’s so important. The data for search retargeting is largely gathered from places like website search widgets, search engines, on-page analytics, and other “third party” sources. Overall, it’s a novel application of retargeting that complements search marketing efforts, and provides a great segue for search marketers to venture into the world of programmatic display.

• Audience — Audience targeting is similar to retargeting in that it relies on cookie data to target specific audiences (as opposed to specific publishers). However, instead of using your own first-party data, you use behavioral data from a third-party data company like BlueKai, Exelate, Bizo, and so on.

The beauty of using data from data companies is that it’s usually broken down into very distinct segments, whether it’s gender, age group, household income, education level, relationship status, banking institution, or hundreds of other specific demographic and psychographic categories. Since it comes from a third party, there is normally a price associated with using this data (typically anywhere from $0.25 to $2.00 CPM) in addition to the cost of inventory. For this reason, it’s important to keep an eye on your metrics to ensure that the cost of data is positively balanced by increased performance.

• Contextual — With contextual targeting, you can choose to show your ads only on pages that pertain to a specific topic or category. For example, you can show an advertisement for snowboards to people reading a page about snowboarding, or accounting software on pages related to finance.

Contextual data technology algorithmically analyzes pages, and categorizes them by topic, allowing you to limit your ad spend to pages that match your selected criteria. Like retargeting, it’s easy to get too narrow with your targeting criteria if you combine contextual targeting with too many other options like retargeting, demographic targeting, or too small of a geographic region.

Contextual targeting works best with text-rich publishers like blogs, news sites, and forums (which also happen to be the type of sites that have tons of pages on a wide variety of topics). Typically, the cost for contextual campaigns should only be marginally more than a standard Web campaign, with contextual fees ranging from $0.05-$0.10 CPM, in addition to the cost of inventory.

2. Mobile Tactics

Mobile RTB is a very exciting frontier at the moment. There are unique advantages with mobile that allow for extremely precise location targeting, and in the near future, extremely precise behavioral targeting that won’t depend on the previously-dominant cookie. Look no further than Twitter’s acquisition of MoPub for a glimpse into the future of mobile.

Source: BI Intelligence

• Mobile App — Most of the time people spend on their mobile devices is within apps. This is why mobile app inventory is among the most attractive inventory for advertisers: it’s where most of the market attention is focused. Not only that, many app publishers also have valuable first-party behavioral and demographic data that advertisers can use for enhanced targeting. As a result, mobile app inventory is the highest priced segment of the mobile inventory spectrum.

• Mobile Web — When I use the term “mobile Web,” I mean mobile optimized Web inventory. Not regularly-rendered websites shrunken to the point of illegibility on a mobile screen with equally illegible advertisements, but properly-formatted pages designed for mobile devices.

This means your creatives will be shown in their proper mobile sizes, resulting in a higher rate of viewability and engagement. Mobile Web inventory is similar to mobile app inventory in terms of creative sizes (320×50 primarily), but simply in a Web environment rather than an app environment.

• Hyper-Local — Out of all the mobile targeting tactics, perhaps the most exciting is the advent of hyper-local targeting, which uses GPS coordinates and a specific radius to target a very precise area. This is made possible by app publishers that have location-sharing capabilities and choose to expose the latitude and longitude of each impression.

For local businesses and events, hyper-local targeting allows for extremely novel ad campaigns. One example could be a local coupon campaign that targets a specific neighborhood based on its demographic profile. For entertainment districts, event promotion is another industry that could leverage the power of hyper-local ad targeting.

Imagine your target audience reading through their social news feed and seeing your ad for a limited-time promotion at an interesting establishment down the street. The same concept could also be applied to wearable computing such as smart glasses and the ad inventory that may appear in their “digital view” of the world. There is no question that advertisers will be able deliver a relevant message to the right person at precisely the right place and time.

3. Video Tactics

Video ad campaigns over RTB are a relatively new frontier that allow advertisers to apply the same precise, impression-level targeting, but with video creatives and in native video environments. In fact, video advertising is one of the fastest-growing segments of online advertising in terms of dollars spent.

However, the two largest barriers to getting started with video are the cost of inventory and the resources necessary to produce the video ads.

• In Stream — As the name implies, in-stream video ads allow you to serve up video ads (of 15 or 30 seconds) to either the beginning (pre-roll), middle (mid-roll), or end (post-roll) of online videos (pre-roll being the most popular, by far).

This is the closest that a digital marketer can get to reproducing the same advertising experience as television, which makes it an appealing tactic for brand advertisers. As a result, CPM rates for video inventory are currently the highest of all RTB inventory, with average rates hovering around $10 CPM.

• In Banner — Video ads can also be served within standard IAB ad units, like the medium rectangle (300×250), which happens to be the most common. There are also several attributes associated with video ad units, such as automatic audio, audio on rollover, audio on click, and muted by default. The ad quality restrictions of each publisher ultimately dictate which of these audio options can be employed.

The benefit of in-banner video ads is that they generally come at a fraction of the cost of in-stream video ads. In fact, the cost of in-banner video is basically the cost of traditional display inventory, and can be combined with most other display tactics.

Source: eMarketer

General Targeting Criteria

There are also general attributes associated with every impression that passes through RTB systems, which are often layered with each of the tactics mentioned above. Some of these overarching targeting criteria include:

• Time of day
• Day of week
• Frequency
• Geo Location
• Operating System
• Device
• Browser

You could consider these universal targeting attributes. They apply to each and every impression available to advertisers, and are generally safe to combine with any of the larger tactics.

Testing To Find The Right Tool

Targeting tactics are your tools in the programmatic world, and it’s always important to understand the tools in your toolbox. When you know what tactics you have at your disposal, you have a much better chance of choosing the right ones to meet your objectives.

Discovering the right tool ultimately comes down to which campaign tactic performs best for you in the real world. You can use this primer to help you narrow down the initial tactics to test with, but then it’s up to you to analyze the results, cut the weak elements, and scale the tactics that perform the best.