Blinding me with science: Ad optimizers

Inplace #2

science_small.jpgADOTAS — When one takes a step back and looks at the online ad market today, there doesn’t seem to be a lot of good news: CPMs are dropping, display ads are on life support, the marketplace is incredibly inefficient, advertisers are demanding more transparency. The list goes on.

Despite all this, not all hope is lost. In fact, there is a lot of innovation happening today, primarily coming from a bevy of startups who are adding a tremendous layer of value between the supply and demand. These companies are still getting funding when other sectors have gone dry. Some even recently got acquired.

This innovation is coming from the optimizers. And they nest in what we in the space like to call the “ad optimization stack.”
This ad optimization stack is made up of an increasing number of companies. In fact, it can be quite confusing, so here’s a Beginner’s Guide to understanding the optimization players and in what area of the stack each plays:


Ad container optimization encompasses placement, ad creative, and reach. Placement has to do with a set of technologies that optimize the most appropriate place for any particular ad creative on a publisher site, right down to the x and y coordinates. For example, knowing if an ad is below or above the visible fold of a web page is an important (and often unknown metric).
Creative optimization refers to any technology that performs testing related to colors, ad styles, engagement interactions, and ad creative. Performing optimizations on creative attract the right types of consumer behaviors for whatever offer is being presented.

Last, reach optimization focuses on analyzing the distribution path of where ads are being seen. Whether the ad is being distributed directly to a publisher site, an exchange, or an ad network (or all three), is important to ensure that the appropriateness of a campaign is maintained (e.g, some brands do not want to be on user generated content sites or blogs).
Some interesting companies in this space: SnapAds, Tumri, Clearspring, Gigya, WidgetBucks.


There are many form forms of targeting, including behavioral, demographic, contextual, category, and interest-based/affinity. Basically all of these forms of targeting address different types of objectives that primarily tradeoff between reach and scale. For example, behavioral targeting, which is a set of segmentation techniques to use consumer behavioral data to capture and use intent data, is one of the best ways to target a consumer who is far down the conversion funnel.

On the opposite end of the spectrum, category-level targeting is a classic way in which traditional display ad networks target customers using broad segmentation buckets (like “technology”, “business”, etc). Thus, behavioral offers the most targeted but the least reach while the inverse is true for category targeting. There are a bunch of technologies in the middle of these approaches that use a combination of publisher keyword content to derive interest in a specific product or topic. These techniques fall under the contextual and interest-level targeting. Some interesting companies in this space: BlueKai, Exelate, Lookery, OthersOnline,


Yield management technologies vary from companies that are focused on yield ad inventory, ad networks themselves, or specific ad offers within an ad placement. Some of the larger optimization companies in the space have focused on being today’s answer to what broader ad exchanges are purporting to offer. They are focused on using a single dashboard across multiple geographies to manage hundreds of ad networks.

They also provide additional services like account management, other types of specific optimization (e.g, text ad optimization), as well as vetting technology (e.g, ensuring quality ad content being served thru one ad placement). Additional optimization technologies provide predictive and advanced A/B testing technologies that focus on allocating specific inventory or offers inside the advertisement. Often times, these technologies utilize sophisticated algorithms that take cookie (person), domain, or aggregate behaviors to predict the most appropriate ad offer.

Some interesting companies in this space: Pubmatic, Rubicon Project, YieldBuild, YieldX, WidgetBucks


The key distinction between marketplaces and ad exchanges are that marketplaces aggregate advertiser demand on their own and exchanges aggregate demand from other ad networks. Ad exchanges are quickly becoming the favorite way to manage multiple remnant ad networks versus the standard “daisy chain” approach that is applied today. Daisy chain management is a process in which a publisher queues up multiple ad networks behind one ad tag (or placement) in order to to get the highest eCPM of each ad network, as well as managing the highest fill-rate for one ad placement.

Some interesting companies in this space: Ad Exchanges: Right Media’s Publisher Media Exchange, Microsoft’s AdECN, DoubleClick

Advertising Exchange

Auction-based marketplaces: ContextWeb ADSDAQ, AdBite, Turn, Google AdSense, and Yahoo! Publisher Network
There has been well over $300 million dollars invested in the ad optimization space in the last year. As you look up-and-down the ad optimization stack, you will notice that these players focus on math version people.

The new players in this space have realized that the value added services are those that take advantage of the huge hegemonic moves that have been made by the larger players on both sides of the supply and demand aisle. We are now in the phase of refining the online advertising story. We’re in the age of bringing more science into the existing landscape.