Above: one example of GumGum's in-image advertising.

Above: one example of GumGum’s in-image advertising.

With all the discussion of native advertising, one format that gets little attention is in-image advertising. Some may argue whether these in-line ad units are in fact native, but GumGum, the company that invented the category, has made an argument for why they are. Regardless of the outcome of a semantic or ontological debate, new data coming out of the company suggest that the units work — and after all, that’s the only thing that matters.

By way of definition, in-image ads are ads served either as an overlay or take-over of an online image. GumGum uses proprietary technology to evaluate the pixels of each image on its premium publisher sites, combined with facial recognition and image clustering methods to identify any people or items. The company then applies its ad targeting to deliver contextually relevant ads. Through its technology, the company not only determines the subject matter of the image, but also whether it is brand-safe — based on nudity detectors and other features. GumGum ads tend to be visually engagaing, with interactive annimations, video content, offers, or other experiences.

Typical GumGum advertisers are global automotive, CPG, and entertainment brands. For example, Chevrolet may choose to deliver an ad for its new Camaro over an image of a recent or classic model of the same car. Ford, for that matter, may bid to deliver Mustang ads over the same images of its competitor’s vehicles. Another common scenario is that studios promote upcoming films over highly-trafficed pictures of celebrity cast-members. Also CPGs often advertise their household products over images of food or celebrity families.

The six-year-old company has identified over 100 billion image assets across its network of premium more than 1,000 publishers – a number it strategically reduced from 6,000 in the last year to increase quality – and reaches 60 million unique visitors per month. These publishers include New York Daily News, Alloy Digital, US News & World Report, and Meredith, among others. Among brands, it has worked with General Motors, Toyota, Disney, Home Depot, AllState, Wells Fargo, L’Oréal, and other giants.

More importantly, data suggests that the company’s ad units perform, according to founder and CEO Ophir Tanz. GumGum’s image takeover product has seen average click through rates (CTR) between 1 to 2 percent, Tanz says, with some campaigns returning as high as 6 percent, compared to industry averages 0.1 to 0.4 percent. Its standard in-image ads typically generate 0.4 to 0.8 percent CTRs according to the CEO.

As the company wrote in a recent blog post, “In-image ads work because they get seen, appearing on contextually relevant images, in-line with editorial content where a consumer’s attention is actively engaged.“ Another advantage of in-image ads is that they allow publishers to reclaim the areas site real estate often dedicated to traditional display advertising — which can often be as high as 30 percent of a given screen.

Today, 25 percent of GumGum’s traffic comes on mobile devices and, like most other adtech companies, it continues to evaluate its strategy on this new platform. Limited screen real estate and attention spans make traditional display advertising more difficult on mobile devices. GumGum’s image takeover units have outperformed on mobile, according to Tanz

GumGum has raised a total of $10.8 million across four rounds of angel and Venture funding, with its backers including New Enterprise Associates, GRP Partners, First Round Capital, and Crosscut Ventures. The company spent its first four years focused on building its technology and only began assembling a salesforce in the last year. Given the effectiveness of the ad product, it shouldn’t be surprising that the company saw its revenue increase by 500 percent year-over-year in 2012 following this move. GumGum is operating profitably today and has no immediate plans to raise additional funding according to Tanz.

While roughly one third of all content published online is images (several trillion per year), the analysis required to determine the content of those images is a difficult problem – so much so that even display adtech giant Google has not offered a solution (although it’s invested in some). Despite this, GumGum is not alone in the category. It faces competition from several well-funded companies, including Google Ventures-backed Luminate and former GumGum partner Vibrant. Each company has approached the problem with a different combination of automated and human-powered tools. Tanz, for what it’s worth is confident that GumGum’s technology is superior to that of each of these companies, but it nonetheless makes for a more competitive customer acquisition process.

Today’s savvy online consumers demand that digital advertising be non-interruptive. Publishers, on the other hand, are perpetually searching for ways to drive engagement. And while many other forms of native or promoted content can be thought of as deceptive or manipulative, in-image ads avoid this issue while offering a solution to the needs of both consumers and publishers.

The challenge for GumGum is that today’s winning digital advertising technologies are tomorrow’s old news. In order to maintain the attention of consumers, and thus the business of brands and publishers the company will need to continue to innovate. Tanz and his team have proven capable in this regard in the past and have teased a number of new products currently in development. If they stick to the same non-interruptive and contextual ethos of its current products, it’s fair to expect similarly positive results.

[Image via demo.gumgum.com]