Brian O'Kelley
May 19, 2016

Programmatic is extinct: It's evolved into something way more powerful

Programmatic is dead. For those of you who rely on digital marketing to engage customers, don't be alarmed! Brian O'Kelley, CEO of AppNexus explains.

Programmatic is extinct: It's evolved into something way more powerful

To clarify: I don’t mean to suggest that "programmatic advertising" – a term that Cowen and Company helpfully defines as "digital ad buying that involves automation and data-driven decision-making, frequently in real-time" – will cease to operate. I invented the real-time auction, and rest assured, it’s not going anywhere.

Rather, what we know today as "programmatic advertising" is going to look much different. It’s going to get a lot smarter – it’s going to get a lot easier – and it’s going to contribute far greater value to marketers, publishers, app developers, and consumers.

In short, programmatic is going to become programmable.

 In its time, programmatic advertising was good enough. It did the trick. It was the best way we knew to monetise digital apps and websites. But it was built for the one-channel world of desktop web browsers and email – a world that hardly exists today. Programmatic is unwieldy, inefficient, slow, and limited.

Programmable advertising is the very opposite: seamless, efficient, quick, and infinite in its possibilities.

As we consign the Programmatic Age to the annals of history, it’s time to usher in the new, Programmable Age of digital marketing. And one of its key pillars is the new data economy.

The new age of marketing is here. Come to the Programmable Marketing Forum on Wednesday 8th June and you're already ahead of the curve.

The data economy
In the Programmatic Age, data was fragmented, cautiously guarded, and of limited use. Marketers and advertisers were mostly unable to use their data assets to customise campaigns. Instead, they relied on standard, black-box algorithms. Publishers cast a wary eye at Google and Facebook – partnership with these behemoths seemed inescapable but sharing user data with competitors was unthinkable. Even if buyers and sellers could use their data free of worry, it had sharply limited applications.

The Programmable Age is different. Marketers are able to harness their data to build customised algorithms that inform smart "decisioning" – not just about when (what time, what frequency, what pacing) or where (what screen, what geography, what format, what channel) to deliver creative, but in some cases, what creative to deliver. 

Not every brand can – or even wants to – hire a team of data scientists and developers to capture the potential of programmable advertising. But in this new programmable age, data as a service opens up considerable growth opportunities.

Complex algorithms, simple bidding
One of the standout early adopters of data as a service is BannerConnect, a highly innovative advertising technology company that’s part of the WPP family. They’re helping brands and agencies solve a core problem.

Until recently, advertisers either juggled multiple, micro-targeted campaigns manually, or worked with multiple ad-tech partners to achieve separate campaign goals. BannerConnect uses data analysis to sort multiple user segments into groups, each with its own bidding strategy. It then builds targeting based on time and recency and applies different bidding strategies for each audience segment. And there is a strong element of machine learning involved.

For one campaign on behalf of a leading automotive brand, BannerConnect studied nine distinct target segments, each containing consumers who viewed specific car models online. Historically, these segments performed similarly and what minor differences existed between them didn’t merit the cost and labour involved in establishing nine bidding strategies.

BannerConnect did two things. First, it created a benchmark campaign that treated all nine segments the same: the control group. Second, it established a competitive campaign that grouped the segments into three cohorts and layered onto these cohorts nine immediacy settings. In other words, it used data indicating consumer interest and immediacy of action to collapse 27 different campaigns into one seamless algorithm.

The results: BannerConnect’s automotive client realised 76 percent lower effective cost per action (eCPA) with its consolidated campaign than with its benchmark campaign. The client ultimately spent 60 percent less time on campaign management.

And time equals money.

When it rains, it pours
One of my favourite examples of data as a service is Accordant Media’s work on behalf of an e-commerce client whose sales volume and return-on-advertising spend is positively correlated with inclement weather.

Accordant, a programmatic media company specialising in targeting and optimisation, built a custom data feed that imported and analysed live weather information from the National Oceanic and Atmospheric Administration, a division of the United States Department of Commerce. It then constructed a proprietary algorithm that calibrated its client’s campaign bid and budget pacing according to localised changes in the weather. The result: a weather-targeted campaign that reduced cost-per-scale by 20 percent and increased overall volume by almost 40 percent.

And they discovered that when it rains, it really does pour – at least for Accordant’s client. Targeting users in inclement weather was up to 33 percent more effective than targeting users in favorable weather conditions in three US test markets.

Working together – seller turns buyer
For publishers, the Programmable Age ushers in new opportunities to pool data with non-competitors. Already, we’re seeing some of the most innovative publisher coalitions – for instance, the Local Media Consortium – syndicate subscriber data, with baked-in protections against leakage or exploitation of proprietary assets. 

Technically, the Local Media Consortium is a seller rather than a buyer of digital advertising. It represents over 65 iconic media companies in the United States, including Belo, Cox, McClatchy, Gatehouse, Chicago Sun Times, and The Boston Globe. But through its data syndicate, member organisations collect, organise, target, and syndicate audience data between participating members. In effect, it is a tool for audience extension, which makes buyers out of sellers.

In the same way that innovative companies offer data as a service, the concept of the data syndicate has exciting potential. Authentication providers – meaning, any website or app that asks users to remain logged in – could pool and sell their assets, offering them at scale, with privacy provisions for the end user. In the new data economy, Facebook and Google won’t be the only actors capable of people-based, rather than cookie-based, targeting.

The takeaway
The new data economy is a powerful engine of innovation and growth in the Programmable Age of digital marketing. For brands and agencies, the opportunity to engage consumers with customisable and relevant creative – and to do so with smarter frequency and pacing – will improve return on advertising spend.

The old programmatic world, with its one-channel approach, just got a lot smarter.

Brian O'Kelley is CEO of AppNexus

The new age of marketing is here. Come to the Programmable Marketing Forum on Wednesday 8th June and you're already ahead of the curve.


Related Articles

Just Published

41 minutes ago

Live today: Performance Marketing Playbook

Register now to join today's performance-marketing conference, which starts at 9:30.

2 hours ago

Neil Heymann returns to Accenture after brief stint ...

The former Droga5 global CCO decamped to Publicis less than a year ago to open its New York-based creative collective.

8 hours ago

MSQ acquires design agency Elmwood to deliver ...

The design agency will merge with MSQ's existing brand design agency, Holmes & Marchant, creating a new entity under the Elmwood name.

9 hours ago

Qatar World Cup is advertising’s biggest ‘set-piece’...

Brands should start planning for Fifa's first winter World Cup.