Programmatic for the People — Working with the Machines

Posted Jul 11, 2017

One of the terms thrown around on the advertising side of mobile is the term “programmatic.” It sounds simple: automated advertising, but the reality is much deeper. The term “programmatic” describes advanced marketplaces that utilize complex algorithms to analyze massive amounts of metadata, then use real-time bidding methodology to bid on impressions from thousands of supply sources to deliver super-targeted and dynamic ads to users, yielding maximum earnings to publishers and efficiencies to advertisers.

That’s a mouthful, so we’re going to have a deeper look.

Pro·gram·mat·icˌ (prōɡrəˈmadik/)
There have been a plethora of articles in industry publications about what programmatic does and its potential, but with it being a new development amongst a constantly evolving industry, sometimes the basic facts can be lost as details gain focus.

Put simply, programmatic ad buying refers to the use of some form of software to automatically purchase advertising.

This is as opposed to the traditional direct sale advertising process that involves RFPs, back and forth negotiations, and manual insertion orders. Programmatic advertising can have a number of possible goals, including views, clicks, or other engagements defined by the advertiser.

Have you ever wondered why you see an ad for products from Amazon across the web and apps, right after you decided maybe you don’t actually need a sous-vide cooker? Or see tourist attractions for cities you’ve recently searched flights for? On a fundamental level, those are both examples of programmatic-powered retargeting & behavioral campaigns at work.

At its heart, programmatic advertising is designed to increase efficiency.

Working with machines
Normal campaigns are bought and sold by human ad buyers and salespeople. Humans, the organic beings that we are, have more nuanced issues to think about — things like what’s for lunch, who’s messaging them on Slack, and on a basic level, how to survive — and we can only analyze so much data at once.

According to our most recent App Install Marketing Survey, the sheer number of targeting parameters advertisers are taking into account makes having humans micro-manage every step of the process impossible.

OS, geo, device, demographic, look-alike, source, interest, and connection all are considered important factors in placing an ad. That doesn’t even include behavioral targeting, which is considered an important factor by 90% of advertisers and can include thousands upon thousands of data points on its own.

How can advertisers take all their targeting considerations into account in real time? Enter the machines.

Machines are the savvy assistants that can analyze more data in a minute than a human can read in a lifetime, and make decisions based on that data within a matter of milliseconds — without even having to turn on their cooling fans (read: breaking a sweat).

While there’s still a place for high-level site-wide targeting for many campaigns, laser-focused individual targeting can often yield better results for both advertisers and monetizing publishers, while also making sure end users don’t see ads that might bore them, or that simply aren’t relevant.

In the past, ads were traditionally purchased and sold in bulk at the site level with sorely limited audience targeting finesse, in a manner similar to buying full page ads in traditional media, or commercials during certain TV shows.

Since then, the data capabilities of in-app mobile advertising has enabled more advanced, device-specific targeting that yields better results for both the advertiser and the publisher alike.

Programmatic mobile advertising takes this sophistication even further  — campaigns can now thrive in dynamic, RTB environments at an efficiency never before possible.

Programmatic advertising software does this optimization at the speed of light.

Completely automated mobile ads?
At first blush, it may sound like programmatic is replacing every human in the chain of advertising, but this isn’t the case. Artificial Intelligence is still a long way off from replacing the talented people making ad creative, as well as making the high-level decisions about campaign costs and top-level targeting parameters.

Leveraging direct-sold campaigns is still the dominant methodology for most performance advertisers. The human element offers a unique perspective and optimization layer that accounts for the uniqueness of inventory and more complex, less binary win/lose campaign objectives.

However, navigating the sheer volume of mobile devices and connected users across ALL inventory sources in the mobile ecosystem makes precise, individual-level targeting impossible to manage for human beings.

Using programmatic advertising, advertisers can target individual end users utilizing mountains of data to serve ads that are the most relevant to those users, which speeds up the buying and selling process to an almost instant rate.

Types of Exchanges
An exchange is a system where programmatic software makes the decisions between all the different ad creatives and campaign objects. For advertisers, exchanges for in-app inventory offer a higher quality, brand-safe means to buy programmatically. For monetizing apps, those same exchanges open up inventory to more advertisers, driving revenues up as bids are placed.

  • Open Marketplace (OMP) — A true, open exchange where Publishers send their impressions out for purchase by way of an open auction. The bidders, also known as a DSP (Demand Side Platform), compete for each impression with their best qualifying bid – the highest bidder gets the impression. End of story. In a programmatic OMP, it’s a 100% holistic auction where bidders usually pay on a 2nd price auction dynamic.
  • Private Marketplace (PMP) — PMPs are more exclusive ways that advertisers/DSPs can buy inventory programmatically. A PMP is more of a more “luxury” service – bidders get prioritized bid preference on premium inventory. It can potentially include other value-adds as part of the deal, such as competitive pricing, creative resources and access to first-party data for more targeted campaigns. Unlike OMP, PMPs are contracted deals signed between a bidder and the exchange. This type of exchange is a mid-point between direct programmatic exchanges and fully open ones.
  • Direct — Direct programmatic allows an advertiser to programmatically buy impressions directly from a specific app. In this way, they’re very similar to a manual, human-powered I/O based buy – the rates and volumes are agreed on by both the advertiser and exchange/publisher. The impressions are all still bought programmatically but in a guaranteed and forecasted way. This means that the ads themselves are placed programmatically in front of the right audience.

Working with Programmatic
All that said, the human part of advertising is still indispensable.

It’s important to maintain the quality of user experience, and for that, even the most advanced programmatic systems need human beings. Advertisers have to trust ad networks that the results they’re delivering are genuine, and being seen in the right brand-safe contexts, a problem some social ad networks have run into recently when removing human interaction from the equation entirely.

There’s no better way to ensure honesty and build positive relationships that boost creative collaboration and increase results for every part of the funnel than good old fashioned human-to-human interaction.

We believe strongly in programmatic advertising as the next huge step in mobile ad tech, but we also are firmly committed to using it as a means to an end, not an end in and of itself. We’re still driven by our mission to deliver great results to our advertising partners and providing personal, human-to-human customer service as part of that process.

The human teach is still needed, but optimization is now squarely in the hands of machines.


Latest at AdColony