Industry Perspectives

Real-Time Advertising Just Got a Bit More… Instantaneous

Blog Post
The AppNexus Team

But as Chris O’Hara points out  in a recent piece in AdExchanger, the truth is somewhat more complex. Can advertisers engage successfully with their individual consumers in near-real time? Absolutely and yes. But there’s a catch: They can only make it happen under certain circumstances.

Allow us to explain.

For argument’s sake, suppose you’re one of the top e-commerce retailers in the United States. On any given day, you’ve got hundreds of thousands (or millions) of visitors browsing your digital properties in search of things to buy.

For every visitor who makes it to the final lap of the sales funnel, hits “purchase,” and becomes an official customer, plenty of other visitors don’t. After a few trial-and-error online searches for waffle irons, vintage headphones and Legos, they drop suddenly from your website, without so much as saying goodbye… or abandoning any items in their shopping carts.

Your obvious goal is to get those customers back to their senses; nudge them back into the sales funnel; urge them into filling their carts up with all the things they were searching for just five minutes ago. But time and recency are ticking – and consumers’ memories are already fading. Whatever website, app or cat video they’ve moved onto, now would be the right moment, if ever there was, to hit them with some real-time advertising.

Given the ideal circumstances, an advertiser can do just that. Their demand-side platform and their data-management platform already have a cookie-sync in place. In other words, their DMP is able to cross-reference the new user ID data (i.e. the recent actions of the shopper) and add it to the pre-existing list of audience segments the shopper belongs to already on the DSP.

If a brand or agency can put enough of these puzzle pieces together, and transition the individual into a new audience segment relatively quickly, then they stand a fighting chance of engaging that person in real time…

But let’s emphasize that if again. Chances are they don’t.

Transferring a customer from one segment to another can only happen if there’s an actual cookie match already in place. If the advertiser’s DMP and DSP haven’t yet made the match, then an advertiser will have to wait for a data transfer to occur before a match can be made. Using most of the standard technology available on the ad-tech market today, they’ll be waiting a long time: The standard method for shifting one person from one segment to another, a process known as batch data transference, often takes hours – sometimes even overnight – to complete.

That’s bad news for advertisers. The sooner a brand can regain the interest of consumers, the likelier those consumers will find their way back into the purchase funnel. But without any reliable, timely means of shifting potential consumers to new audience segments, the greater the chances are that those consumer will forget all about ever browsing an e-commerce site in the first place.

For an e-commerce marketer, batch transfers of data put an end, essentially, to the possibility of meaningful real-time advertising. After all,  an “overnight transfer” doesn’t exactly resolve itself in seconds or minutes.

Putting the real back in “real-time” with AppNexus Instant Audience

Now that we’ve entered the era of the “programmable internet”, where millions of interconnected devices, platforms and channels have enough collective data to help businesses make better-informed decisions about finding new customers (not to mention help individual customers make smarter purchasing decisions), advertisers have begun examining ways of unlocking and unleashing programmable’s intelligence.

In his AdExchanger article, O’Hara posits a hypothetical future of “dynamic segmentation”, where platforms can track the actions of online consumers as they navigate their way between various devices, websites and apps in real time. In the future, argues O’Hara, such platforms would be able to translate those real-time actions into real-time data that advertisers are able to follow up on instantly and meaningfully. Such platforms would also give advertisers the actual means of instantly targeting those audiences in real time, regardless of DSP or device type –without any overnight waiting for a batch transfer to occur.

Mr. O’Hara, the future of real-time advertising is… well, now.

A new feature called AppNexus Instant Audience allows advertisers who use the AppNexus platform to move users between different audience segments in near-real time.  Unlike large-scale batch transfers, streaming transfers can take place quickly and on a user-by-user basis. That means the number of user IDs being added to a new audience segment no longer needs to be added as a single, giant batch overnight. In the space of two minutes or less, Instant Audience essentially lets advertisers practice O’Hara’s definition of “dynamic marketing.”

Granted, Instant Audience doesn’t add audience segments at speeds comparable to, say, the high-frequency trades conducted by Wall Street supercomputers (which can be executed within the space of nanoseconds). But then again, real-time follow-up ads don’t need to take place at the speed of light in order to bring customers back to a purchasing state of mind. It’s enough that a tool like Instant Audience reminds internet users that only five minutes ago, they were searching for a new pair of skis on a e-commerce retail site ; and that those skis are more than likely still waiting for them to hit “purchase”.

Harsh Jiandani is the Senior Manager of Client Advisory in AppNexus' Advertiser Technology Group. He heads the Growth division which focuses on building strategic partnerships with innovative companies in the industry. He joined AppNexus after running his own startups for several years and has an MBA from NYU Stern.