"AdExchanger Politics" is a recurring feature that tracks developments in politics and digital advertising.
Today’s column is written by Jordan Lieberman, politics and public affairs lead at Audience Partners.
Happy election season, everyone. I’ve got good news and bad news. The bad news is that many of those who enter the political advertising space right now will exit within the year. The good news is that all they need to do succeed is solve this problem: Peak cookie targeting is in our rearview mirror.
Political and public affairs digital targeters, or really any digital targeting shops that must rely on deterministic data, must face this mounting problem. The best political and public affairs shops will adjust tactics; the worst will fake it.
In 2010, differentiation meant cookie matching the voter file. By 2014, if you weren’t cookie targeting a voter file, you were doing it wrong. The political cookie spent nearly a decade driving digital persuasion and get-out-the-vote efforts, while email, search and Facebook were the go-to for acquisition. But moves by giant tech players have all but ensured cookies will be outmoded during this campaign cycle.
In this campaign season, if political advertisers are only cookie targeting a voter file, they’re in trouble. Unlike some commercial targeting, politics relies stringently on a specific set of vote history and polling data. Lookalike targeting – especially in lower-turnout environments – has less value since a likely voter can pretty much look like an unlikely voter.
We’ve passed peak cookie-to-voter-file match
Last fall, Apple made changes to prevent the use of most cookies on its mobile web browser. It’s good for major players like Amazon, but it’s bad for virtually every other type of third-party advertiser, including anyone using a voter file.
Separately, on Feb. 15, Google began blocking ads on Chrome that don’t meet standards developed by the Coalition for Better Ads. During the first two months of the initiative, the browser will block 7.5% of page views with noncompliant ads. This threshold will decline in the ensuing months as political targeters’ pain increases.
Even if Apple and Google hadn’t gone after the third-party cookie, we’d see precision declines from device fragmentation and a race to the bottom with voter data quality.
So, post-cookie, now what? As voter file cookie targeting falters in this campaign season, other options remain. We’re not going back to email spam, given better filters and worse match rates to real – not throwaway – emails. Most in the community are not planning to trust the duopoly with bigger budgets. So then what do we do, post-peak cookie?
First, there is ISP-authenticated IP targeting with an attached device graph. Using IP addresses that are provided by ISPs means that the targets’ postal addresses can be matched to their IP address with 100% certainty since the ISP is the source of the IP address. ISP-authenticated IP targeting reaches users on IP-addressable set-top boxes and beyond the home due to its device graph anchored in a home’s IP address. This is only possible if the programmatic platform has an IP bidder, allowing measurement of reach and frequency because all the IP addresses come from one source, unlike cookies.
Another solution many will look to is blurring the lines between voter-targeted digital and voter-targeted cable. Addressable television’s rise is choppy and uncomfortable, but in many places, a voter file-targeted pre-roll ad can end up on a television screen instead of computer. In the end, we’ll see more digital political vendors pushing addressable – or least linear optimized – cable to supplement shrinking cookie pools.
We’ve also seen a spike in interest in geofencing for political and public affairs goals. The challenges with scaling, third-party data matching and privacy will limit this solution. I do not expect large campaign dollars to end up there, but for precisely targeted public affairs geofencing campaigns, it works.
Worst case, political targeters will party like it’s 2009. Some did just fine with direct site buys, Pandora, Facebook and search ads.
Political tech usually follows consumer tech. We don’t like change, and we value relationships over merit in too many cases. But in this instance, political targeting will need to lead the way since it has the rare combination of hard deadlines (also known as elections), a non-negotiable need for deterministic data (also known as highly targeted segments on the voter file) and relatively tight geographic boundaries (also known as statehouse and congressional districts).
Differentiate and win, or stay on autopilot and write political out of next year’s business plan.