Memoirs On the Move From Programmatic to Paid Social

Lauren Clawson Oct 6 2017

The Move from Programmatic to Paid Social Advertising
In the (relatively) short time I’ve been in digital advertising, I’ve had the good fortune to work on what a lot of the industry would call “both sides” of paid media – social and programmatic. In truth, paid media sports more options than a Crayola starter-pack, but these two broader categories frequently make up the opposing sides of an ongoing argument that revolves around granularity of targeting vs. scale. It’s also understandable that social and programmatic would take up such a large portion of the digital conversation with the duopoly that is Facebook and Google.

If you’re anything like I was before I came to Portent, you may be curious about the pros and cons of “the other side,” whichever that is for you.

Here, I’ll offer my impressions as a young, front-line advertiser, who’s worked in-house at a programmatic DSP (demand side platform), and at Portent where we do a lot of advanced paid social targeting and advertising.

How are Social and Programmatic Alike?

Data is Key

The greatest advantage in paid media is the wealth of data that comes with it, and the ability to refine both messaging and targeting as a result. As with nearly all forms of digital, one of the biggest advantages is the immediate feedback from click-stream and conversion data.

Marketers have an immediate, quantifiable response to the content they deliver to users, whether that’s a display ad, a piece of native content, an in-line social post, etc. Industry-standard metrics like click-through, cost-per-acquisition, and engagement are available through both programmatic and social and allow marketers to change their targeting and creative strategy within minutes.

And although much of this will change as a la carte digital delivery becomes more of a reality for traditional formats, unlike traditional ad mediums like print, TV, and radio, with programmatic and social advertisers can see exactly who their ad is served to and at exactly what cost. Where collecting user feedback may have previously taken a lot of legwork, these platforms provide it in real time.

Audience Optimization

With all this data comes the opportunity to customize content directly to new or expanded target audiences of your choosing.

Social media platforms segment their audience based on what you engage with inside their platforms (or the external networks they own). Ad servers segment their audiences based on cookies or device IDs. Both vehicles give the marketer an opportunity to look at a group of behaviors and customize their content to appeal to that group. Or to find a new audience where their offer is most likely to resonate.

Constantly Changing Environment and Targeting Options

As digital media continues to mature and to steal share from traditional media, technology must change to keep pace with rising advertiser expectations of accountability and dependability. Both ad servers and social platforms are constantly working on technology to fix issues that keep clients up at night. Increasing brand safety, increasing viewability, audience targeting, all require that these platforms up their game to meet advertiser expectations.

How Are Social and Programmatic Different?

Variance in the Quality of Your Ad Placement

Running an advertisement through a Facebook Newsfeed and running a similar advertisement through a programmatic network may have drastically different results.

For most publishers (think news sites, sports, Flute Enthusiast Magazine online, etc.) the layout of a page necessarily creates different places to show ads which are of higher and lower quality or value. Top of page vs. right rail vs. further down the page. When you’re making a programmatic ad buy through a DSP, or even directly with the publisher site, you have to specify the type of placement that you expect, and hope that you’ve picked the right partner who will monitor your placements’ position on the page, its proximity to relevant content, distance from irrelevant or audience-unsuitable content, etc.

With a Facebook Newsfeed ad by contrast, content and ads are loaded only as a user scrolls, so there’s a lot less variance in viewability and engagement based on page placement.

Take a moment to realize this doesn’t mean programmatic display is evil, or lower quality. It simply means there’s more variability in the quality and viewability of placements, both good and bad.

Even though it’s got an edge in the phrase “Yes, but we know your ad showed up on screen”, where Facebook encourages rapid scrolling to keep up with your many friends and family, a great display ad placed next to content that will be stared at for minutes at a time could be significantly more impactful.

Black Metrics

Within the Facebook platform, there are several metrics and targeting objectives available that are specific to their proprietary algorithm. Meaning they’ll use an algorithm to optimize for the kinds of impressions and clicks that show the highest likelihood of delivering your desired action for a particular campaign, but you can’t take that optimization with you to other channels.

For example, the metric “Brand Awareness” is based on user feedback, relevancy, and historical brand surveys conducted by Facebook. They also specify that this metric can be imprecise, and give a relatively vague explanation when pressed further. As a marketer if can be incredibly frustrating, if not dangerous, to evaluate a campaign on a metric that does not have a concrete description, or industry-wide standard. If you’re working at a DSP or you’re an advertiser working with a team at that technology partner, metrics used to measure success of a campaign are relatively similar across the industry.

Fraud

Fraud within programmatic advertising (really any large-scale digital display advertising) has been cause for great concern and plenty of industry hand-wringing in the past few months. Social platforms certainly have an issue with fake or inactive profiles by contrast, but the number of bots or crawlers being served worthless impressions is significantly lower. Facebook has an advantage here as a user must actually log in to the platform and complete a series of questions to verify their identity before consuming the content and ultimately viewing any ads.

Misconceptions About Programmatic Advertising

”Every Digital Strategy Needs Programmatic”

Programmatic is not a one size fits all approach. Given the huge scale and reach of most platforms, it is scarily easy to spend an incredible amount of money on a programmatic platform without a justifiable ROI. It’s on you and your team to set clear goals, and make sure the investment is paying off at every step.

If your company or your client is only spending a minimal amount of ad dollars at present through Facebook, PPC, etc., it may not be necessary or wise to include an additional channel if you are not maximizing the ones you’re already using.

Caveat: this advice applies differently to a long-term sustained programmatic ad buy vs. a one-time programmatic direct campaign.

“Fraud Is Guaranteed In Your Ad Buy”

Although I just discussed how fraud is a key difference between programmatic and social media, the truth is that you’re not realistically going to have a huge portion of your ad buy eaten by evil bots (great mental image) as some alarmist reports would have you believe.

evil fraud bots destroying all of the ad budget in programmatic advertising

My mental image of evil fraud bots destroying all of the ad budgets.

As with any product, the choice exists between premium vs. value alternatives, and quality control is a known issue and an ongoing effort. (News-flash: your pant-legs aren’t the same length by accident. Someone periodically checks.) Programmatic advertising is no different, and you’ll see outliers on both the good and bad end of the spectrum. It’s the marketer’s job to weigh all factors (fraud, viewability, targeting, price) when developing a programmatic strategy. If your team wants the best inventory, you must be willing to pay for that quality.

”Your Social Strategy Will Work For Programmatic”

Social media offers a robust set of targeting and ad serving options, while programmatic has its own. User behavior (read interaction with ads) on a social media platform is different than behavior while browsing the wider internet. Use your brain and your data to tailor a strategy that fits the context of the moment you’re trying to reach a new potential customer.

Observations in My Own Move From Programmatic to Paid Social Advertising

Facebook Has Enormous Market Share For Good Reason

Before even accounting for the enormous consumer appeal of Instagram, I’d argue that Facebook’s giant and rapidly growing share of wallet with advertisers is warranted. Facebook makes nearly all of their revenue from on-platform advertising, so it is critical that their platform be easy to use for advertisers as well as end-user consumers.

The Facebook content team dedicated to their ad platform has created an incredible amount of literature that addresses both the goals and specific issues a marketer would have in using the platform. Their interface is easy to use, and their reporting is easy to understand. With the exception of Google and its several-year head start, any individual software company or individual platform (looking at you Twitter, LinkedIn) looking to create something that can compete on simplicity and utility for advertisers clearly has tough hill to climb.

Amount of Data Being Collected on Users Is Outstanding. And Yeah, Creepy

Because programmatic ad servers collect data via cookies (and sometimes device IDs) people can clear their cache and easily erase part of their online history, making it potentially challenging to serve an ad that’s most relevant. By contrast the data collected by social networks is tremendous compared to a typical programmatic lookback window.

Issues With Conversion Attribution

Every platform, whether its Google, Facebook, or a standalone ad server has issues with conversion attribution. “Last-click” is the standard model that many digital platforms still use, but most of the industry is in agreement that this model is inadequate for explaining what’s really working in your ad mix.

At their annual conference Google announced an initiative to better model the impact of each individual impression on a final conversion, which is great. My $0.02: after working in both social and in programmatic, I believe there will be no way for marketers to determine the accuracy of this attribution until all digital advertising platforms find a better way to share anonymized user data.

Where to From Here?
I’ll withhold comment on where we’re heading, but if the recent AdWeek conference is any indication, the points here are not lost on the leaders of the programmatic and digital display advertising side of the industry. For now, we’re extremely happy with the scale, granularity, and dependability of platforms like Facebook in achieving reach and frequency, but we’ll continue to watch both sides closely.