© Copyright 2017
Baker Street Advertising
November 1, 2016
Brian Kelleher

Programmatic Collaboration, What?!?

By:

Big-brand agencies have complained that there is a lack of Cannes-worthy creativity in programmatic and little integration among the creative, media and data operations that go into a modern ad campaign. This is typically due to the agencies separating their departments into specific silos with little collaboration. Holding companies like WPP are missing many opportunities due to a lack of communication between their creative and media agencies.*

Audience Data Visualization

“We sit on top of the data, see all the signals and know how they are being used, but we don’t have direct access to content creators, so we often need to script up the piece of content we need in order to accomplish the goal that the advertiser wants. It would be amazing if the two can go together in a meaningful way. We are constantly reaching out to WPP agency leads directly or through GroupM. But the [collaboration] hasn’t happened significantly yet.” – Tim Bagwell, VP of Xaaxis Labs, a division of GroupM.

Programmatic creative is enriched with some form of data, either retargeting or audience data, and is served with the intent of advertising something the audience have recently shown interest in while web browsing. For example, a user goes onto Zappos, looks up a pair of Nikes and then leaves the site. That person is then retargeted with those same Nike shoes shown in the unit, or possibly similar shoes they may have missed. This tactic can go deeper to the point that when our media is looking to serve an impression on a given site, if the data for this person also shows this person lives in a windy area, we can serve that person a unit for a Nike windbreaker. You can also use these tools to set up automatic tests to optimize toward the best performing creative for that audience. Audiences are formed within a dynamic creative optimizer. The combinations of what we can learn of a person based on their browsing history, and how we can use this to serve them the most appropriate ads at the best time, are virtually limitless.

Dynamic Messaging

Baker Street Advertising, being a nimble, full-service agency has a great advantage here.  Our ability to make adjustments quickly will allow us to swim circles around the larger agencies where such a change will take much more time to implement.

BSA Image

Examples:

If person A is 3 months away from buying a car, we can serve a broad “see your NorCal Honda Dealers now!” message.

If person B is 2 months away from buying a car, and we know they are interested in midsize cars, we can serve them a carousel unit where they can scroll through various offers for Accords, Civics, or Insights.

If we know person C went to a dealership in the past month from their phone’s GPS, we can pair that with their browsing history to serve them a hard-hitting message, like “that Civic you were eyeing is now only $169/month!”

“A really simple observation I had when I moved from media to creative is that they’re two totally separate worlds. Media people are spending their days figuring out optimization and targeting, but it’s done completely separate of the actual idea.” — Adam Cahill, Anagram Founder, a programmatic agency.”**

Data Analysis

Part of the problem does stem from brands not being willing to share their data. While we do experience the issue of our clients being unwilling to share customer data (a very understandable stance considering privacy issues), our partners are able to gather enough information within their pixels that allow targeting to be highly focused. Pairing this targeting with templated creative management tools that will allow us to create dozens of variations of ads with different copy and backgrounds at the flip of a switch, allows us to break down which audiences are responding to each variation. Our teams need to continue to educate each other on what is available within our targeting capabilities, and why a certain message might resonate with a specific audience.

Of course, this route will bring up questions on the cost of producing so many variations for so many audience segments. How niche should we go with targeting and creative before we hit zero return? We need to test, retest, and continue testing audience segments with various creative. Start by finding a new, micro audience that is responding to a more generic message and adjust units to fit that audience. If media is able to generate a new micro audience that is showing promise, creative can then move to create and assign specific creative to that audience (i.e. targeting football moms with a Honda Odyssey unit with a tag line “Space for all your gear and the kids too.” – [sorry, not a copywriter]. This approach will also help us to avoid creative burnout.

BSA already demonstrates excellent collaboration, and as we work towards adopting new creative tools, such as Rich Media and Dynamic Creative, our teamwork, communication, and education across specialties will make our client’s campaigns that much more successful. Pairing data enabled targeting to custom content through dynamic creative is fast becoming the next big tactic, and we are on a great path to become leaders of this approach.

Brian Kelleher is an Associate Media Director at BSA, specializing in digital strategy, planning and buying.
 
* Digiday – “ The Programmatic Divide Still Persists” By: Yuyu Chen http://digiday.com/agencies/gap-programmatic-creative-media-planning-still-exists/

** Digiday – “Programmatic Creative Inches to a Fix” By: Shareen Pathak http://digiday.com/agencies/programmatic-creative-inches-toward-fix/

Analytics, Digital Marketing, Programmatic, Uncategorized

September 5, 2016
Howie Leibach

Can Data Predict “Correct” Advertising?

By:

 
In 2008, Nate Silver predicted Barack Obama as the next president of the United States way before the election took place. While the pundits relied on “gut-feelings,” Nate was skeptical. He rolled up his nerdy sleeves and crunched probabilities that leveraged historical data and micro trends within each municipality — something TV pundits simply weren’t doing.
 
Nate_Silver_2009
Pictured: Nate Silver, Statistician
 
In 2013, Netflix famously predicted “House of Cards” would be a huge hit, way before ever having the green light to produce it. Instead of guessing what show to fund first, Netflix looked at the completion rates of loyal users. Using Venn diagram data, they discovered that the melding of David Fincher’s “The Social Network”  and the British political thrillers was scoring off the charts. The analysis was so strong that they threw $100 million into season 1 — and the rest is history.
 
house-of-cards-kevin-spacey
 
Data is prophesizing BIG things like U.S. Elections and hit TV shows. So you might be wondering: Are we at the point where data can spit out the perfect ad?
 
The answer is, almost.
 
In 2016, IBM created promotional content based on machine learning. With little to no human assistance — the A.I. processed what it considered to be the best message for humans. While thought-provoking, there was a huge problem with it. The ad sucked.
 
This means one thing. Human creativity still matters immensely, and it will for decades. But data is growing stronger, and if used correctly can make a huge difference.
 
In 2016, the BEST creativity isn’t just head turning — it’s action-oriented; inspired by data to work harder. This means stronger segmenting, better ideas, and proof.
 
Baker Street takes this to heart.
 
Like Nate Silver, we embrace historical data for pin-point accuracy.
 
Like Netflix, we predict campaign success (with algorithms and velocity demand) before anyone writes a big check.
 
It’s a rare creative process that gets vetted from start to finish.
 
So while agencies are years away from churning out “the perfect ad,” Baker Street ensures that by coupling data with creativity, it’s more “correct” than ever before.  

Analytics, Data Analytics, Qualitative Research, Quantitative Research, Uncategorized