For those of you who were at Ad Week in NY a few weeks back, you can probably attest to the fact that Programmatic is not just a buzzword any more, its a resounding theme. This morning, we’re going to talk about why measurement is important for branding in a programmatic world.
To kick off our discussion, I think this video captures the theme for this morning.
I liken that video to programmatic….it’s fascinating, confusing, and can pop out a few surprises along the way.
Do any of you remember what the media space was like before automation? I sure do. media buying was conducted face to face, over phone, over lots of emails, sometimes over lots of drinks (with a lot of spreadsheets emailed in between!). With this, there were limitations in to how quickly you could execute (not just putting together the buy, but then trafficking), how granular you could get, and how effectively you could make improvements to a buy.
Programmatic is the automation of buying and selling media inventory, where data is processed in real time (literally milliseconds) to inform decisions. And while it certainly brings speed, precision and control to the buying process, Programmatic is not perfect, but it has come a long way in a very short amount of time.
And it’s evident by how much we’re seeing brands invest in Programmatic. Sell Platforms have spent the last year+ laying down infrastructures that have now moved beyond the testing phase, giving comfort to Marketers to participate in Programmatic . 2014 is clearly evident of this with eMarketer projecting just over $10B in progrommatic digital display spend by the end of this year! And note, this doesn’t account for video (digital and linear) which is now taking form very quickly.
There are still some reservations about programmatic though. Brand marketers want more transparency and quality controls, such as where the inventory is coming from/where their ads are being placed, based on what criteria, is it brand safe, and what are the real costs (meaning, what is the actual media cost vs. labor or technology cost). They want to understand the players in the ecosystem that may contribute to these costs, i.e. DMPs, DSPs, SSPs, Trading Desks…what are their roles in my media buy, do I need them, what am I paying for their use.
At the end of the day, brand marketers are still looking for proof points that programmatic can deliver on the criteria/metrics brand advertisers have always been looking for.
Though the way media is being bought and sold has changed, advertiser’s fundamental marketing objectives have not changed. Brand marketers still need their ad to reach the right audience, resonate with the audience, and drive them to the desired action (e.g. purchase, viewership). Programmatic’s promise to brand advertisers have largely been in reaching the right audience, but I’m not sure we’re quite there yet in bringing more efficiency around resonance and reaction.
We mentioned on the outset that Programmatic is about automation and machine based learning. The idea that the “right formula” is already built into the algorithm may make measurement unnecessary. However, an algorithm can sometimes be wrong. Because of it’s automated nature, and because of the convoluted chain of players involved, anything that is “off” can replicate itself very quickly, turning small problems into very big ones. Aside from algorithms, human error can enter the scenario with erroneous inputs. And what about data providers? Not every data provider is created equal.
This is where neutral third party measurement comes in, validating a campaign and providing proof points based on the Marketers’ metrics.
Here are some examples of how our customers are using measurement in a programmatic world
Marketers have used solutions like Nielsen Online Campaign Ratings to ensure that their programmatic campaigns are actually meeting their brand objectives, and reaching their target audiences. It’s tempting in a programmatic world to “trust in the algorithm”, but the data sources that the algorithm is based on isn’t always accurate. Platforms, you can use this approach to audit your data sources, and get a better internal understanding of where certain audiences are on your network
Measurement doesn’t just have to be about accountability though. What’s exciting about programmatic platforms is the increased ability to take hundreds of data points to improve campaign delivery. One of those data points can be the measurement itself.
For example, one of our clients, a video advertising platform (TubeMogul), wanted to provide TV comparable metrics to attract more brand advertisers to their video platform.
Solution
Used OCR to measure GRP, age/gender metrics
Used API to pull those metrics into their interface
Run this on everything, allowing them to start to predict, at the placement level, age/gender reach
Built this into their platform to allow
Customers to plan and buy based on GRP
Platform to automatically optimize age/gender audience delivery
Doesn’t make sense to optimize using DR tactics, like clicks…many instances where campaigns optimized using CTR and negatively impacted brand awareness objective.
We saw an instance in which a selling/buying algorithm may not have been built using the most accurate data. Another “watch out” brand marketers have in a programmatic world is to make sure that what they’re focused on is even part of the decisioning factors that inventory is bought and sold with – that it’s even part of the algorithm.
Many different information sources can be used as a “proxy” for brand opinion, but unless this is verified with a brand-focused measurement solution, it won’t be clear if the campaign actually achieved the brand marketer’s objectives.
Measurement doesn’t just have to be about accountability though. What’s exciting about programmatic platforms is the increased ability to take hundreds of data points to improve campaign delivery. One of those data points can be the measurement itself.
For example, one of our clients, an agency trading desk (Annalect) wanted to ensure that their campaigns were consistently achieving their brand KPIs – brand lift.
Solution
Used OBE to measure brand lift
Used API to pull those metrics into the programmatic platform they were working with
Platform auto-optimized campaign delivery, shifting impressions to placements, sites, frequencies, and creatives that were driving higher brand lifts
Platform also pixelled positive brand responders and built a look alike model to deliver the campaign to similar positive brand responders, increasing reach.
This allowed the trading desk to ensure they were driving brand lift (via optimization) at scale (via look alike modelling).
The promise of programmatic is that it is supposed to be more efficient. No matter how advertising is bought, advertisers at the end of the day need to ensure that it’s driving the desired reaction, whether that be sales or tune in, in order to justify continued investment.
Fraud and viewability are obviously concerns for marketers, but if you Close the loop and understand whether the campaign delivered against your objective and how it delivered your objective, you can identify what to replicate and what to change.
Marketers are always looking for opportunities to get the most from their advertising budgets. One of the challenges of successfully executing an integrated campaign across platforms is ensuring budget is not being wasted by reaching the same audience with the same message too many times, across their devices.
CHALLENGE: At the beginning of the busy holiday season, a luxury auto brand wanted to maximize the effective reach of their national TV advertising using digital media to fill in audience gaps. They wanted to minimize the impressions wasted on their A25-54 audience already exposed to the TV campaign 3 or more times.
More than half of TV GRPs were delivered to just 14% of the population, yielding an over-exposure
61% of the target population remained un-exposed or under-exposed
SOLUTION: Use TV viewing behavior data to focus delivery of digital campaign to unexposed and under-exposed audience.
Create custom segments based on audience exposure to TV campaign
Review the TV plan programming details and audience based on Nielsen panel – the networks, genres, shows, day-parts and exposure
Develop custom “TV Extension Audience Segments” – apply look-a-like models based on thousands of characteristics including demographics, technology and location
Deploy digital programmatic video campaign to deliver TV extension
Measure audience
RESULTS: 3% incremental reach from digital advertising.
32% of digital impressions were served to audience never exposed to TV campaign
Frequency Extension to under-exposed audience: 6% reach for audience exposed less than 3 times to TV advertising
What does the future of programmatic look like? Brand marketers will continue to invest in this area. The perception from remnant to premium will resonate more.
Marketing is shifting from screen-focused to consumer-focused, who spans all screens. Both advertising in general and programmatic in particular will do the same. Media budgets are shifting towards multi-screen campaigns. In fact, in a survey conducted with the ANA, by 2016, 50% of an advertisers media spend will go toward integrated multi-screen campaigns.
Data generated in one screen/platform can be used to improve marketing in any other platform/screen, making marketing more relevant and cross platform.
As the media landscape continues to evolve, so will the Programmatic practice