6. Beyond the Tipping Point…
Doing Programmatic Not Doing Programmatic
93%
7%
Agencies
87%
13%
Advertisers
Source: Attitudes towards Programmatic Advertising survey , IAB Europe, 2016
7. If you don’t have a
competitive edge,
don’t compete
Jack Welch
7 AppNexus Inc. - Confidential – Not for Distribution
CEO GE 1988 - 2001
8. AppNexus Inc. - Confidential – Not for Distribution8
How will you differentiate?
9. AppNexus Inc. - Confidential – Not for Distribution9
from Programmatic
to Programmable
10.
11. AppNexus Inc. - Confidential – Not for Distribution11
While on the train
For Social For Messaging
Whilst watching TV
NATIVEBANNER VIDEO INTERSTITIAL RICH MEDIA
Handsets Data Mobile Ad Formats Cell Coverage
20. example
sales funnel
< 1 hour ago
1-6 hours ago
6-24 hours ago
1-3 days ago
4-7 days ago
7-14 days ago
14-30 days ago
landing page
< 1 hour ago
1-6 hours ago
6-24 hours ago
1-3 days ago
4-7 days ago
7-14 days ago
14-30 days ago
product page A
< 1 hour ago
1-6 hours ago
6-24 hours ago
1-3 days ago
4-7 days ago
7-14 days ago
14-30 days ago
product page B
< 1 hour ago
1-6 hours ago
6-24 hours ago
1-3 days ago
4-7 days ago
7-14 days ago
14-30 days ago
product page C
< 1 hour ago
1-6 hours ago
6-24 hours ago
1-3 days ago
4-7 days ago
7-14 days ago
14-30 days ago
new user
21. results
The APB campaign
delivered 6.5 times
more specific
targeting strategy
The APB campaign
needed 20% less
time on campaign
management
The APB campaign
delivered a 35%
increase in
performance (eCPA)
Last year at this conference, I talked about the need to simplify programmatic, how technology is making peoples lives harder, not easier;
Our industry is already facing enough challenges today, many or all of which you could argue that technology has contributed to and you can’t go to a media conference today with references to the challenges facing pubs, marketers, agencies.
I’ve just observed a great many companies, globally and here in Australia, who are doing the exact same things, licensing the exact same pieces of commodity technology and expecting some kind of amazing result, when in reality the tech has basically stuck them in neutral. I think some people have been waiting for ad tech itself to do the hard work, like there’s some magic button that is going to stem your share losses to Google or Facebook, or address your agency’s fundamental problems creating value for your client. But there is no magic button. And a lot of you are NOT winning. So in all seriousness, what’s next?
Recap – you need to understand where we have come from to get here today, lots of confusion but markets growing fast
Reference – Now we have reached the tipping point, programmatic now accounts for more than 50% of display, and search and social have always been traded programmatically
Programmatic adoption has reached the tipping point because of those many benefits we’ve just talked about, efficiency, scalability. But all this growth combined with a lack of education and understanding, has brought with it major challenges; Brands realized there had been a shift that they did not understand and started asking tough questions that the black box model could not answer – where is my budget going, why are supply chain costs so high, who is taking a margin and where, how did my brand appear on a torrent or porn site, how do I know my ad was viewable by a real person and not a bot; Consumers took action to manage their ad experience by experimenting with adblockers and finding not only did they like the ad-free experience but that their first screen – the mobile, was faster and cheaper without ads.
So how is this relevant to mobile
So let’s get into it with some context and controversy. First off, the ad-tech boom is over. The venture capital money hose that fuelled the last decade or so of ad tech has run dry. There are too many companies out there fighting for the same dollars with undifferentiated products and confusing everyone in the process.
Let’s start with product. Product is the “what” – the media package, the data, your service.
If you’re a good product manager, you might ask yourself “how can I deliver a product of superior quality that marginalises competition?”
After product, the second area you can grab an edge is process. Process refers to the “how” – how you build your product, how you deliver it. Now I have to tell you something about this graphic. One of our brilliant Implementation guys pointed it out to me: these gears will never turn. But it looks good, right?
I know process seems boring, right? Process is the stuff that gets introduced when you become a big company. It’s like a TPS report, if you’ve seen Office Space. But what I’m talking about here is process that can give you a sharp edge that makes you better than the other guys.
The last but far from least is PEOPLE, the “who” behind it all.
It’s a point that I make time and time again: invest in your people. Nothing is possible without skilled, passionate, amazing people.
Now there’s a very common misconception that technology is there to replace people. I don’t think that’s true, but I do fundamentally believe that the skills of the people inside tomorrow’s winning publisher, agency, and marketer need to fundamentally change so that we can put the machines to work for us. And sometimes that means looking in new places for your talent.
I’m not going to go into detail here, but this is a science problem that has now been solved. With advances now and throughout this year, we are going move from the programmatic era of automation and commoditisation, where you all work for the machines, to the programmable era and a return to differentiation, where the machines work for you.
The next generation DSP won’t be the commoditised, media vending machine, one-size-fits-all Demand Side Platform. It’ll be a Data Science Platform. It’s a programmable marketing machine that is unique to every brand. And smart marketers, agencies, and publishers who adapt from programmatic to programmable, and who hire people like data scientists to put the machines to work, will win.
And when we think about where the puck is going – beyond the first screen to a range of screens as we enter what Gartner described as the Programmable Economy but you may know better as the Internet of Things, it doesn't make any sense for advertising to be commoditized.
The mobile user is operating in a very different context from the desktop browser of 2009,
focused more intently on a specific task whether that’s booking a car, finding a restaurant, researching a purchase, sharing photos with friends or simply killing time. Difference in user experience – use it while watching tv, on the train, use it for social, messaging
The difference in and the importance in consumption - handset types, coverage, data packages, formats
Marketing in these moments should be different for each of the brands on this slide – yet in a programmatic world they would typically all operate on a black-box DSP that offers standardized functionality and the same algorithms behind a limited set of optimisation tools. These tools are enormously effective at buying media at low cost and have made the ad buying process itself much more efficient than it was but as brands seek to be relevant and useful across thousands of devic types, a plethora of media and a world of apps, sites and media owners today’s buyers are starting to push at the boundaries of what is possible with DSPs designed for simplicity and efficiency. That is why most buyers today are using multiple DSPs (video, mobile, regional etc) which creates complexity and destroy agility as well as reinflating costs. Performance for any one brand is limited by the functionality offered to all brands and when everyone uses the same tools in the same way its difficult for anyone to create a competitive advantage.
While the industry has made huge strides in eliminating bad actors and benchmarking for viewability, the black-box remains a fundamental block to buyers having genuine control and transparency.
What does Bannerconnect do, for those that don’t know us
Execution and insights for brands that are looking to take control over programmatic, whether it be via in-housing or outsourcing.
Storyline
What does programmable mean to us
Programmatic = using commodity intelligence built for you
Programmable = adding your own unique intelligence
An algorithm is a list of rules to follow in order to solve a problem.
Algorithms need to have their steps in the right order. Think about an algorithm for getting dressed in the morning. What if you put on your coat before your jumper? Your jumper would be on top of your coat and that would be silly! When you write an algorithm the order of the instructions is very important.
How do we use it / how does it change our business
People | skillset | execution -> Lean Six Sigma, data analyst (querying)
People | skillset | insight -> data scientist and engineers
Data science -> beat us in chess, beat is in go, beat our best fighter pilots, will beat our traders
Coding in R
What does it mean for our clients (slide 6+7 from SummitNYC)
Results: faster, deeper targeting, better results
Adding 20/20 research into the bid logic
Exposure time
Storyline
What does programmable mean to us
Programmatic = using commodity intelligence built for you
Programmable = adding your own unique intelligence
An algorithm is a list of rules to follow in order to solve a problem.
Algorithms need to have their steps in the right order. Think about an algorithm for getting dressed in the morning. What if you put on your coat before your jumper? Your jumper would be on top of your coat and that would be silly! When you write an algorithm the order of the instructions is very important.
Storyline
What does programmable mean to us
Programmatic = using commodity intelligence built for you
Programmable = adding your own unique intelligence
An algorithm is a list of rules to follow in order to solve a problem.
Algorithms need to have their steps in the right order. Think about an algorithm for getting dressed in the morning. What if you put on your coat before your jumper? Your jumper would be on top of your coat and that would be silly! When you write an algorithm the order of the instructions is very important.
How do we use it / how does it change our business
People | skillset | execution -> Lean Six Sigma, data analyst (querying)
People | skillset | insight -> data scientist and engineers
Data science -> beat us in chess, beat is in go, beat our best fighter pilots, will beat our traders
Coding in R
What does it mean for our clients (slide 6+7 from SummitNYC)
Results: faster, deeper targeting, better results
Adding 20/20 research into the bid logic
Exposure time
Storyline
What does programmable mean to us
Programmatic = using commodity intelligence built for you
Programmable = adding your own unique intelligence
An algorithm is a list of rules to follow in order to solve a problem.
Algorithms need to have their steps in the right order. Think about an algorithm for getting dressed in the morning. What if you put on your coat before your jumper? Your jumper would be on top of your coat and that would be silly! When you write an algorithm the order of the instructions is very important.
How do we use it / how does it change our business
People | skillset | execution -> Lean Six Sigma, data analyst (querying)
People | skillset | insight -> data scientist and engineers
Data science -> beat us in chess, beat is in go, beat our best fighter pilots, will beat our traders
Coding in R
What does it mean for our clients (slide 6+7 from SummitNYC)
Results: faster, deeper targeting, better results
Adding 20/20 research into the bid logic
Exposure time
Storyline
What does programmable mean to us
Programmatic = using commodity intelligence built for you
Programmable = adding your own unique intelligence
An algorithm is a list of rules to follow in order to solve a problem.
Algorithms need to have their steps in the right order. Think about an algorithm for getting dressed in the morning. What if you put on your coat before your jumper? Your jumper would be on top of your coat and that would be silly! When you write an algorithm the order of the instructions is very important.
How do we use it / how does it change our business
People | skillset | execution -> Lean Six Sigma, data analyst (querying)
People | skillset | insight -> data scientist and engineers
Data science -> beat us in chess, beat is in go, beat our best fighter pilots, will beat our traders
Coding in R
What does it mean for our clients (slide 6+7 from SummitNYC)
Results: faster, deeper targeting, better results
Adding 20/20 research into the bid logic
Exposure time
Storyline
What does programmable mean to us
Programmatic = using commodity intelligence built for you
Programmable = adding your own unique intelligence
An algorithm is a list of rules to follow in order to solve a problem.
Algorithms need to have their steps in the right order. Think about an algorithm for getting dressed in the morning. What if you put on your coat before your jumper? Your jumper would be on top of your coat and that would be silly! When you write an algorithm the order of the instructions is very important.
How do we use it / how does it change our business
People | skillset | execution -> Lean Six Sigma, data analyst (querying)
People | skillset | insight -> data scientist and engineers
Data science -> beat us in chess, beat is in go, beat our best fighter pilots, will beat our traders
Coding in R
What does it mean for our clients (slide 6+7 from SummitNYC)
Results: faster, deeper targeting, better results
Adding 20/20 research into the bid logic
Exposure time
Storyline
What does programmable mean to us
Programmatic = using commodity intelligence built for you
Programmable = adding your own unique intelligence
An algorithm is a list of rules to follow in order to solve a problem.
Algorithms need to have their steps in the right order. Think about an algorithm for getting dressed in the morning. What if you put on your coat before your jumper? Your jumper would be on top of your coat and that would be silly! When you write an algorithm the order of the instructions is very important.
How do we use it / how does it change our business
People | skillset | execution -> Lean Six Sigma, data analyst (querying)
People | skillset | insight -> data scientist and engineers
Data science -> beat us in chess, beat is in go, beat our best fighter pilots, will beat our traders
Coding in R
What does it mean for our clients (slide 6+7 from SummitNYC)
Results: faster, deeper targeting, better results
Adding 20/20 research into the bid logic
Exposure time
Storyline
What does programmable mean to us
Programmatic = using commodity intelligence built for you
Programmable = adding your own unique intelligence
An algorithm is a list of rules to follow in order to solve a problem.
Algorithms need to have their steps in the right order. Think about an algorithm for getting dressed in the morning. What if you put on your coat before your jumper? Your jumper would be on top of your coat and that would be silly! When you write an algorithm the order of the instructions is very important.
How do we use it / how does it change our business
People | skillset | execution -> Lean Six Sigma, data analyst (querying)
People | skillset | insight -> data scientist and engineers
Data science -> beat us in chess, beat is in go, beat our best fighter pilots, will beat our traders
Coding in R
What does it mean for our clients (slide 6+7 from SummitNYC)
Results: faster, deeper targeting, better results
Adding 20/20 research into the bid logic
Exposure time
http://www.bannerconnect.net/exposure-time/
http://www.bannerconnect.net/20-20/
http://www.bannerconnect.net/2020-market-research/
The greatest example is our 2020 research product, whereas we fuel insights on optimal exposure time.
Exposure time is our time-based digital advertising metric that evaluates at how long a campaign needs to be in-view to a consumer for it to be effective. This proxy for consumer attention is not new to advertising, but with Bannerconnect you can see how long a unique user is exposed to your advertising across several touchpoints. Viewability is a hot topic that’s crucial to consider in ensuring your ads are seen. Exposure time is the evolution of viewability. As well as being in-view, the consumer should be exposed to the ad for a certain amount of seconds for best impact
http://www.emerce.nl/research/exposure-time-behoefte-nieuwe-metric
https://www.exchangewire.com/blog/2016/02/01/exposure-time-a-new-standard-for-measuring-digital-effectiveness/
Image 2 (top-right): AlphaGo A.I. beats world’s best Go player
Image 3 (bottom-right): New A.I. beats tactical experts in combat simulation