One British Airways ad urged consumers to “escape the commute and discover the Indian Ocean" not long after the missing Malaysian airline MH370 was rumoured to have crashed there (Source: Getty)
Automation has transformed the marketing industry. Last year, programmatic trading reached an all-time high, making up 60 per cent of the £3bn of transactions around UK display advertising, according to the Internet Advertising Bureau.
Programmatic can be more efficient and cost-effective method for both buyers and sellers of online advertising. But it has its limits. As with any form of automation, programmatic trading is only as useful as the humans behind the wheel.
This is easier said than done when marketers are under increasing pressure to deliver measurable results and in a shorter period of time.
But a number of infamous cases illustrate how algorithms can produce messaging which is accidentally offensive. Microsoft’s chatbot Tay was issuing racist and sexist messages shortly after its launch earlier this year, causing the company much embarrassment on social media.
Solid Gold Bomb, a T-shirt making business founded in Melbourne, set up an algorithm to generate thousands of variations on the “Keep Calm and Carry On” slogan for its T-shirts. Unbeknownst to the firm, one slogan encouraging rape had been published. When this came to the public’s attention in 2013, Solid Gold Bomb went bust.
It is universally accepted that retargeting customers with products they expressed an interest in makes sense. It is also accepted that spamming those customers with the same message (sometimes after they’ve actually bought the product) causes irritation.
But machines have no idea about these things. They lack the ability to predict and detect subtle nuances and changes in circumstance or opportunity which might affect how someone could respond to an ad. When seasonality, fraud, or sociopolitical events threaten to change the context of a message or call to action, it’s up to us to bring the technology up to speed.
Remember the US presidential ads, which ran alongside Isis propaganda on YouTube? Or when one of British Airways’s ads urged consumers to “escape the commute and discover the Indian Ocean" not long after the missing Malaysian airline MH370 was rumoured to have crashed there? The company subsequently apologised and explained that it was "pre-scheduled content".
This is what can happen when algorithms go AWOL. For a brand, it can be unexpectedly damaging.
What tech can accomplish in hyper-local targeting and optimisation. But humans must still lead with their knowledge of language, ideas and originality.
It’s easy to spot the difference between an ad “designed” solely by the tech, and one that has been collaboratively developed through both artificial intelligence and human insight. Machines don’t produce great ideas; this is the job of your creative team or agency.
Read more: AppNexus CEO Brian O'Kelley lays out his vision for the programmable age
If successful automation is your goal, the answer is to create teams with different skill sets. The right mix is one that boasts amazing creative people and data analysts that can collaborate on the end solution, along with a dedicated and knowledgeable ad ops team.
Machines and algorithms have a crucial role to play, but we still need skilled drivers to win in this race.
Programmatic can be more efficient and cost-effective method for both buyers and sellers of online advertising. But it has its limits. As with any form of automation, programmatic trading is only as useful as the humans behind the wheel.
The problems with automation
Consumers are hard to follow. They jump from one channel to another, and in the struggle to keep up, more marketers are turning to automation. What many are yet to recognise, however, is that while machines respond well to patterns, human behaviour is not as predictable. Programmatic may provide marketers with greater scale, but they must not sacrifice quality for quantity in terms of the eyes their ads are reaching.This is easier said than done when marketers are under increasing pressure to deliver measurable results and in a shorter period of time.
But a number of infamous cases illustrate how algorithms can produce messaging which is accidentally offensive. Microsoft’s chatbot Tay was issuing racist and sexist messages shortly after its launch earlier this year, causing the company much embarrassment on social media.
Solid Gold Bomb, a T-shirt making business founded in Melbourne, set up an algorithm to generate thousands of variations on the “Keep Calm and Carry On” slogan for its T-shirts. Unbeknownst to the firm, one slogan encouraging rape had been published. When this came to the public’s attention in 2013, Solid Gold Bomb went bust.
Garbage in = garbage out
The information you put into a machine dictates what you get out. If there are errors or gaps in the data and instructions you input, the results will reflect this. A human-led strategy is the only way to give machines the information they need to work at optimum performance.It is universally accepted that retargeting customers with products they expressed an interest in makes sense. It is also accepted that spamming those customers with the same message (sometimes after they’ve actually bought the product) causes irritation.
But machines have no idea about these things. They lack the ability to predict and detect subtle nuances and changes in circumstance or opportunity which might affect how someone could respond to an ad. When seasonality, fraud, or sociopolitical events threaten to change the context of a message or call to action, it’s up to us to bring the technology up to speed.
Remember the US presidential ads, which ran alongside Isis propaganda on YouTube? Or when one of British Airways’s ads urged consumers to “escape the commute and discover the Indian Ocean" not long after the missing Malaysian airline MH370 was rumoured to have crashed there? The company subsequently apologised and explained that it was "pre-scheduled content".
This is what can happen when algorithms go AWOL. For a brand, it can be unexpectedly damaging.
Creative context
Automation provides efficiency, but this shouldn’t come at the expense of creativity. Programmatic advertising will help identify and precisely target a chosen audience, but if creatives forgo the human touch, ads won’t perform as well as they could.What tech can accomplish in hyper-local targeting and optimisation. But humans must still lead with their knowledge of language, ideas and originality.
It’s easy to spot the difference between an ad “designed” solely by the tech, and one that has been collaboratively developed through both artificial intelligence and human insight. Machines don’t produce great ideas; this is the job of your creative team or agency.
Read more: AppNexus CEO Brian O'Kelley lays out his vision for the programmable age
If successful automation is your goal, the answer is to create teams with different skill sets. The right mix is one that boasts amazing creative people and data analysts that can collaborate on the end solution, along with a dedicated and knowledgeable ad ops team.
Machines and algorithms have a crucial role to play, but we still need skilled drivers to win in this race.
Automation has transformed the marketing industry. Last year, programmatic trading reached an all-time high, making up 60 per cent of the £3bn of transactions around UK display advertising, according to the Internet Advertising Bureau.
Programmatic can be more efficient and cost-effective method for both buyers and sellers of online advertising. But it has its limits. As with any form of automation, programmatic trading is only as useful as the humans behind the wheel.
Read more: Why publishers need to keep pace with programmatic
This is easier said than done when marketers are under increasing pressure to deliver measurable results and in a shorter period of time.
But a number of infamous cases illustrate how algorithms can produce messaging which is accidentally offensive. Microsoft’s chatbot Tay was issuing racist and sexist messages shortly after its launch earlier this year, causing the company much embarrassment on social media.
Solid Gold Bomb, a T-shirt making business founded in Melbourne, set up an algorithm to generate thousands of variations on the “Keep Calm and Carry On” slogan for its T-shirts. Unbeknownst to the firm, one slogan encouraging rape had been published. When this came to the public’s attention in 2013, Solid Gold Bomb went bust.
It is universally accepted that retargeting customers with products they expressed an interest in makes sense. It is also accepted that spamming those customers with the same message (sometimes after they’ve actually bought the product) causes irritation.
But machines have no idea about these things. They lack the ability to predict and detect subtle nuances and changes in circumstance or opportunity which might affect how someone could respond to an ad. When seasonality, fraud, or sociopolitical events threaten to change the context of a message or call to action, it’s up to us to bring the technology up to speed.
Remember the US presidential ads, which ran alongside Isis propaganda on YouTube? Or when one of British Airways’s ads urged consumers to “escape the commute and discover the Indian Ocean" not long after the missing Malaysian airline MH370 was rumoured to have crashed there? The company subsequently apologised and explained that it was "pre-scheduled content".
This is what can happen when algorithms go AWOL. For a brand, it can be unexpectedly damaging.
What tech can accomplish in hyper-local targeting and optimisation. But humans must still lead with their knowledge of language, ideas and originality.
It’s easy to spot the difference between an ad “designed” solely by the tech, and one that has been collaboratively developed through both artificial intelligence and human insight. Machines don’t produce great ideas; this is the job of your creative team or agency.
If successful automation is your goal, the answer is to create teams with different skill sets. The right mix is one that boasts amazing creative people and data analysts that can collaborate on the end solution, along with a dedicated and knowledgeable ad ops team.
Machines and algorithms have a crucial role to play, but we still need skilled drivers to win in this race.
Programmatic can be more efficient and cost-effective method for both buyers and sellers of online advertising. But it has its limits. As with any form of automation, programmatic trading is only as useful as the humans behind the wheel.
The problems with automation
Consumers are hard to follow. They jump from one channel to another, and in the struggle to keep up, more marketers are turning to automation. What many are yet to recognise, however, is that while machines respond well to patterns, human behaviour is not as predictable. Programmatic may provide marketers with greater scale, but they must not sacrifice quality for quantity in terms of the eyes their ads are reaching.Read more: Why publishers need to keep pace with programmatic
This is easier said than done when marketers are under increasing pressure to deliver measurable results and in a shorter period of time.
But a number of infamous cases illustrate how algorithms can produce messaging which is accidentally offensive. Microsoft’s chatbot Tay was issuing racist and sexist messages shortly after its launch earlier this year, causing the company much embarrassment on social media.
Solid Gold Bomb, a T-shirt making business founded in Melbourne, set up an algorithm to generate thousands of variations on the “Keep Calm and Carry On” slogan for its T-shirts. Unbeknownst to the firm, one slogan encouraging rape had been published. When this came to the public’s attention in 2013, Solid Gold Bomb went bust.
Garbage in = garbage out
The information you put into a machine dictates what you get out. If there are errors or gaps in the data and instructions you input, the results will reflect this. A human-led strategy is the only way to give machines the information they need to work at optimum performance.It is universally accepted that retargeting customers with products they expressed an interest in makes sense. It is also accepted that spamming those customers with the same message (sometimes after they’ve actually bought the product) causes irritation.
But machines have no idea about these things. They lack the ability to predict and detect subtle nuances and changes in circumstance or opportunity which might affect how someone could respond to an ad. When seasonality, fraud, or sociopolitical events threaten to change the context of a message or call to action, it’s up to us to bring the technology up to speed.
Remember the US presidential ads, which ran alongside Isis propaganda on YouTube? Or when one of British Airways’s ads urged consumers to “escape the commute and discover the Indian Ocean" not long after the missing Malaysian airline MH370 was rumoured to have crashed there? The company subsequently apologised and explained that it was "pre-scheduled content".
This is what can happen when algorithms go AWOL. For a brand, it can be unexpectedly damaging.
Creative context
Automation provides efficiency, but this shouldn’t come at the expense of creativity. Programmatic advertising will help identify and precisely target a chosen audience, but if creatives forgo the human touch, ads won’t perform as well as they could.What tech can accomplish in hyper-local targeting and optimisation. But humans must still lead with their knowledge of language, ideas and originality.
It’s easy to spot the difference between an ad “designed” solely by the tech, and one that has been collaboratively developed through both artificial intelligence and human insight. Machines don’t produce great ideas; this is the job of your creative team or agency.
If successful automation is your goal, the answer is to create teams with different skill sets. The right mix is one that boasts amazing creative people and data analysts that can collaborate on the end solution, along with a dedicated and knowledgeable ad ops team.
Machines and algorithms have a crucial role to play, but we still need skilled drivers to win in this race.
No comments:
Post a Comment