Marketers, you might be starting to get a bit fed up of seeing the acronym “AI” in the headlines of your go-to industry publications.
We get it. It’s everywhere. But we’re going to keep talking about it.
Why?
Because AI is disruptive. It’s changing how the majority of industries across the globe operate. And it’s even threatening to make certain job roles obsolete.
Is ‘ecommerce marketer’ one of those jobs? Not quite.
But it is inevitable that things will change: much of it for the better (think less admin, more creativity).
Human vs machine – which wins?
In this post, we’ll begin by exploring the decisions that marketers can trust a machine to make, before taking a look at those which—now more than ever—need the human touch.
Machine
As we know, the “I” in “AI” stands for “intelligence”. And it’s true: machines can be intelligent. Very intelligent. But in a different way to human beings.
Machines are able to process information much quicker than us, and automate laborious tasks that would take any homosapien hours.
This puts them in a strong position to take on the following ecommerce marketing tasks:
Customer insight
AI-powered algorithms, through unsupervised and responsive learning, can independently process large sets of data and spot patterns and/or similarities amongst customers that a marketer might easily miss.
It’s almost pointless for a data scientist or marketer to spend time manually going through large data sets if a computer can do it much much faster.
Segmentation and taste profiling
Through clustering processes, an unsupervised learning algorithm can then group customers together – based on their shared traits – to create very focused segments.
An algorithm can also use all of the various data sets available to create (or refine) a single customer view (i.e. a consolidation of all of the information available on a customer into a single record).
This in turn can be used to execute accurate taste profiling and predictive marketing; for example, knowing that customer A is most likely to be interested in category [X] (let’s say ‘trainers’, and therefore sending a message featuring a new trainer launch to boost engagement.
Cross-channel execution
When it comes to cross-channel campaigns, the frequency, discount and channel can all be decided via a machine-learning algorithm using a feedback loop (which basically means perpetually learning from past actions).
Monitoring (and optimising) campaign performance
The number of messages being sent to each customer will also fall under the “machine” category, as an artificial intelligence model can efficiently ensure a brand sends out the optimum number of messages (with the right frequency, via the right channel) to each customer—and that each message compliments, not contradicts, the other.
This brings to question whether or not A/B testing will continue to exist in an industry where campaign performance and optimisation is taken care of.
Human
It’s fair to say, then, that the desired skillset of an ecommerce marketer is set to alter as a result of AI.
As mentioned in the introduction to this post, most administrative and manual tasks are likely to be wiped from the future marketer’s job spec, making room for them to focus more on creativity, customer service and strategy.
Listed below are the ecommerce marketing decisions that will continue to fall under the remit of humans.
Campaign creation
Aspects of campaign creation will still be decided by marketers, not algorithms. For example:
The “look” and “feel” of a campaign | While an algorithm can be used to optimise which images are most effective at getting which customers to convert, the actual creative direction of those images will still be up to a human.
Why?
A human brain is more likely to understand what will engage a fellow human brain, both creatively and emotionally.
Copy and tone
Until someone invents an algorithm that can write like a human and adapt to brand tone of voice (a prospect that is, admittedly, scarily close), humans will remain in charge of copy.
Customer experience
Whilst there are numerous ways artificial intelligence will be able to contribute towards improving a brand’s customer experience, the overall vision for how a customer will be treated by a brand will still fall within the hands of the ecommerce marketer (at least for the next decade or so).
From fun content to customer service to in-store events, it will still be up to a marketer to figure out creative ways to ensure a customer’s journey is as smooth and enjoyable as possible.
Strategy
The automation of short-term administrative tasks will give marketers the chance to start making important decisions about a brand’s long-term strategy. The amount of industry knowledge and first-hand experience needed to look this far ahead this transcends the capabilities of an algorithm—at least for now.
Human & Machine
As you’ve probably guessed, there are definite overlaps between man and machine when it comes to delegating decision making in ecommerce marketing. For example, in order for a brand to even get started with artificial intelligence, a marketer will need to:
Make sure customer data is actually available
The sort of datasets an intelligence layer will need at its fingertips before it can start mining for insights includes:
Behavioural, online, offline, marketing and email interaction, social interaction, demographic, order, customer service, device, product.
Invest in the right technology stack
Ideally, you’ll be looking for a platform that can:
Process all of your data in one place
Create a single customer view for each customer
Create predictive customer profiles (in real-time)
Action this information across multiple channels
Revamp the company’s organisational structure
With a tech stack similar to the above, everyone working within an ecommerce marketing department will be able to log into the same platform and access the same data. This will bring siloed teams closer together, making it much easier for them to carry out seamless cross-channel marketing campaigns.
Keen to learn more about AI and ecommerce marketing? This October, we’ll be hosting our annual Lifecycle conference at 8 Northumberland Avenue, London. With a host of industry-leading speakers, live case studies and in-depth analysis of the latest retention marketing strategies, you won’t want to miss out. Details & ticket information here: http://lifecycle.ometria.com/2017/
Ivan Mazour is chief executive and founder of Ometria
You are in: Home » Guest Comment » GUEST COMMENT What ecommerce marketing tasks can we trust machines with?
GUEST COMMENT What ecommerce marketing tasks can we trust machines with?
by Ivan Mazour
This is an archived article - we have removed images and other assets but have left the text unchanged for your reference
Marketers, you might be starting to get a bit fed up of seeing the acronym “AI” in the headlines of your go-to industry publications.
We get it. It’s everywhere. But we’re going to keep talking about it.
Why?
Because AI is disruptive. It’s changing how the majority of industries across the globe operate. And it’s even threatening to make certain job roles obsolete.
Is ‘ecommerce marketer’ one of those jobs? Not quite.
But it is inevitable that things will change: much of it for the better (think less admin, more creativity).
Human vs machine – which wins?
In this post, we’ll begin by exploring the decisions that marketers can trust a machine to make, before taking a look at those which—now more than ever—need the human touch.
Machine
As we know, the “I” in “AI” stands for “intelligence”. And it’s true: machines can be intelligent. Very intelligent. But in a different way to human beings.
Machines are able to process information much quicker than us, and automate laborious tasks that would take any homosapien hours.
This puts them in a strong position to take on the following ecommerce marketing tasks:
Customer insight
AI-powered algorithms, through unsupervised and responsive learning, can independently process large sets of data and spot patterns and/or similarities amongst customers that a marketer might easily miss.
It’s almost pointless for a data scientist or marketer to spend time manually going through large data sets if a computer can do it much much faster.
Segmentation and taste profiling
Through clustering processes, an unsupervised learning algorithm can then group customers together – based on their shared traits – to create very focused segments.
An algorithm can also use all of the various data sets available to create (or refine) a single customer view (i.e. a consolidation of all of the information available on a customer into a single record).
This in turn can be used to execute accurate taste profiling and predictive marketing; for example, knowing that customer A is most likely to be interested in category [X] (let’s say ‘trainers’, and therefore sending a message featuring a new trainer launch to boost engagement.
Cross-channel execution
When it comes to cross-channel campaigns, the frequency, discount and channel can all be decided via a machine-learning algorithm using a feedback loop (which basically means perpetually learning from past actions).
Monitoring (and optimising) campaign performance
The number of messages being sent to each customer will also fall under the “machine” category, as an artificial intelligence model can efficiently ensure a brand sends out the optimum number of messages (with the right frequency, via the right channel) to each customer—and that each message compliments, not contradicts, the other.
This brings to question whether or not A/B testing will continue to exist in an industry where campaign performance and optimisation is taken care of.
Human
It’s fair to say, then, that the desired skillset of an ecommerce marketer is set to alter as a result of AI.
As mentioned in the introduction to this post, most administrative and manual tasks are likely to be wiped from the future marketer’s job spec, making room for them to focus more on creativity, customer service and strategy.
Listed below are the ecommerce marketing decisions that will continue to fall under the remit of humans.
Campaign creation
Aspects of campaign creation will still be decided by marketers, not algorithms. For example:
The “look” and “feel” of a campaign | While an algorithm can be used to optimise which images are most effective at getting which customers to convert, the actual creative direction of those images will still be up to a human.
Why?
A human brain is more likely to understand what will engage a fellow human brain, both creatively and emotionally.
Copy and tone
Until someone invents an algorithm that can write like a human and adapt to brand tone of voice (a prospect that is, admittedly, scarily close), humans will remain in charge of copy.
Customer experience
Whilst there are numerous ways artificial intelligence will be able to contribute towards improving a brand’s customer experience, the overall vision for how a customer will be treated by a brand will still fall within the hands of the ecommerce marketer (at least for the next decade or so).
From fun content to customer service to in-store events, it will still be up to a marketer to figure out creative ways to ensure a customer’s journey is as smooth and enjoyable as possible.
Strategy
The automation of short-term administrative tasks will give marketers the chance to start making important decisions about a brand’s long-term strategy. The amount of industry knowledge and first-hand experience needed to look this far ahead this transcends the capabilities of an algorithm—at least for now.
Human & Machine
As you’ve probably guessed, there are definite overlaps between man and machine when it comes to delegating decision making in ecommerce marketing. For example, in order for a brand to even get started with artificial intelligence, a marketer will need to:
Make sure customer data is actually available
The sort of datasets an intelligence layer will need at its fingertips before it can start mining for insights includes:
Behavioural, online, offline, marketing and email interaction, social interaction, demographic, order, customer service, device, product.
Invest in the right technology stack
Ideally, you’ll be looking for a platform that can:
Process all of your data in one place
Create a single customer view for each customer
Create predictive customer profiles (in real-time)
Action this information across multiple channels
Revamp the company’s organisational structure
With a tech stack similar to the above, everyone working within an ecommerce marketing department will be able to log into the same platform and access the same data. This will bring siloed teams closer together, making it much easier for them to carry out seamless cross-channel marketing campaigns.
Keen to learn more about AI and ecommerce marketing? This October, we’ll be hosting our annual Lifecycle conference at 8 Northumberland Avenue, London. With a host of industry-leading speakers, live case studies and in-depth analysis of the latest retention marketing strategies, you won’t want to miss out. Details & ticket information here: http://lifecycle.ometria.com/2017/
Ivan Mazour is chief executive and founder of Ometria
Read More
You may also like
Register for Newsletter
Receive 3 newsletters per week
Gain access to all Top500 research
Personalise your experience on IR.net