OI Discussion: Artificial Intelligence in Email Marketing

Jenna Tiffany: 

It's a pleasure to host another OI discussion with a topic that I’m sure all of you will have thoughts on as well as experiences to share. 

Artificial Intelligence in email marketing

So let's start with, have you used AI in your email marketing activities? What did you learn from doing so? 

If you’re yet to utilise AI, what challenges are you facing for implementing? 

For those of you that have already contacted me off list following my OI article, please join in. 

My experience of utilising AI particularly for generating email subject lines has been hugely successful for clients, both from an open rate perspective but also in generating revenue. The vital area from my perspective is that AI can enable marketers to test at scale for example 10+ subjects lines or different types of content and then learn from the results for future campaigns. Where AI has provided the most impact has been where there is a clear strategic need. 

I'm looking forward to what I hope will be an interesting discussion this week and finding out what you think and your experiences to. 

Over to you all!

David Baker: 

I sense that the industry in general is still grappling with "does it really pay off" and if it does pay off, "can it be operationalized to how we work today?  or requires major change process wise?   and if you go deep on AI, does it lock to you a vendor? given the value of algorithms is over time...    We aren't any where near the point of "singularity" (where machines are equivalent to humans), but I do see this explosion of investments in this category impacting the type of work you may do 5 years from now as an email marketer.   i think the term "smart marketer" will take on a new meaning over the next few years.

I wrote about this and covered it to some degree a few months ago on OI.. here's the original article.

Jordie van Rijn: 

It is very good that we focus on getting a better understanding of thepractical applications.

Marketers that are asked about the trends and future of email marketing AI is one of 
the Big Topics for the coming 5 years are: Next to Data, Personalisation, Content, GDPR and Automation. Just have a look at this Word Cloud created from their predictions. Artificial Intelligence (AI) is the hottest topic of all.

jordie image

Looking at it another way, it means the concept is hyped.  

Logically, because the outlook / promise is fantastic. Which until we are over the first hype gives an
opportunity to be more practical than the average marketer. Display a heavy dose of realism not get caught up. Plenty of what we are told is “AI”, is a blackbox – which makes it hard to see what I really going on under the hood. 
Could very well be mostly logic and business rule based.

Tink Taylor: 

There is no doubt this AI and machine learning is the way of the future. These days there are so many data points that can be used, leveraged, tested & displayed etc that it can be over whelming for the ‘human’ marketer to be able to deploy them at their most optimum, enabling marketers to deliver true one to one value delivering the right message, at the right time, to the right person and nowadays in the right channel.

I like David’s point on the term ‘"smart marketer" as this hits the nail on the head on a key issue right now. AI is such a broad term and often spoken about across numerous different industries. Therefore perhaps it’s not clear to present day marketers what they will actually be getting, what it will achieve and what it would take to implement.

There are numerous venders in the mar-tech sector… and how do I put this… some are better at ‘mar’ than they are ‘tech. I’ve seen lots of spin around certain features & technologies that are badged as AI but are they really? For example send time optimization, is this really AI, I’ve seen certain venders claim that this is, but it’s just a new name for the same feature that have been around for a while.

To become a "smart marketer" I think we as an industry would certainly do ourselves justice in defining specifically in our world what we all agree AI is, as perhaps some marketers may become skeptical to the marketing spin and buzz words that are being thrown at them.

Nick Crawford: 

Hi to all – perhaps slightly late to the discussion party, but a topic I looked at in some detail prepping for a DMA UK Email Council debate a few months ago.

The full blog write up is here

What struck me from the comments post event from the audience was the pure confusion around the topic, but with almost a guilt that everyone should know more given the proficient use of the term.

And yes there does seem to be a lot of spin around AI solutions.

For me the stand out vendors are those that use AI as an enabler to the marketer rather than a feature. For example the move from ‘people who bought x also bought y’ to ‘what people like me like’ enabling next best purchase/action on far more than just past transaction. Executing simply the power to use and link multiple data points across multiple moments where our human brains struggle to grasp the enormity of what’s possible (to paraphrase the CTO of Ometria)

Mike Austin: 

Great points Nick, totally agree that the best use of AI at the moment is to enable marketers - to do more, to do better. 

The other point I'd like to make relates to the volumes of data required to make AI work properly. If you have vast quantites of data - like Google, Amazon, eBay and other global platforms, then you can make AI solutions do wonderful things with relatively little human intervention.

However, most companies just don't have those volumes, so the reality is that due to those limited data volumes, most AI solutions require a lot of human intervention from data scientists to (1) clean the data (2) construct and test hypotheses for which data may be related to which outcomes, (3) Identify significant clusters (4) Choose appropriate algorithms to build the models. Once all that's done, you'll be in a position to start to deploy the models.

This means that "general purpose" AI really isn't much use for most companies out there. Companies like ours (and many others) are busy creating models to handle specific applications, usually based on data scientists and aggregated data from tens or hundreds of clients. 

AI is fascinating, and has huge potential, but it's very important to realise that we're only in the foothills of capability and of sophistication. It's very early stages, and it's one of the most exciting areas, but there's a long way for the industry to go. 

Jenna Tiffany: 

Thank you all for joining in this week's discussion, in brief summary, interesting points from the "smart marketer" to the possible need that as an industry we needs to define precisely what we all agree AI is. 

I think we all agreed that AI is exciting and provides a lot of potential but as highlighted there does need to be data and lots of it to reap huge benefits.