Fatboy Slim released a track in 2010 entitled ‘Machines Can Do The Work.’ The next line of the song continued with, ‘People can think.‘ For me, that sums up the future of AI and MA perfectly. I have no doubt that right now, some very smart people are probably trying to visualise advanced data models that link together the how, what, why, when and where of pre-purchase behaviour and are trying to underpin that with a subtle layer of nudge theory. My brain begins to melt at the thought. But as the future possibilities unwind before us, the one thing we can be certain of is that AI is coming to MA. In the near future, it seems that if we can think of it – then a machine will be able to do it.
Many will shout ‘hype’ and most of what I read today about AI I would tend to agree. Nothing administers a little fear into the world quite like a story about machines taking over the humans’ role on earth. But scaremongering aside, AI is already weaving its way into our world and is rapidly accelerating thanks to its scientific application to data. Netflix, Facebook, Google and just about any corporation that wants to remain relevant is knee-deep in AI. Why? Because however smart your staff may claim to be, they could never match the pace of a machine running an algorithm for finding useful data insights.
A Brief History of AI
Artificial Intelligence is based on the idea that you can process human thought in a mechanised way. That is a fairly simple concept to grasp, but a baffling theory to put into practice. We make decisions based on context, emotions, beliefs and many more random inputs all being processed through a variable of 100 trillion neural connections. Therefore, the dreams intellectual philosophers had of coding the human mind remain unfulfilled. Most AI since the 1940’s has either been high energy progress or long winters of bugger-all. The real progress made by AI (and it is very much a work in progress) is computers communicating in more natural language. However, Arthur C. Clarke’s HAL 9000, which was imagined for a 2001 world where a machine matched a human’s intelligence, is quite a way off. What is exciting, and is happening in the now, is the capability of AI to enhance the data for MA.
Bringing AI to MA
So, it is definitely worth digging deeper at when it comes to AI and MA. Us marketers need to be ready to adopt and adapt, so we can deliver better value to our clients. Salesforce recently announced its very own AI data scientist, called Einstein. Einstein promises to learn from your data and automatically prioritise leads based on those most likely to convert, as well as a host of other features. In a nutshell, Einstein finds actionable patterns in the data. It’s a smart move by Salesforce – it demonstrates that they’re applying advantages that can be made when you add intelligence to data.
Act-On CMO Michelle Huff recently announced that their marketing automation software with AI-powered capabilities was geared to better understand an individual. They plan to use tracking, scoring and learning to enable individuals to have a unique journey, rather than being on a generic track.
IBM Watson, which has been using its cognitive intelligence to crunch numbers since 2013, is also set to play a part in the next AI phase of marketing as it allows marketers to add a cognitive layer to their data processing. That cognitive layer, at present, can identify patterns in the data and create strategies and content much faster than humans.
The vendor promise of AI in MA is to offer a number of benefits including: more intelligent segmentation, tailored messaging, delivery of that messaging at the right time on the right channel, moving buyers more efficiently along the funnel, knowing when they are ready to buy and how much more they may be able to spend all with a minimal margin of error. For business, it means more efficient sales, reduction in marketing waste, higher customer satisfaction and more accurate revenue forecasting. But the starting point for this journey to utopia is data. I cannot stress enough how much that means to success. A MA vendor may promise a layer of AI will create more leads, but if the data sucks, then so will the leads.
MA + AI = ?
So if the exam question is MA + AI =? Then what should I write as my answer? I believe (and bear in mind I am one person sitting in my office in England thinking out aloud, so feel free to challenge me) the answer is the ability for machines to aggregate and leverage all data to identify likely buyers based on buyer signals, as well as known and compared behaviour, to deliver a personalised journey to purchase. In short, complete customer satisfaction.
The real value of the marketing technology stack is when it helps a buyer choose you over a competitor. The best way to achieve that goal is to help the buyer on their journey to purchase in every conceivable way. So if we drill deeper into my answer we see that pattern come through: the ability for machines to aggregate and leverage all data, identify likely customers based on known and compared behaviour and deliver a personalised and adaptive journey to purchase.
This is ‘all’ data, known to your datacentre and beyond. This includes buyers past behaviour, what people like the buyer have done in the past and are likely to do next, known peer purchase behaviour, buying cycle, content consumed, content ignored, social status, travel, weather, time of day, channel, financial situation and so on. That’s a lot of data. 90% of the data in the world today has been created in the last two years and we are producing 2.5 quintillion bytes of data every day. How much is a quintillion? Well, that will depend on whether you are in America (1 followed by 18 zeros) or the UK (1 followed by 30 zeros) and anyway, either one is too large to really comprehend.
With so much data being produced, and more that is sure to come, machines will soon be able to build a deep profile of a buyer. This data can be used dynamically against all kinds of real-time events to determine what message needs to be delivered next, in which format, on what channel and at what time to best help the buyer make a choice.
The AI boost given to MA will overcome the shortfalls that todays MA (automating repetitive marketing routines) fails to deliver on. AI + MA will equal a great experience that enables the buyer to achieve their goals with the minimum of effort. AI will enable your MA to make the buyer journey so awesome that they will thoroughly enjoy the process of buying from you.
Apply a little HI (human intelligence)
In the meantime, there is no harm in marketers applying a little HI to their MA customer journey. Ask yourself, is that next piece of content you are planning to serve useful for the buyer? Is it too soon or too late? Would a follow-up email benefit from having the original content they downloaded attached to it in case they didn’t get the chance to read it first time or misplaced it? Machines can do the work, but we, the people, will still need to think about what we want the machines to do. So we may as well get some practice in now because it would seem the future is approaching faster than we can imagine.
We believe in simplicity. It’s proven to be beneficial. Especially in a complex world where new ideas need to be explained. That doesn’t mean we won’t work with clients who have complex offerings. We love the challenge of unpacking layers of complexity in order to communicate your offering in a memorable way. Like all these things, it starts with a conversation. So let’s talk.