This article was previously published on LinkedIn.
From personalisation to individualisation
When I started at O&M in 2004, I was working for an IT client. The split between offline advertising and online activities was 80% / 20%. By 2007, it was the opposite.
With the rise of online, we started the long race of personalisation; digital marketing helped us to adapt the way we communicate on brands. We could segment better and send more specific communication for a lower budget!
The ability to precisely measure the efficiency of digital marketing allowed us to start A/B testing and answer the crucial question: ‘Does the customer prefer a green call-to-action button or a red one?’. What looked like details proved to be key success factors. But that was not enough, we needed to talk directly to the customer.
We then witnessed the start of matching target users databases with websites or newsletters, the famous ‘Dear Mrs X’. What looked good at the beginning quickly became unproductive. People knew the trick.
Let’s move forward in time and get to remarketing. This technique combines the information gathered on what the user has seen on several websites to deliver an incremental and optimised message. Results were positive, but we were still missing something. We focused on the user/customer/visitor without taking into consideration the world they live in. We needed context, and with it a massive volume of data.
That’s when artificial intelligence started its journey in marketing.
From individualisation to context
Two years ago, we started IBaround, a small start-up dedicated to contextual marketing. It began with a French retailer who wanted to slow down the sales drop of a very seasonal product. We worked on splitting the season change into measurable streams; temperature, rain, light etc. Crossing this data with sales results, we realised that the strongest sales influencer was hygrometry. A key info to optimise advertising spends.
However we needed more than just past context integration. In a World of real time optimisation, we had to have live feeds. We natural next step was awareness.
From context to awareness
Two years ago, when I was visiting the mobile site of a well-known retailer in Sweden, I saw that the top page product was a swim suit. That was normal, since we were in the middle of July. Well, the problem was that when it comes to the summer, the weather in Sweden is anything but ‘normal’. That specific day, it was 12°C and raining. I would have preferred to see either a more suitable product, or a message saying something like ‘We know it rains, but don’t worry, tomorrow the sun is back!’
Now platforms like Google Awareness API are doing the job pretty well. They capture contextual info, transmit them to the site which will then adapt accordingly.
But again, that is not quite enough as the approach is still static, we need incremental actions. We need intelligence.
From awareness to intelligence
One of the most visible examples of AI within digital are chatbots. They are interfaces aimed at having a discussion with a human, with the ability to learn from and during the interaction. If you think about it, this has existed for quite a while but in a more manual way; when you phone any administration and you navigate through a menu by pressing keys.
Just like a young mind, a chatbot still needs to be supervised by its ‘parent’. You might have read about Microsoft’s experiment last year about a chatbot which went rogue becoming racist and insulting people.
So until we reach this intelligence maturity, what’s the immediate next step? More controlled bots, for sure, but the answer might also be to let the user do the job in coming up with optimised marketing and new products. In other way, tap into the most perfect computer there still is, the human brain. Asking the user to come up with new ideas, that is the grail of the any brand. The future client now designs (and sometime even finance) the products they will buy. The H&M and Google Awareness API project is a good example. The API monitors the mobile user’s daily activities, lifestyle etc. After a week, the context signals are passed through an algorithm which creates a digitally tailored dress design for the user to purchase.
So what could be next?
During the 19th century, the invention of the sewing machine triggered riots, with people in fear of losing their jobs. The same thing happened with every major revolution, from the factory automatisation to the uberisation of services. But each time an industry was threatened by innovation, Schumpeter’s creative destruction opened opportunities and innovations.
Artificial intelligence is a signal for a revolution in advertising agencies, it is also a warning for a much needed full mindset change in e-commerce, or even state administrations management. The user finally becomes fully in the centre, engaging in a one-to-one relationship with the brand, impersonated by the AI.
A more democratic economy?