I attended the SAS Analytics Experience which was this year held in Rome. The conference is very hands-on and you pick up a lot of useful tips and tricks, new analytical methods and how they can solve issues we all face in marketing analytics.
There were a lot of interesting keynotes at the seminar. Three things came up as major themes which everyone was talking about: IoT, artificial intelligence, and open source.
IoT – The Internet of Things
The Internet of Things has for several years been a big topic in almost every industry. So far, it has mostly been associated with operational benefits such as logistics and production maintenance. However, the big break in marketing is coming. IoT can provide real-time, contextual data from many touchpoints giving us as marketers invaluable insights into customer journeys and at which stage specific customers are at any given point in time.
Artificial intelligence is another big topic these past few years and in marketing, AI has enormous potential in marketing and we have only seen the beginning of this. In the seminars, the topic was discussed in relation to Twitter bots, image recognition, speech recognition, product recommender systems and the analytical techniques used to create them.
Most of the analytical models behind all of this are based on some sort of deep learning algorithm (which is really just another name for neural nets). The big promise of AI, or deep learning, is that it could replace many of the handcrafted features that typically go into models built by analysts with algorithms that can handle semi-supervised or unsupervised feature learning. What this really means is that once the model framework is set up, it can continue to learn and improve as long as it is being fed with new data to train on.
Google released a Pictionary-style game called “Quick, Draw!” just a couple of weeks ago using a neural net to identify what you are drawing. Give it a try here.
There is clearly a lot of investment going in to developing artificial intelligence from pretty much all major players on the market in one way or another. We have seen both Salesforce (Einstein) and Adobe (Sensei) implementing AI components into their marketing cloud and Google, Facebook and Amazon are investing huge amounts to develop their AI offerings.
Applications in marketing are huge and with many companies giving away the work they have done (one example is Amazon that has released its framework for product recommendations), it is really just up to marketers to figure out how to use it for their own benefit.
Interestingly, SAS raised open source as one of the major themes of the conference. For SAS, this represents a big shift being a proprietary software vendor. SAS now integrates with Python, Java, R, Hadoop, and there is even a kernel developed so that you can run SAS code from Jupyter. It is now possible to develop models in R and deploy them within the SAS framework. It will be quite interesting to see how this pans out going forward.
Finally, I was surprised how notable the change in the audience was compared to two years ago when I last attended the SAS seminar. Two years ago the seminar was attended by pretty much analysts only. This year there was a wide range of professionals, ranging from CEOs to CMOs to management consultants. It just goes to show how analytics and the use of data is not just a competitive edge for a few large companies, but a necessity for businesses to survive.