Tom Blah
09-Aug-2017 10:26:57

The future of AI in marketing and beyond

It’s not here to replace marketers, it’s here to help them

 

 

 

main-qimg-a97528af669a30935df02c4a2c72098d-c

 

Today, consumers can already expect, and do find relatively tailored ads served to them. For example, each person’s Facebook feed is extremely personalised, most likely based on personal search histories or the pages they follow. Sometimes, this kind of targeting misses the mark because of its simplistic approach: you searched for a widget six months ago, which translates into an understanding that you are interested in that widget, and now you cannot stop seeing ads for it. The problem is that this approach is not effective, because you already bought that widget when you first searched for it six months ago.

AI can take marketing to a deeper, less intuitive, yet more effective level. Marketers are increasingly eager to use data to make sophisticated predictions based on previous behaviour, which is not necessarily related to a product or brand. This kind of analysis goes beyond the Facebook widget example. Marketers want to learn about what someone does and what kind of person they are so that they can gain an empathetic understanding of them, and better influence behaviours.

Here are the key trends, shifts, and areas where we feel AI will have the biggest impact in the marketing industry:

Audience segmentation

Gone are the days of traditional demographics being an effective way to segment. Because of the prominence of social media platforms, people, mainly millennials and the younger generation, are watching less and less TV. They are getting most of their news, entertainment, and consuming most content on social media platforms, which has led to a fractured media landscape, and more need than ever for marketers to think about tribes. Enter AI: AI can help marketers take a wide group of people and find the tribes that actually exist within that wide group. Taking it a step further, AI can segment these audiences around their actual content interactions online. For example, take a large asset management company trying to target to a demographic of elderly, wealthy people looking to invest large sums of money. Using AI to discover this demographic’s content interactions, you can discover that people who fall into this demographic also enjoy cycling or luxury car brands. The asset manager can now create content around those things the audience is actually interested in, instead of assuming they are all similar just because of their age or income.

Audience insights

The next step in enhancing marketing through AI, is richer audience insights. Bigger data sets, and more sophisticated algorithms allow for more unintuitive insights about target audiences to be uncovered. For example, if you uncovered a strong correlation that showed people who enjoy cycling and interior design also frequently watch historical documentaries. Instead of broadcast marketing around cycling gear, this understanding may help direct marketers at cycling brands to sponsor historical documentaries. AI can also help determine personality types and dominant personality traits through natural language processing (NLP). This works by analysing all the text written online by users and creates a robust profile based on these data points. Lastly, deep neural nets and computer vision can help analyse images to determine what types of images are being well engaged with - e.g. from beach/holiday photos, to pictures of awesome city architecture: perfect for giving marketers the creative sparks or mood boards for their target audiences.

Bots

Advancements in artificial intelligence, coupled with the proliferation of messaging apps, are fueling the development of chatbots — software that uses messaging as the interface through which to carry out any number of tasks, from scheduling a meeting, to reporting weather, to helping users buy a pair of shoes. Although bots are seen as an ongoing, growing trend and potentially the future for customer service, we don’t really believe that chatbots are advanced enough to completely take over consultative customer support - yet. Currently, though, they are useful for simple demands and helping reduce certain companies operating costs and customer support costs.

 

Amazon-Echo-vs.-Google-Home-Make-The-Best-Choice

 

Alexa and home assistants

The trend of reducing transactional friction has always been front of mind for retail companies, but for consumers it is increasing now more than ever. People are more concerned with convenience and lack of hassle, and they have begun to put their trust more and more into technology. For example, in the past, people would walk to the store, buy what they need, walk home - it would take about twenty, or so, minutes. The next step was ordering it online, which required making an account and entering your payment details and then waiting up to 3-5 days for a delivery. Now, you can just speak to your home assistant and ask them to order you whatever you desire, with same day delivery.. Brands are lining up to try and forge partnerships with these voice assistants, create dash buttons, and just generally take advantage of this reduced transactional friction.

Democratised use of AI and ML algorithms

In the future, the marketplace for AI and ML will change. Instead of the smaller brands having to hire expensive in-house data science teams, of which they have no experience doing or managing, tech companies will just sell off-the-shelf AI and ML algorithms direct to the brand. Therefore, the brand doesn’t necessarily need to know how this technology works and can just use the already established algorithm to help understand how to better market and create content for their target audiences.

Decline of survey & focus group data

Surveying and using focus groups to gather and analyse data is extremely outdated. AI provides better ways to analyse big data. Firstly, there is much more data available than ever before. Secondly, surveying and using focus groups is not all that conclusive or meaningful, and often leads to skewed results, as most marketers know. In the future we will see a large decline in these methods as the majority of people will switch to using AI to sort through big data and provide us with more conclusive results and better, richer insights.

Next post