Technology is evolving and the world of online marketing is evolving along with it. Artificial intelligence and machine learning now make a wide variety of marketing tasks easy; even those which would have been impossible only a few years ago.
We’re still in the very early stages of using AI in marketing but even so, there are some impressive applications that are already available and in use today.
Here are just a few examples of how AI is currently being used in digital marketing.
Most people will already be familiar with the tailored recommendations that are offered when you log into a site like Amazon or Netflix.
These recommendation engines have become increasingly sophisticated over the years, and can be startlingly accurate, particularly for users who have had an account for several years so the service has been able to collect lots of data.
For example, Amazon has a record of:
- Every purchase you’ve ever made
- Your product browsing history
- The addresses you’ve lived and worked at
- Items you’ve wished for
- TV shows and music you’ve played
- Apps you’ve downloaded
- Product ratings you’ve made and reviews you’ve left
- Devices you’ve used to watch movies or download ebooks
- Everything you’ve asked Alexa if you have an Echo
It can use this information to deliver product recommendations based on your interests, past purchases, and what other people have purchased who also bought the same items as you.
For example, if you’ve previously bought a printer then Amazon is quite likely to recommend you print cartridges and paper. If you’re expecting a baby and you’ve ordered stretch mark cream and pre-natal vitamins, don’t be surprised if baby clothes and toys start popping up in your recommended products.
All this is powered by an AI framework called DSSTNE that has been released as open source software to improve its deep learning capabilities.
Providing discounts is a surefire way to increase sales, but some customers will buy with a smaller discount, or if there is no discount at all.
AI can be used to set the price of products dynamically depending on demand, availability, customer profiles, and other factors to maximize both sales and profits.
You can see dynamic pricing in action using the website camelcamelcamel.com, which tracks the price of Amazon products over time. Each product has a graph showing just how much the pricing fluctuates depending on season, popularity, and other factors.
If you’ve ever searched for a flight and then gone back to buy it a couple of days later only to find it’s gone up a few hundred dollars, this is also a good example of dynamic pricing at work.
Customer Service Chatbots
Facebook Messenger, Whatsapp, and other messaging apps have become a popular and convenient way for customers to contact companies, but ensuring the accounts are constantly staffed with customer service agents can be expensive.
To reduce the workload and provide a faster response to customers, some organizations are now using chatbots to deal with common customer queries and provide instant replies at any time of the day or night.
With virtual assistants like Siri, Google Assistant, Alexa, and Cortana, we’re getting more comfortable with chatbots and in some cases even preferring them to a real person.
Chatbots can be programmed to provide set replies to frequently asked questions and to direct the conversation to a human agent if the question is too complex. This means that customer service time is reduced and the workload lifted, leaving the agents free to deal with conversations that need a more personal response.
Machine generated content has been around for quite a while but the first unsophisticated attempts were pretty unreadable – they may have fooled the search engines (temporarily) but not humans.
AI for content creation has now become incredibly sophisticated to the point where Stylist magazine published three automatically generated articles created by Articoolo in its special “Robots” edition.
Automated content software is now able to generate news stories and reports in a matter of seconds that would take a human writer hours or days to create.
Even if you don’t trust machines to take over your content creation process entirely, they’re still useful for smaller tasks like generating your social media posts. The Washington Post uses in-house reporting technology called Heliograf to write basic social media posts and news stories.
Computers are also pretty good at coming up with formulaic headlines, particularly those that can be classed as “clickbait”.
Search algorithms are improving all the time in every aspect from small database product searches on ecommerce sites to search engines like Google that are used by millions of people every day.
Integrating AI into search can pick up misspellings and suggesting alternatives (“did you mean…”) and may be influenced by your past browsing or shopping behavior.
Google is becoming increasingly sophisticated at working out searcher “intent” For example if someone searches for “Apple” are they looking for information about the fruit, the technology company, or the record label?
Most search engines know if a user is on their mobile phone and searching for “coffee shops” they’re looking for a coffee shop within a few miles, rather than researching coffee shops in general.
Special results such as shopping and Google My Business results are also providing a better user experience for searchers, and voice search is becoming more commonplace as the number of AI-powered devices and assistants continues to grow.
A/B testing is the traditional approach to optimizing marketing messages and display ads, but it’s a painstaking process with an infinite number of variables to try out, and therefore takes up a lot of time and resources.
With AI algorithms you can continually and automatically optimize your ads depending on conversions and interactions.
AI ad optimization is already in use on social networks such as Instagram. Algorithms analyze the accounts that a particular user is following and will show the ads most likely to be relevant to this user. This provides a better experience to the user and a better ROI for the advertiser as fewer ads are shown to people who aren’t interested in them.