[vc_row el_id=”aiblog”][vc_column][vc_video link=”https://www.youtube.com/watch?v=sHbY5kAdGu0″][vc_column_text]
This video is about AI for Marketing. Learn about the application of AI and machine learning in marketing. For each application I am going to give you examples, inspirational stories and uses cases.
Advancements in machine learning has led to wide adoption of Artificial Intelligence. Some see this progress as a source of danger that could lead to tech overtaking humanity. Others see AI as a way to improve society, work, and life.
I believe in the latter but at the same time I know marketers, who are limited by technology, will not be able to compete, and would be superseded. This makes it extremely important for marketers to up-skill themselves. Being aware how AI can be leveraged for marketing could open new and exciting doors for your career. Top companies are already using AI to stay ahead and gain competitive advantage. Companies that will take time or cannot apply AI in their business will lag behind.
Ok now so let’s talk about the application of A.I. in marketing. I have identified 6 major areas where AI is making life easier for marketers.
1) Predictive Analytics
So the most popular and probably the most required area in which AI is being used is for Predictive Analytics.
It is used to make predictions about unknown future events. Predictive analytics encompasses a variety of statistical techniques from data mining, statistics and machine learning, that analyse current and historical facts to make predictions about future.
In business, predictive models exploit patterns found in historical and transactional data to identify risks and opportunities.
Some application of predictive modelling are:-
• Determine pricing structure
• Calculate and predict Customer lifetime value
• Predicting Churn
• Market basket analysis
• Predict advertising campaign performance
• Marketing Mix
• Lead scoring
• Product portfolio prediction
• Sales forecasting
• Predict will the user recommend your product or not
• One to one marketing
Fox Sports Australia –
In February 2019, Fox Sports Australia launched ‘Monty’ — the greatest cricketing mind ever created. Monty is a custom machine learning model, designed to spot when and how wickets would fall in live games, in real time, to boost the fan experience. It has observed every ball bowled by the Australian cricket team in the past eighteen months to work out its algorithm. Monty can spot an approaching wicket up to 5 minutes before it falls and alert viewers through push notifications to watch the live match. This lead to a double-digit increase in weekly sales growth.
Is using its loyalty card and mobile app to collect and analyse customer data including purchases, where they are made, and at what time of day. The company uses predictive analytics to process this data in order to deliver personalized marketing messages to customers including recommendations when they’re approaching their local stores, and offers aimed at increasing their average spend. A virtual barista service on the app powered by AI also allows customers to place orders directly from their phone via voice command. As well as delivering a more personalized customer experience, Starbucks uses their data from 90 million transactions every week to inform business decisions such as where to open new stores, and which products they should offer.
In general segmentation refers to a division of a whole into subsets of similar units.
The meaning of the term segmentation changes based on the context in which it is being referred.
It is common for companies to create several segments across different departments.
This is how different departments in a company can use segmentation.
Research & Development
• Research and development (R&D) may have segments to better understand consumer preferences and purchase behaviors to drive tailored product enhancements.
• R&D might also develop a product segmentation to understand product similarities and types of products that are usually purchased together.
• Finance can segment customers to aid in revenue forecasting.
• Marketing dept. can use segmentation to understand who is responding to various marketing channel campaigns to refine targeting and improve campaign response.
• Marketing has yet another use of segmentation. Three popular ways to segment your market are:- Demographic segmentation, Behavioural segmentation and Psychographic segmentation
• Human resource can create talent based workforce segmentation. Here the focus is on identifying most productive employees so they can be retained and incentivised. Another category can be of consistent and satisfactory performers. They can be improved via training. Another category can be of poor performers they are most likely to be separated.
3) Recommender Systems
A recommender systems also known as recommendation engine is a system that identifies and provides recommended content or digital items for users.
Recommendation engines are becoming an integral part of applications and websites as users love custom recommendations. Recommender systems are known to boost engagement and customer retention.
They are being utilized in a variety of areas like:-
• Playlist generators for video and music services like Netflix, YouTube and Spotify
• Product recommenders for services such as Amazon, Walmart etc
• Content recommenders for social media platforms such as Facebook and Twitter
• Google uses recommendation systems to suggest news, books, and search queries
• Recommendation engines are also being used by Job portals, food ordering site,
dating sites etc.
4) Natural Language Processing
Natural language processing (NLP) is the ability of a computer program to understand human language as it is spoken.
Natural language processing involves speech recognition, natural language understanding, and natural language generation.
You have already witnessed speech recognition if you have ever interacted with any voce assistant like Siri, Google Assistant or Amazon Echo.
Natural language understanding is being used for things like sentiment analysis, competition research, trend analysis, customer feedback analysis. Basically, you can monitor and analyse text on the web.
NLU is used by Gmail for spam filtering.
It is also being used for self-learning search which is must for major e-commerce player as it adds contextually relevant synonyms to a catalog that can result in 3x the depth of search results.
Machine translation – Google translate is yet another use case.
Conversational UI or chatbots which are being used in several industries for support and automation.
Natural language generation is being used for content marketing, like Market Muse is one such content strategy tool that is powered by NLP and AI. The software analyses articles as you write them, giving detailed directions to writers so that content is the highest quality possible. Market Muse also analyses the current events and recent stories, allowing users to instantly create content that is relevant and ranks in Google News.
Natural language generation is also being used for text message suggestions, related keywords suggestions on search engines, autocomplete, spell check and auto correct in search bars.
5) Discover Psychographic Personas To Deliver Personalisation
Psychographic segmentation divides consumers into sub-groups based on shared psychological characteristics, including subconscious or conscious beliefs, motivations, and priorities.
It tries to understand consumer decision-making process, consumer attitudes, values, personalities and lifestyles.
By understanding user psychographics you can predict user behaviour and deliver hyper-personalised experiences. Later in the course we will see how to personalise your website in real time based on your marketing personas.
Users expect brands and companies to know them – not just by their first name, but by the products they prefer to purchase, their interests, and lifestyle.
AI platforms are capable of collecting and analysing prospect and user data at oceanic scale, far beyond the capability of human beings, by drawing from internal databases, third-party sources social media and web.
IBM Watson Personality Insight service lets you discover user shopping preferences by analysing their personality through written text.
Here, (https://www.ibm.com/blogs/client-voices/ai-personalizes-japan-airlines-travel-experience/) Japan Airlines decided to use IBM personality insights to deliver personalised travel experience. They built a chatbot called Makana-chan which analysed users personality when they logged in with Facebook and Twitter account. The chatbot then assigned one of nine personality types and give advice based on their interests and preferences.
Another interesting case of Influential –
An A.I. powered influencer technology that matches brands and agencies to social media influencers. How it works is that Influential will first, analyse the brand and learns who their followers are, their interests and affinities. Then, they analyse influencers by pulling their posts and comments. They also discover insights about their followers to ensure brand safety. This allows them to answer questions about which influencers to choose, which messaging to create and which media to use.
6) Computer Vision
Computer vision is closely linked with artificial intelligence, as the computer must interpret what it sees, like humans can. Computers can be made to gain high-level understanding from digital images or videos.
Computer vision tasks include methods for acquiring, processing, analysing and understanding digital images. Aim is to extract high-dimensional data from the real world in order to produce numerical or symbolic information, e.g. in the forms of decisions.
Let’s look at its application in marketing
A) Original Image Generation With GANS
Generative adversarial networks (GANs) are an exciting recent innovation in machine learning. GANs are generative models: they create new data instances that resemble your training data. For example, GANs can create images that look like photographs of human faces, even though the faces don’t belong to any real person. These images were created by a GAN:
DataGrid – A Japanese tech company uses generative adversarial networks to create realistic images of fake fashion models. Instead of hiring new people every single time, brands can generate their own original content in a budget-friendly and efficient way.
B) Visual Social Listening
Visual listening leverages object recognition to uncover insights about a brand. Visual listening is essentially social listening, but for images.
With 3.2 billion images shared each day, visual listening is essential for brands.
In fact, 80% of the images online that includes a brand’s logo don’t mention the brand in the accompanying text. Here’s an example:
This tweet includes the Coca-Cola logo and an interesting looking hamburger, but it doesn’t reference Coca-Cola in the tweet. Without visual listening, Coke would miss this mention. It also helps them answer questions like ‘what food do our customers drink cola with’.
Brands use visual listening to uncover threats and identify opportunities
Visual listening can uncover threats like an old altered advertisement that’s potentially damaging:
Or maybe a protest around your brand
Or maybe unpleasant misuse of your logo:
It can also help you spot first copy products of your original product.
Visual listening can also uncover opportunities to improve marketing content, produce better advertisements, and even build better products.
Like analyse user generated content:
Work out the ROI on your advertisements or sponsorships:
Find out how your customers really use your product:
Discover product advocates:
See growing viral content:
Understand Product Feedback –
Image credit – Brandwatch
C) Visual Search
When it comes to online shopping, customers mostly use the search bar or a filter function to discover new products. This usually requires extensive use of tags, all of which are manually assigned to products. Since tagging depends entirely on the retailer, this can be very confusing and inconvenient for consumers, especially if they don’t know brand jargon.
Pinterest has an AI-based tool called Visual Search for visual product discovery, eliminating the need for manual tags. Instead of standard filtering systems, consumers – via mobile app or browser extension – can select any image they want, and they will be shown a whole roster of similar items. This way, consumers don’t have to know brand jargon to find what they’re looking for.
D) Optimise Conversion Rates With Images
Equipped with deep learning algorithms, Yelp can curate the most beautiful photos for any establishment to maximize their conversion rates. Instead of using the number of likes to determine the best photos, they judge photos based on the characteristics that actually matter: contrast, depth of field, and alignment to name a few.
Yelp used convolutional neural networks to design a photo scoring model. In their datasets, DSLR photos served as positive examples, while non-DSLR images were negative examples. They fed their data into the deep learning model, enabling it to recognise the qualities of good photos.
E) Visual Data For Personalisation
Using computer vision, companies can gather real-time visual data on customers to personalize experiences and inform marketing strategy. Select McDonald’s locations have implemented camera-equipped kiosks that suggest menu items based on the customer’s perceived age and gender.
In another example, some analysts theorise that the addition of a camera-equipped smart speaker to the Amazon Echo lineup could give Amazon the ability to gather customer data for more effective cross-sells. By observing what people wear and what they bring into their homes, the company can learn which products to restock or suggest for purchase.
F) Contextual In-Image Ads
When Google AdSense or Google Display Network is embedded on a site, users will see a text or image ad that’s either (a) relevant to the text on that page, or (b) based on retargeting data of that particular user.
But what about images? As it turns out, there are companies like GumGum is one of them that can display advertisements over images, by contextually identifying what is in the image and displaying relevant ads on the image itself.
For example, an image featuring playing kittens might be a good place to advertise a cat food brand.
G) Facial Recognition, Tracking User Attention and Emotions
Advancements in face analysis algorithms are now powerful enough to assess consumers’ facial expressions and measure their emotions. Disney developed an algorithm called Factorized Variational Autoencoders to determine how their audience responds to their films. Infrared cameras detect and capture people’s reactions during movie screenings. The software identifies complex facial cues and even predicts how moviegoers would feel at certain parts of the movies. This helps Disney understand what provokes certain emotions.
Want to learn more about AI For Marketing