Artificial intelligence in digital marketing

Artificial intelligence in digital marketing

Currently, the processes associated with artificial intelligence (AI) have a vital impact on marketing and internet advertising as these facilitate the study of the market in brands. This has allowed to make delivery of advertising information appropriate to the characteristics and interests of users, a process also called behavior targeting. AI then serves to study, define, and segment users to create speeches and strategies that respond to the demands and attributes of their audience. To this end, AI also includes different forms of tracking and collection of information such as tracking cookies or data capture behind different free platforms that feed its databases. Thus, depending on the quantity and quality of the information, and the objectives and times of the companies, there is a certain type of learning for the optimal AI machine to develop the task. However, the generation of "solutions" and results it is a repetitive and incessant process as companies always seek to anticipate reality.

Data collection

Data collection is the set of processes by which it is possible to obtain the information to create or enrich a database; this can occur in a similar or virtual way. In the non-virtual world, people share their information when they fill out an application, register to vote, register a product for warranty, purchase a driver's license, or participate in a raffled. Sensitive data such as transactional data can also be known when people use their credit card or pay an invoice with a check. Virtual information, on the other hand, is mainly favored by the internet since in it cookies record every click, conscious searches in the browser and the interaction of people on social networks are recorded, the cell phones record their owners if they say "hello siri" or "ok google", there are cameras on the streets that keep records in databases etc. And, in the future, when the internet of things to everyday life is developed and integrated more, the amount of information will be much more detailed.

The following is the list of data typologies that an organization can collect:

Demographics: Name, Gender, Age, Race, Address, Phone, Fingerprint, Heart Rate, Weight, Device, Government ID.
History: Education, Career, Criminal Background, Press Exposure, Publications, Awards, Association Memberships, Credit Score, Legal Affairs, Divorce, Travel, Loans.
Preferences: Settings, Promotion of Ideas, Political Party, Social Groups, Social Likes, Entertainment, Hobbies, News sources, Browser History, Brand Affinity.
Possessions: Income, Home, Automobiles, Devices, Clothing, Jewelry, Investments, Subscriptions, Collections, Social Relations.
Activities: Keystrokes, Gestures, Eye Tracking, Part of the day, Location, IP address, Social posts, Departures to eat, Watch TV, Heart rate over time.
Personality: Religion, Values, Donations, Political Party, Skepticism/Altruism, Introverted/Extrovert, Generous/Greedy, Adaptable/Inflexible, Aggressive/Passive, Opinion, Mood.

Data hygiene

To feed AI with data it becomes necessary to know how it was collected, cleaned, sampled, added, segmented and what transformation is required before combining it with other data streams.1 This process is paramount to ensure that the outcome of the analysis serves the desired goal and can influence the outside world: for example find the perfect owner to induce a group of people towards shopping. For this reason a data expert is recommended to decide which bits (information) should be included and which are rectified.

See also Data Mining

Data features

For a marketing expert it is necessary to know the typology of the data with which we work in artificial intelligence. At this point there are two key concepts: cardinality and dimensionality. The first refers to the uniqueness of the items in a database column. For example, an e-mail has high cardinality because it should be unique while living in "Paris" has low cardinality because more than one shares that characteristic. As far as dimensionality is concerned, it is recognized as the number of attributes obtained about an individual; when you have information from more than one individual, a database is generated where each attribute becomes a dimension. AI is key in the treatment of multidimensional database since through its artificial neural network it is possible to find connections and patterns with statistical basis. This multidimensional data can be mapped and studied using support vector machines that use algorithms to predict the category of a new data.

See also Entity-Relationship Model

Types of learning in Artificial Intelligence

In general there are 3 types or levels of learning in Artificial Intelligence: supervised, unsupervised and by reinforcement; depending on the specific need of the company, some may be applied.

Supervised learning

Main article Supervised learning

It's about teaching the machine certain rules (training data) to create a profile and recognize the results that those entries meet. For example, if you are taught to identify cats through a group of images, in the future you should be able to identify them on their own. Or, if a brand already has its best user type defined, those features can be used to locate them all.

Unsupervised learning

Main article Unsupervised learning

The machine makes associations and draws new conclusions from the information it already contains: If it already identifies cats then it can study its context and recognize that these are found in chairs and sofas as a trend. Or, for example, you may find that the person who searched for the camera "Sony DSC W830 20.1 Megapixels digital camera" after having searched for "digital camera", "digital camera reviews" and "wi-fi cameras" is 50% more likely to buy than the one that only sought "digital camera," "digital camera reviews" and "digital cameras for sale."

Associations

By rules of association machines can infer for example if a person is prone to buy something with the logic of "those who bought this also bought that". Data analysis by association has two key concepts and are: support and trust (confidence). The first would refer to the number of times an item has appeared in the shopping bag and the second relates the number of times two items have been purchased together. For example, if a person bought toothpaste 400 times and floss 300, and 300 times they bought the products together it means that the trust is 3/4 or 75%, but the association between the two is 100%.

Anomalies

Contrary to the patterns are the anomalies, which should be paid special attention because these are unexpected changes that must be explained to take action. For example, a fraud can be detected if a purchase is made at a location that does not match the person's actual location. However, there may also be beneficial anomalies to make marketing decisions such as being a trend on Twitter and being able to take advantage of fame to induce shopping.

Learning by reinforcement

The machine, based on its own learning process, generates outputs and conclusions that it tests to learn and improve. Reinforcement learning differs from supervised because in supervised man must indicate when the machine is wrong while "by reinforcement" the machine creates its own mental model of the world in which, for example, it decides which poster is most impactful for a certain group of people.

How unsupervised learning works

This learning system is composed of neural networks that function like a brain where each neuron transmits information to others to generate a result. Each artificial neuron has its limits because at the individual level it has certain inputs and outputs, however, if there is a situation with high support and high confidence (confidence) it sends the message to others since given the situation makes them prone to spread the signal. Considering the one-way situation the tickets would be the factors that may influence whether or not it occurs: weather, effort, cost. These inputs are not binary but operate on a grayscale (since sentiment about the weather or effort does not have an answer to or b; for that reason the outputs are a percentage (for example, 65% chance of going to the cinema).

This number of layers and factors create different decision layers that become deep learning or deep learning. This learning combines many layers of information (e.g. level of education, the likelihood of buying pasta and dental floss etc.) to enrich each neural unit and thus provide new outputs and conclusions. In this way, it is no longer the human who establishes these relationships, but from the data the machine creates a learning process. The strengthening and robustness of this learning becomes what was previously called "reinforcement learning".

Browser marketing

Google

Google through its "Google display network" composed of different websites (also called publishers) supports its service/program "Google ads". In it Google receives ads from advertisers to then select the websites (publishers) associated with the ad depending on criteria such as the relevance of the content, the offer price and the revenue it would get. Thus, in Google's targeted advertising model, publishers are used to track users while browsing the Internet (via the DoubleClick cookie whose domain belongs to google) and at the same time to profile users when they visit their pages. For example, if a user who frequently visits a football website will be tainted in the category of "sport" and in the "football" subcategory. This crosses the demographic information that Google owns (such as age, gender, location) creates a user profile that will be used to show you advertising (the behavior targeting method). According to research, 88% of the tags/categories with which an individual is profiled (such as "football") receive targeted ads that are directly associated with the keywords that define them. These tags/categories that define individuals are updated in a 1- and 2-minute range and can be seen on the Google Ads preferences page.

Social media marketing

Facebook

Facebook, like other platforms such as Amazon, use the 1-1 marketing method where it is used: the history of pages visited, information collected by data brokerage firms (such as Experian, Acxiom and Epsilon digital data profiling) and user data and interactions on the platform to profile and advertise them according to their interests.4 Specifically Facebook had an evolution in its targeted ad technology when on May 6, 2015 it partnered with IBM in order to give its users a more personalized and relevant experience. In practice, Facebook's personalized ads have made the platform an indispensable tool for advertisers (92% of marketing companies use it) as it is less expensive than other media and also has wide reach (at least 1.39 billion active users per month).

Instagram

Concerns about the use of AI in Marketing

Business models such as Google's where personal information acquires monetary value raises major concerns about users' privacy. Sensitive categories such as sexual orientation, health, religion and political ideology are being used to display targeted advertising even though in many places it is forbidden to use that information. According to research between 10% and 40% of ads shown to people profiled with these sensitive conditions, they correspond to ads that appealed to those characteristics.