Last Updated: November 16, 2020

Most of the advertising you see is completely irrelevant. Watch an hour of television and you are bombarded with ads for prescription drugs you don’t need, cars you can’t afford, food you don’t like to eat and politicians you would never vote for.

This illustrates one of the oldest challenges for marketers: most marketing messages fall on deaf ears. Because there has never been a good way to connect specific marketing messages with specific customers, many people tune out most advertising efforts. Traditionally, advertisers have to cast a wide net and hope that they can reach at least some of their intended audience. This leads to a lot of wasted time, money, and energy.

All of that changed with the rise of the Internet and e-commerce (See also Internet Marketing). Online environments allow retailers to create commercial spaces that are specifically tailored to individual shoppers. Now customers can seek out only the things they need without wandering through massive stores filled with irrelevant products. This allows marketers to create custom advertising messages that speak directly to their intended audience. Additionally, online activities can be tracked to get a better idea of a customer’s preferences, creating hard data that marketers can use to craft more compelling advertising messages.

The Internet has ushered in a new era of marketing in which advertisers can market to individuals rather than demographics(See also Personalized Marketing). This leads to increased sales, improved customer satisfaction, and more efficient marketing strategies. The guess work and intuition which have classically been used to create ad campaigns has lessened over time. Marketers can now rely on empirical evidence and customizable environments to create a truly personal shopping experience for the customer.

What is Computational Marketing?

Computational marketing is an emerging field that uses the power of computing to assist with marketing strategies. Computation is a way to give form and function to blocks of information too large for any person to analyze. Online activities generate huge amounts of information about users. Everything from your Facebook page to your search history creates a record of who you are and what you like. This data is high sought by advertisers who want to connect consumers with the products they are most interested in buying (See also Analytical Marketing). In order to do this, they must create technologies that can process and respond to all this information.

The main components of computational marketing involve algorithms, mathematical formulas, computer programs and other analytic tools to refine the shopping experience for customers. The mission is not to create advertising, but to find better ways to connect customers with advertising. The greatest advantage of computational marketing is that it automates many of the traditional duties of marketing. Instead of an actual marketer choosing what ad will appear in front of a consumer, a computer makes this choice based on the data it has collected. This leads to greater efficiency for marketers and higher profits for retailers. (See also Geomarketing)

You have probably been involved with computational marketing and not even known it. When you do a Google search for fishing, ads for fishing poles show up at the top of the page. After buying a book off of Amazon.com it recommends other books that you might like. Netflix displays movies to you based on the movies you have watched already. All of these marketing messages are based on your past behavior online. This behavior has been tracked and analyzed to provide consumers with a more personalized experience the next time. Computational marketing gives advertisers an unprecedented way to connect their customers with the products that are most relevant to them.

The Science of Computational Marketing

Computational Marketing is a highly technical field that draws on several disciplines within computer science, and one large area of economic theory.

• Information Retrieval – The Internet contains an unimaginable amount of information. Accessing the specific information you want can often be a challenge. Information retrieval tries to create stronger, faster and more efficient connections between users and the information they want. In a commercial sense, the easier a customer can find what they want, the more likely they are to buy it.
• Machine Learning – This is a branch of artificial intelligence that tries to create ways for machines to recognize and react to complex patterns. It is, quite simply, teaching machines to learn things from their environment. In the context of computational advertising, machine learning is used to pick up on patterns of customer behavior and offer relevant online experiences. An example is BestBuy.com, which learns from your past purchases to get a better sense of what products you will want to buy in the future.
• Optimization – Not all websites work the way they were intended for all users. Optimization is a process for maximizing the accessibility of a website. This process takes many forms. Mobile optimization involves tailoring your website to be accessible on smart phones. Search engine optimization enhances your website’s ranking in search results. The goal of optimization is for your website to be easy to use and easy to find for the largest number of people.
• Microeconomics – Microeconomics is the study of the economy on a small scale. It is concerned with the way individuals make choices about what to buy. Computational advertising tries to connect personalized marketing messages with individual shoppers. Using the theories of microeconomics, computational advertising can find better ways to engage with consumers on a direct level.

Who employs Computational Marketing?

Large retailers and popular online portals are companies that are most interested in computational marketing. Retailers want to use this strategy to increase sales and customer satisfaction. Companies like Amazon.com and Walmart.com use computational advertising to match their products with their customer’s needs. If the customer’s shopping experience is tailored specifically to their consumption habits, they are likely to buy more, and buy more often. Computational advertising allows retailers to recommend products based on past purchases, offer personalized deals, and tailor marketing messages to the customer’s demographic. (See also Closed-Loop Marketing)

Online portals such as Facebook and Pandora use computational advertising to create value for their brands. Pandora, for example, creates radio stations based on a user’s music preference. The more that person uses the website, the better the program understands their preferences and the more relevant the song selections become. If a user enjoys what they are hearing, they are likely to return to the website over and over. The value of this traffic can be leveraged to sell advertising space and to create other commercial products like smart phone applications that support the brand.

Smaller websites and businesses don’t have the technical infrastructure to harvest and analyze customer data, but that doesn’t mean they cannot benefit from computational advertising. Businesses can exploit the algorithms that guide major search engines to get higher rankings for their websites. If a website is easier to find, it will be visited more often. Rather than designing their own algorithms, small businesses can take advantage of the ones already in place.

Consumers with a strong presence online have the most to benefit from computational marketing. The more data that has been collected about them, the more relevant their online experience will be. Users with a smaller online footprint will have to sort through unwanted products, irrelevant ads and misguided recommendations.

Google vs. Bing

The success of an online ad is measure by its click through rate (CTR). The CTR compares the number of times an ad is clicked on with the number of times it is displayed. The chart below is based on data collected in a 2011 study that compared the CTR of Google with Bing. Using the same keywords, the study measures the percentage of times that the first entry in a search page was clicked on, and then the second and so forth. The chart illustrates that Google has a much higher click through rate than Bing, suggesting that they are returning more relevant results to their users. For now, Google is winning the battle of computational marketing.

Computational Marketing Moving Forward

Computational marketing has only been possible for a little over 15 years and has only recently been seriously researched. There is enormous potential for the field to grow and expand. Here are some of the main goals of computational marketing researchers moving forward:

• Better Ad Relevance – The ads that appear in search results are based on the search terms that were entered. Creating the strongest link possible between the terms and the ads is the best way to get users to click on those ads.
• Social Media and Behavioral Marketing – Researchers want to use the biographical content logged in social media sites, user profiles, and search histories to offer personalized ads to customers. Maximizing the potential of all this intimate information is one of the largest areas for growth.
• Reducing Costs – An ad is only worthwhile if it generates more revenue than it costs. Researchers want to use computational techniques to lower advertising costs and increase returns from each ad.
• Automated Marketing Models – This involves using intelligent models to predict the outcomes of marketing decisions. These models help forecast the behavior of real consumers in actual markets.
• Harvesting User Information – Advertisers want to get the most complete picture they can of their consumers at the moment of purchase. Finding new sources of detailed information is the best way for advertisers to understand their customer’s needs.
• Multi-Platform Analysis – The Internet isn’t the only advertising sphere that matters. Marketers want to integrate computational analysis into their print, TV, and radio ad campaigns.

How is a Computational Marketing Plan Developed and Deployed?

Computational marketing is less of a self-contained strategy and more of a tool that is used to compliment traditional online marketing plans. Marketers will begin by analyzing any and all information that has been collected about customers to find out who they are and what they want. Technicians will mine databases and other information sources to provide a portrait of the consumer to the creative team (See also Technical Marketing). Using this data, marketers can draft copy, establish design schemes, and tailor price points to their target audience.

Once an ad campaign has been deployed, computational marketing strategies will be used to evaluate and refine that campaign. The first step will be to identify all of the areas where customers interact with the products being sold. This includes everything from a company’s official website, to affiliate retailers, to search engines where advertising appears. Once the landscape has been defined, marketing engineers can find the best ways to optimize the technology they have to work with.

It will then be necessary to define clear benchmarks for success or failure. If the numbers are determined to be too low, it is probably because customers are not finding what they are looking for. Computational marketing strategies are used to tweak the process to better serve the needs of customers. Marketers can redefine the algorithms they use to try and return more relevant search results. Customer data can be reanalyzed to get a better sense of purchasing trends and demographics. Products can be tagged with descriptive information that makes them easier to find during searches. None of these strategies involve changing the aesthetic or core message of the ad campaign. They only refine the choice of message that is delivered to customers.

Careers in Computational Marketing

Computational Advertising Researcher

Data Analyst

SEO Expert

How can a marketing school help you succeed?

Computational marketing is a growing field, but most of the jobs opening up require training in computer science and engineering rather than marketing. Many marketing departments are filled with creative people but desperate for employees with technical skill. They do not need copywriters; they need scientists and engineers who can respond to the technical challenges of online advertising. Earning a computer science degree focused on computational marketing is the best way to find work in the advertising industry.

Creating links between products and customers is not a purely technical process. The traditional strategies of marketing also apply. It will be important to understand consumer psychology and how that translates to online environments. Even the most carefully designed algorithm cannot accommodate all of a customer’s wants and needs. Training in marketing will help aspiring computational marketers create strategies that use customer data to create more appealing shopping experiences.

Skills and Attributes of a Computational Marketer

First and foremost, a computational marketer has to have incredible technical skill. This is a discipline that draws on many fields within the sciences and applies them to the vast and dynamic space of the Internet. It will not be enough for a computational marketer to have expertise in one area. They must be able to draw on research and theory produced throughout the field of computation.

The ability to think abstractly is another crucial skill. Computational marketing tries to make sense of vast and often unrelated sources of information. Identifying the most important sources and extracting meaning from them requires marketers to think outside of the box. It is not enough to fall back on established marketing strategies and received wisdom. Computational marketers have to be able to take empirical data and transform it into compelling advertising.

Finally, successful computational marketers must have a wealth of personal experience using the Internet. All the technical training in the world cannot teach you how the Internet works in practice. Having experience with Facebook, Youtube, and a variety of search engines is the best way to understand how people use the Internet in the real world. Any computational marketing strategy has to be based on the behavior of true life users.