HOW AI AND SOFTWARE DEVELOPMENT WORK ON E- COMMERCE WEBSITES: THE EXAMPLE OF INSIDER
Umut Can Gürsoy 191112102
GRADUATION PROJECT
Business Administration Department e-MBA Programme
Project Advisor: Asst. Prof Melek Erdil
Istanbul Maltepe University
Graduate School
January, 2020
HOW AI AND SOFTWARE DEVELOPMENT WORK ON E- COMMERCE WEBSITES: THE EXAMPLE OF INSIDER
Umut Can Gürsoy 191112102
GRADUATION PROJECT
Business Administration Department e-MBA Programme
Project Advisor: Asst. Prof Melek Erdil
Istanbul Maltepe University
Graduate School
January, 2020
ACKNOWLEDGEMENT
I would like to thank Asst. Prof. Melek Erdil for supporting me and helping me to complete this thesis from the very beginning. I am glad to know that the knowledge she has given me will have the most positive effect on my business and personal lives.
Lastly, I would like to express my gratitude to my family and fiance who never let me feel alone during my journey.
Umut Can Gürsoy January, 2020
ÖZ
YAPAY ZEKA VE YAZILIM GELİŞTİRMENİN E-TİCARET SİTELERİNDE ÇALIŞMA ŞEKLİ: INSIDER ÖRNEĞİ
Umut Can Gürsoy Mezuniyet Projesi İşletme Anabilim Dalı
e-MBA İngilizce Tezsiz Yüksek Lisans Programı Danışman: Dr. Öğr. Üyesi Melek Erdil
Maltepe Üniversitesi Lisansüstü Eğitim Enstitüsü, 2020
Teknolojideki gelişmelerle birlikte yazılım geliştiricilerin yapay zekanın makine öğrenimi yeteneklerini geliştirmesi kaçınılmaz hale gelmiştir. Günümüz dünyasında en önemli konulardan biri olduğu için, e-ticaret siteleri bu gelişmelerden yararlanmaya başlamış ve bu değişiklikleri ve gelişmeleri web sitelerine uygulamışlardır. E-ticaret sitelerinin kullanımı her geçen gün artarken, bireylerin online alışveriş alışkanlıkları ile iyi tasarlanmış yapay zeka yeteneklerinin birleşimi, yeni bir alışveriş deneyimine yol açmıştır. Yapay zeka, şirketlerde işgücünün yerini alsa da hem müşterilere hem de şirketlere fayda sağlamıştır. Müşteriler için alışverişi daha kişiselleştirilmiş ve daha az karmaşık hale getirmiştir, ve iş için ise pazarlama giderlerini azaltmak ve bütçeyi daha iyi yönetmek anlamına gelmektedir. Bu durumda Insider gibi yapay zeka hizmetleri sunan firmalar gün geçtikçe daha popüler hale gelmiş ve geliştirdikleri teknolojiler ile e- ticaret kavramı pazarlama, satış ve müşteri ilişkileri bağlamında farklı bir anlam kazanmıştır.
Anahtar Sözcükler: E-Ticaret; Yapay zeka; Alışveriş alışkanlıkları
ABSTRACT
HOW AI AND SOFTWARE DEVELOPMENT WORK ON E- COMMERCE WEBSITES: THE EXAMPLE OF INSIDER
Umut Can Gürsoy Graduation Project
Business Administration Departmant e-MBA (Without Thesis) Programme Project Advisor: Asst. Prof. Melek Erdil Maltepe University Graduate School, 2020
With the developments in the technology, it has become inevitable for software developers to improve machine learning capabilities of artificial intelligence. As it is one of the most important subjects in today’s world, e-commerce websites started to benefit from these improvements and they have implemented these changes and developments to their websites. While usage of e-commerce websites has been growing each day, combination of online shopping habits of individuals and well-designed artificial intelligence capabilities has caused a new ways of shopping experiences. Even though artificial intelligence has taken place of workforce at companies, it has also brought benefits to both customers and companies. For customers, it has made shopping more personalized and less complicated and for business, it meant cutting down marketing expenses and managing the budget better. In this case, companies that provide artificial intelligence services such as Insider has become more and more popular and, with the technologies that they have been improving, e-commerce concept has gained a different meaning in sense of marketing, sales and customer relations.
Keywords: E-Commerce; Artificial Intelligence; Shopping Habits
TABLE OF CONTENTS
ACKNOWLEDGE ... iv
ÖZ ... v
ABSTRACT ... vi
TABLE OF CONTENTS ... vii
LIST OF FIGURES ... ix
ABBREVIATIONS ... xi
CURRICULUM VITAE ... xii
CHAPTER 1. UNDERSTANDING E-COMMERCE WEBSITES ... 1
1.1. What Is an E-Commerce Website? ... 3
1.2. Importance of E-Commerce with Statistics ... 4
1.3. How Do They Work? ... 6
CHAPTER 2. SOFTWARE DEVELOPMENT AND AI ON E-COMMERCE ... 9
2.1. Why Is AI Important in Our Lives? ... 11
2.2. Which Sectors Will Be Affected by AI? ... 12
2.3. Changes That Ai Creates Depending on Different Sectors ... 13
2.3.1. Healthcare ... 13
2.3.2. Insurance ... 14
2.3.3. Law ... 15
2.3.4. Manufacturing ... 15
2.3.5. Transport ... 16
2.3.6. Retail ... 16
2.4. Usage of AI In E-Commerce with Examples ... 17
2.4.1. Voice Assistants ... 18
2.4.2. Searching Products with Visuals ... 19
2.4.3. Recommendation System ... 20
2.4.4. Personalized Experience ... 21
2.4.5. Pricing Optimization ... 22
2.4.6. Filtering Fake Reviews and Products ... 23
2.4.7. Chatbots ... 24
2.4.8. Robotics ... 24
CHAPTER 3. INSIDER ... 27
3.1. Channels That Insider Uses with Examples ... 29
3.1.1. Web ... 29
3.1.2. Mobile Web ... 32
3.1.3. Messaging ... 33
3.1.4. Mobile App ... 36
3.1.5. Advertising ... 37
3.2. Customer Feedbacks with 4 Examples ... 39
3.2.1. Toyota ... 39
3.2.2. New Balance ... 40
3.2.3. IKEA ... 41
3.2.4. Samsung ... 42
CONCLUSION ... 44
REFERENCES ... 47
LIST OF FIGURES
Figure 1.1 Covid-19 Crisis Accelerates Shift to Online Retail ………
2
Figure 1.2 Total number of internet sales in £m ………..………… 4 Figure 1.3 Biggest US companies by market cap ……….…
5
Figure 1.4 Number of annual active consumers across Alibaba’s online shopping properties ...
6
Figure 1.5 Homepage of Amazon’s website ………
6
Figure 2.1 The number of professional software developer ……… 9 Figure 2.2 Estimated impact of AI: Job creation and displacement ………
12
Figure 2.3 Percentage of smart speaker owners in the US who have ………
18
Figure 2.4 Alibaba’s visual search feature ……….
19
Figure 2.5 Visual search result on Alibaba’s website ………
19
Figure 2.6 Perceived benefits of using AI for marketing personalization worldwide in
2018 ………
21
Figure 2.7 Robots at Amazon warehouse ……….. 24 Figure 2.8 Picture of Amazon Scout on a street ……… 25 Figure 3.1 Example of banner management system ……….. 29 Figure 3.2 Example of social proof feature ………
30
Figure 3.3 Examples of mobile web features ……….
31
Figure 3.4 Price alert push notification example ………
33
Figure 3.5 Mailing example ………
33
Figure 3.6 WhatsApp business example ……….
34
Figure 3.7 Real-time push notification example ………
35
Figure 3.8 Lifecycle of a customer ……….
36
Figure 3.9 Discount affinity ad example ………
37
Figure 3.10 Banner of New Balance ………..
39
Figure 3.11 Smart menu of Samsung’s mobile website ……….
41
ABBREVIATIONS
AI: Artificial Intelligence
CRM: Customer Relationship Management ECG: Electrocardiogram
ERC: Economic Research Council FBA: Fulfillment by Amazon
GMP: Growth Management Platform KPI: Key Performance Indicator ML: Machine Learning
UK: United Kingdom US: United States
USA: United States of America
WHO: World Health Organization
CURRICULUM VITAE
Umut Can Gürsoy
Business Administration Department
Education
2020 Maltepe University, MBA
2017 Ege University, American Culture and Literature 2013 Kırımlı Fazilet Olcay AL
Job Experience
2020 – Present Assistant Digital Marketing Specialist at Biota Laboratories
2020 – Present English Instructor at Udemy
2020 – Present Founder of EnglishTensePractice.com 2019 – 2019 At Home Advisor at Apple
2017 – 2019 English Teacher at British Time Language Schools
Personal Information
Birthplace and Year : Istanbul, 1994
Sex : Male
Languages : English, Turkish
GSM : 90 554 802 32 04
E-Mail : umutcangursoy94@gmail.com
CHAPTER 1. UNDERSTANDING E-COMMERCE WEBSITES
With the development of the internet technology, there has been inevitable changes in the human behavior. Knowing the fact that almost everybody has access to the internet, it has changed the way that people act, react, communicate and, in general, behave. Now, people can easily send messages to each other, get news from dozens of different media sources, use streaming platforms to listen to music or watch shows and all of these actions are in instant speed, which brings easiness to daily lives of people. In addition, the change is not happening consciously among the people. It is getting its form in a natural way with the environment, so, most of the time we can deduce that people are adapting to the development of the technology and the internet.
When we talk about the changes that the internet has caused on people and their behaviors, it is inevitable to talk about the differences in our buying habits. Until the beginning of early 2000s, it could be considered as a normal act for people to buy their needs from shops, stores, groceries, etc. but after the rapid increase of computer sciences, it has become more and more natural for people to buy things from the internet. Even though it is still considered as normal to buy needs from shops or stores, it can be also considered as a traditional way to trade since e-commerce sector has grown and it has a big role in our daily lives. But why have people changed their buying habits from traditional to online?
According to Michael Solomon et al. (2013), there are several factors and benefits affecting people to buy products online.
People can shop 24 hours a day, which basically means that you do not need to wait for stores to be opened or people can shop even at night.
This situation also means that people do not need to travel to shop. It helps people to save extra time for other activities and also, it gets less tiring for people.
People have more product choices. Local businesses or markets sometimes offer limited products and it may get hard for people to find the exact products they
need. With the help of e-commerce websites, it is, most of the time, one click away from people to order the right product and choice.
Better price options help people to use their budgets more cleverly, which also creates a competition in the market.
For the marketers, it also creates an opportunity to become not only a local seller, but also to become a global seller.
Considering these advantages, it is inevitable for both the customer and the seller to choose online trade and e-commerce. When we look at the graphs, it can be seen that over the years, online retail and e-commerce sales have been increasing enormously.
According to the statistics of Statista, since the very beginning of 2000s, online marketing has been doubling its value up once in every five years. Also, if we think about certain occasions such as Covid-19, we can clearly see that it forces people to buy online and spend more time and money on e-commerce websites and, if we think about the benefits, it is reasonable for people to choose online shopping.
Figure 1.1 Covid-19 Crisis Accelerates Shift to Online Retail
Source: https://www.statista.com/chart/14011/e-commerce-share-of-total-retail-sales/ Date Received: 15th November, 2020
As it is mentioned before, there are several different factors and benefits that cause people to go to online shopping websites. However, it is not only people’s needs that is making them have an active role on online markets. During the journey on an e- commerce website, people may come across with a lot of different factors that trigger their desires and psychology, and these factors are mostly generated by software development and artificial intelligence (AI).
To have a better understanding on how e-commerce websites work and how new technologies affect the consumer when they start shopping online, we should carefully examine what an e-commerce website is and how they work. After finding out how the things work, we should focus on the AI and its importance in today’s world with some examples.
1.1. What Is an E-Commerce Website?
In Encyclopedia Britannica, etymologically, the term commerce is combination of two Latin words; cum, which means together, and merx, which means merchandise.
With this fact, in modern sense, it can be said that commerce is basically an activity of buying and selling goods among people and businesses.
In addition, if we think about the term e-commerce, we can deduce that it is a form of commerce that happens electronically. Considering the improvements in technology and the internet world, with e-commerce, the things that can be bought and sold are not only limited with tangible products. According to Satterlee, e-commerce is defined as “any transaction over the internet involving the transfer of goods, services, or information, or any intermediary function that helps enable those transactions.”
Although e-commerce activities among businesses and people can be through e- mails, messages, etc., the major factor that creates importance for e-commerce is websites. As we all know, there are thousands of e-commerce websites that help both consumer and businesses. People can find the right products as they need, and companies can expend and improve their businesses thanks to e-commerce websites’
globality. Then, what is an e-commerce website?
E-commerce websites can be thought as global supermarkets where consumers can find anything they desire and businesses can create their own audiences or expend their customer profiles. The competition among businesses and people is really high due to price differences, quality of products and the number of companies and consumers in it. Because it is global and it is always important for companies to import new ideas to their regions, new products that bring something new on the table or new designs that can change the concept of a product gain extra importance for businesses. On the other hand, when we think about the consumer side of the e-commerce websites, it is a great
field to find right products you need. Let’s say that an individual needs a niche product that is not produced or sold in their region. It becomes so much easier for the person to find the product globally and order it.
1.2. Importance of E-Commerce with Statistics
As it is mentioned before, e-commerce websites bring easiness to people’s lives.
Especially because it helps people to save extra time and to spend less energy for shopping, it becomes more and more popular each day. On most of the e-commerce websites, consumers get almost every information about the product that they need to buy. It helps to create much more educated customer profiles and as the customer gets educated and conscious, people get more and more used to it, which also makes e- commerce websites more traffic each moment. We can see how much popular e- commerce has become when we look at the data graphics on e-commerce. According to BBC, online sales has been doubling when we look at each year’s November sales.
Figure 1.2 Total number of internet sales in £m
Source: https://www.bbc.com/news/business-46260739 Date Received: 15th November, 2020
As we can see above, there is an enormous increase in online sales each year.
Because the internet can be used all around the world and because of accessibility of e- commerce websites by millions of people at the same time, it becomes inevitable for
psychical stores to lose customer and it becomes an opportunity for e-commerce companies to get chance to improve their sales.
In addition, companies such as Amazon (www.amazon.com) in the US, Alibaba (www.alibaba.com) in China and Trendyol in Turkey have been dominating e- commerce sector in both their regions and in the world. Due to the accessibility to these websites are easy and advantages such as shipment and discounts that they serve helps them to get attention and traffic from everybody. When we look at statistics of some of these websites, we can clearly see how big impact they have in the sector.
Figure 1.3 Biggest US companies by market cap
Source: https://www.bbc.com/news/business-46793466 Date Received: 15th November, 2020
When we check the graph above, we can see that an e-commerce website, Amazon, has become the biggest US company in the world leaving all other well- known technology companies behind, which shows us how trading electronically is maybe the most efficient way to take a place in commerce world. Amazon is a company that has warehouses all around the globe and especially in the USA. They have hundreds of thousands of sellers and hundreds of millions of buyers on the website.
In addition to the situation of Amazon from the USA, Alibaba which is China based company that contains thousands of retailers, suppliers and companies is also one of the biggest companies in the world with its e-commerce power. According to the Figure 1.4., it can be easily said that in 2020, Alibaba has more than 700 million active consumers on their website. The number has doubled since 2015.
Figure 1.4 Number of annual active consumers across Alibaba’s online shopping properties from 3rd quarter 2015 to 3rd quarter 2020
Source: https://www.statista.com/statistics/226927/alibaba-cumulative-active-online-buyers-taobao-tmall/ Date Received: 15th November, 2020
1.3. How Do They Work?
As it can be understood from its name, an e-commerce website is basically a website. Just like any other website, it also has a domain which means a name to be put after “www” part and it has its own unique server that manages the system. Interface of an e-commerce website is no different from any other websites in the world. Each e- commerce website has their own unique design to give the customer a better experience throughout their journey while shopping. Below you may find the homepage of Amazon’s website as an example.
Figure 1.5 Homepage of Amazon’s website
Source: https://www.amazon.com/ Date Received: 16th November, 2020
On the other hand, when we talk about the functionality of these websites, we can say that e-commerce websites are basically shops that try to convince the customer to buy goods. Just like a physical shop, e-commerce websites have their own storages but, because all the interactions are happening digitally, it can be said that they store everything digitally. In a word, they keep data of everything on their website in a database. This data can contain customer information, product photographs, description, encrypted credit card information, etc. Also, most of the time, customers have the option of being a member of that website as a user. In addition to this, sellers must create an artificial store to sell their products.
However, even though general working method of an e-commerce website is similar to the description above, there might be small changes when we talk about how sellers and e-commerce websites work together.
In all e-commerce websites, sellers have their online stores where they present their products with photos and price tags and people can buy them by adding different products to the cart and buy it, but depending on the website’s mechanism there might be several differences. On some websites, sellers also have to have their psychical stores too, which means that sellers themselves must do all the shipping processes and follow a procedure while sending the products. In this case, the e-commerce website has a middleman role that helps both the customer and the seller to come to an agreement.
In e-commerce websites such as Alibaba, the system works with the same way but, because almost every seller is a supplier or retailer, amount of the products that are bought or sold might be really high. In addition, because there are a lot of suppliers manufacturing and exporting their products, sellers may also take care of processes such as custom or export.
However, Amazon, on the other hand, uses a different method to satisfy both its customers and sellers. With the method they call Fulfillment by Amazon (FBA), Amazon takes goods from the sellers, keeps them in its own warehouses, does the packaging for the sellers, sends it to the customers, takes cares of all CRM work and takes a fee from the total price of the product that is sold. Sellers do not have an obligation to use Amazon as a warehouse but this situation shows us how e-commerce
sector has grown up to a level where not only customer satisfactory get pioneer importance, but also sellers are satisfied by the e-commerce platform itself.
As it can be seen, the way that e-commerce websites work is similar to each other and we can easily say that e-commerce websites are basically digital form of malls that have different psychical stores in it. For the customer, the only thing they need to do is to find the correct product according to their needs, enter a debit or credit card information and buy the product, and for the sellers it is a great opportunity and a platform to make their voice heard for wider audiences.
CHAPTER 2. SOFTWARE DEVELOPMENT AND AI ON E- COMMERCE
As technology and internet have grown, our understanding of computers has also changed its vision. Knowing the fact that e-commerce is basically serviced to the customers via internet, it can be easily said that software engineering has a major role in the sector. In addition to this, since software engineering is based on making computers do different types of work for certain purposes, through its development, artificial intelligence (AI) has also its own importance to run on most of e-commerce websites.
On the website IBM (n.d.) describes software development as a set of computer science activities dedicated to the process of creating, designing, deploying and supporting software. As we know, the whole computer systems and internet services are based on software, and in this case, we can say that all the activities on an e-commerce website is designed through software development and, because the sector itself is growing bigger and bigger every day, it can be said that core mechanism of e-commerce websites is also evolving. According to the data of DAXX, currently, there are more than 25 million software developers around the world and the number will be around 45 million in the next 10 years. There are more than 4 million software developers only in the USA where e-commerce websites such as Amazon and eBay not only have a great market value in the country but also, they are one of the biggest e-commerce websites in the whole world.
If we give another example for the relationship between software development and e-commerce websites, it can also be said that Europe proves us that, as the software development gains importance in a country, e-commerce sector also grows.
Figure 2.1 The number of professional software developer In Europe by country
Source: https://www.daxx.com/blog/development-trends/number-software-developers-world Date Received: 8th December, 2020
When we look at the graph above, we can deduce that countries with large amount of software developers have bigger market value in e-commerce. For example, biggest e-commerce websites in different countries such as Otto (www.otto.de) has more than 1.5 million visitors in Germany each day, Cdiscount (cdiscount.com) in France has 22 million users per month and Trendyol (www.trendyol.com) in Turkey has 25 million users and 50 million visitors per month.
In addition to software engineering, there is AI which is also very important for e-commerce websites. John McCarthy (2007) describes AI as engineering of making intelligent machines, especially intelligent computer program. In this case, we can say that AI is basically making of computer programs that can think or act by copying human behavior.
Since AI is made by software engineers and it is a part of software developing, significance of AI has also been increasing in today’s world but how does AI work?
Well, most of the time people think that AI creates values and outcomes with only analyzing one condition, or they think that the idea is a simple concept but actually, AI learns how to act or what to do by analyzing maybe millions of given inputs. It may sound like it is just how basic computer systems work but in AI, the computer tries to figure out what its next step should be with given data. To understand deeply how AI works, we can explain different components of AI with examples. According Otte (n.d.) from Innoplexus which is an Indian based company that provides services to other companies to build AI software and machines, there are several different basic components of AI and we will mention some of them below:
Machine Learning: As it can be understood by its title, machine learning (ML) is a process for AI to understand and analyze given data. With ML, AI understands and analyzes the input and gives an output without being programmed with extra effort. Actually, we, as customers, come across with machine learning’s outputs while using many different applications. For example, Google Maps understands when you park your car and automatically places an icon on the place where you park your car. Or, in streaming platforms such as Netflix, AI calculates what type of movies or TV shows you like and suggests you other shows that you may have interest.
Deep Learning: Deep learning is another application that helps AI to improve itself and come up with better results. In deep learning, the AI is given many inputs and it is expected to give one single output as the result. We can give automated cars and Siri as examples of deep learning. As we all know that Tesla Cars can work automatically, without human interaction. With AI, these cars can differentiate pedestrians, traffic lights, or certain objects that they can come across during travelling. In addition, Siri, which is a voice recognizing assistant developed by Apple, understands what people say to her and gives answers according to the questions that are asked to her.
Computer Vision: Computer vision is another application that helps AI to recognize different images, objects, etc. It basically works like deep learning but with computer vision, we can say that most of the inputs are images. The most well-known example of computer vision is face recognition system on out mobile phones. Your phone can recognize your face, let you unlock the device and it understands the differences among different faces. Another example is that in medical sector, AI can help doctors to find cancer cells or areas easily because it might know how a cancer including organ or a cancer cell looks like.
As it can be understood above, there are lots of different areas that we use AI and software development together. Therefore, while technology and computer sciences keep growing, there will be more and more different results and benefits that we may come across in the future. These applications and technologies also help a lot of companies in e-commerce sector and they drive customers to buy and consume more without realization of the customer. That’s why, it is one of main reasons why software engineering and AI is drawing attention and will get attention in the future.
2.1. Why Is AI Important in Our Lives?
With the improvement of Artificial Intelligence technology, it has been inevitable for individuals and companies to use it for better results for their future. Even though we cannot see working mechanism of a software, an application or a program, people actually come across with AI and experience the benefits of it almost every day.
The first thing that gives importance to AI is automation and its concept of thinking like a human being. With the correct software engineering and coding skills, it
can be said that AI can imitate human behaviors such as thinking, understating, analyzing, planning, etc. In addition to this, AI can analyze vast amount of data way quicker than a person. So, when we add these two facts together, it is not hard for us to imagine its capabilities in almost every business sector. Therefore, when its analyzing skills are combined with the automation, it means that it can decrease general cost for companies, create new opportunities, increase the efficiency, keep sustainability longer and ease the lives of individuals.
2.2. Which Sectors Will Be Affected by AI?
There are a lot of different purposes for which we can use AI. When we think about the business side, there are some sectors that are being affected or will be affected by it more and create issues for individuals but it is also helping businesses to decrease cost and do more efficient job. On the other hand, there are opportunities for individual due to the job creation thanks to AI. When we look at the chart made by Economic Research Council (ERC) below, we can get a clearer vision on impact of AI on certain sectors in UK by the year of 2037.
Figure 2.2 Estimated impact of AI: Job creation and displacement by industry by 2037
Source: https://ercouncil.org/2018/chart-of-the-week-week-37-2018/ Date Received: 9th December, 2020
As we know, there are more software developers in UK than other European countries. So, we can say that the data above can give us a good prediction about the future of the sectors. According to ERC, impact of AI on job creation in UK is around
%18.9 and job displacement is around %20.5. Sectors that are health, social and technical work related have opportunity because it is assumed that AI will create nearly 1 million new job opportunities. On the contrary, sectors that are mostly related to data analysis will experience a great loss. It is expected that around 700.000 people will lose their job due to AI automation. Even though AI comes with great benefit, it might also create a problem for people when we talk about business.
As it is mentioned above, health and social work industry seems that they will not be affected in a negative way as much as other sectors, contrarily, they will gain more job opportunities and grow. Although diagnosis process may be managed by the AI, it is believed that patients will demand real human touch when it comes to the hospital work. In addition, in sectors such as education, science or communication where technical needs are priorities, we can see that AI will not have negative impact.
Therefore, we can also deduce that, in sectors that requires creativity such as art and music, there will not be major chances due to AI’s lack of emotion and incapability.
Because AI is mostly based on analysis and automation, there will be loss in sectors that require repetitive, scheduled and routine duties. We can give Amazon example in this case. Although Amazon has massive warehouses all around the USA, almost half of them are operated by robots that are programmed with AI.
2.3. Changes That Ai Creates Depending on Different Sectors
Due to its infinite possibilities and effect on business side, AI can affect almost every business sectors. Aside from generalizing its benefits and negative effects on some sectors, according to MMC Venture (2017), which is a venture capital based on UK, AI directly affects lots of different sectors from insurance to transport.
2.3.1. Healthcare
As we mentioned before, AI uses its machine learning skills to develop a better analysis logic to give us better results. In this case, it uses variety of applications to helps experts such as pattern discovery or image recognition.
Deep learning technology helps AI to recognize diseases even in the early era of the disease. With the algorithm that is written in it, AI can analyze maybe thousands of visuals and data to differentiate a healthy person and an unhealthy person. For example, when a person is connected to ECG machine, with analyzing the individual’s heart rhythm, AI can give us information about possible negative situation. Or, with analyzing thousands of different visuals of skeletal structure of people’s hands from different ages, AI can deduce the average age of a person with examining and learning certain part of a body, in this case the hand.
AI also helps a lot in radiology area. Because its efficiency in collecting data and analyzing images, AI helps experts to diseases that can be seen through medical imagining. Diseases such as ALS, epilepsy; common cancer types such as breast cancer and colorectal cancer; heart diseases such as cardiomegaly can be diagnosed way easier via AI programmed machines.
In addition, AI is used for discovering drugs. Due to its side and fatal effects, it might take a lot of time for companies to invent and test new medicines for people.
Thanks to deep learning process of AI, in medical area, experts can analyze thousands different ingredients and articles related to the specific medicine to get fast and better results.
2.3.2. Insurance
In insurance sector, AI can be beneficial for several different reasons. First of all, AI can give customers basic information about the cost for their basic needs. If the customer is looking for a specific coverage, he or she can select multiple options and get better price depending on their needs. These types of actions can be done through websites of insurance companies. Also, AI can predict policy pricing depending on, for example, a driver’s age. Safer drivers can pay less and risky ones can pay more, AI may take the control to determine which is good or bad.
When we look at the insurance company’s side, there are also a lot of benefits that can help the company. For example, AI might take to control to detect fraud claims.
It is a well-known fact that insurers may try to take advantage of insurance companies to earn money and they may try to trick the companies for that. AI can trace interactions
and collect data to come up with a feedback related to the issue, which helps companies to cope with complex issues with maybe thousands of different clients.
2.3.3. Law
When it comes to law and compliance related issue, AI has to deal with a large amount of data. One of the most well-known AI usage in law is case prediction. With machine learning capability, AI can analyze previous cases and come up with possible outcomes of certain types of legal cases. With the given information such as type of the crime and former judge decisions, AI can create several different possibilities to predict the end of the case.
If we need to give another example for AI’s help in law; as we all know, constituents and rules in law can differ in each country and also, they are very tiring and time consuming to read. So, for a lawyer, it is a great service to use an AI software to do research about any cases on any written source to find what he or she is looking for.
2.3.4. Manufacturing
Industries and their manufacturing capabilities have been increasing constantly, and thanks to AI, automation in manufacturing process has been easier. It helps companies to plan, analyze, and predict possible problem with less effort.
For example, AI can predict the maintenance time for machines in work place with high accuracy, which makes companies save a lot of time and prevent possible downtimes. Analyzing dozens of different machines in a workplace can give AI a perception to predict the maintenance need. Also, via censors that can be placed on equipments, the chances of downtime can be decreased.
If we give a more specific example for manufacturing, we can also say robotics has a major role in manufacturing products. Fully automated system with robotics can help companies to reduce human mistake and increase efficiency in a workplace. Even in agriculture, it can be seen that picking up fruits and vegetable, and separating bad one from the good one is way easier and faster thanks to AI’s image recognition skills.
2.3.5. Transport
In today’s world, it is a well-known fact that companies can produce cars which are programmed with AI. With AI’s deep and machine learning capabilities, cars are able to track roads, differ pedestrians and car, keep safe distance from other vehicles and carry the passenger from point A to B.
It is inevitable that we will see this private experience in transportation sector too. With the help of AI, it can be easier for companies to arrange their transportation needs with autonomous vehicles. Different types of vehicles such as trucks and busses can be used while delivering goods or carrying employees from various locations to work. Making schedules and predicting delivery time may be easier for the company to manage.
Companies such as Google, Uber and Tesla have been working on improving autonomous vehicle concept and it is believed that they will be one of the key aspects of the future for both individuals and companies.
2.3.6. Retail
When it comes to retail, AI takes a major role to manage how things work.
Although we are going to mention in detail later in this chapter, there are several different concepts that AI mainly focuses on.
As it was mentioned before, retail and e-commerce has been increasing for years and it is inevitable for companies to try to find an alternative method to manage such data for better customer and employee experience. Most of retail companies use AI applications such as price optimization, customer segmentation and content personalization for the customers. With price optimization, AI can trace the price differences of the same type of product around the whole internet and set an optimal price for the certain product, which helps the company or virtual store to keep the competition with other sellers. In customer segmentation, AI tries to find more and more information and data about the customer using his or her different accounts on different platforms. With this way, AI understands what are the interests of the customer, creates a profile of the customer and gives personalized experience to the customer, which leads us to content personalization part. With content personalization,
after AI collects available data about the customer, it tries to give the customer a unique experience. For example, when an individual creates an account on Netflix, he/she select multiple shows that they like and AI suggests others shows with the same content.
Although this seems like a basic computer algorithm, the AI of Netflix can even create same kind of thumbnail with the same kind of graphic design automatically by analyzing customer’s interests.
2.4. Usage of AI In E-Commerce with Examples
AI has a great impact on almost every sector in the world and e-commerce is not an exception for it. So far, it is known that every step you take leaves a mark behind for AI and AI can trace your movements to get better results for its purpose. Therefore, people come across with AI each time they visit a website, and e-commerce websites are one of the places where people experience its positive features and benefits.
AI not only helps to customers to find the right product with the easiest way, but also it provides several different features to the business owners and marketing departments to make their processes effective.
There are tens of different AI applications to support services and products that e-commerce websites serve to their customers. Services and products section of e- commerce websites mostly focus on the customer since they are the things that the website is trying to sell to the customer. Hence, it can be deduced that AI tries to give a better user experience to the customer to increase the sales and find the right product for the right customer. As we know, popular e-commerce websites get millions of views each day, and because it would take large amount of time and requires a lot of workforce for companies to deal with millions of people at the time, e-commerce companies use AI to create a better experience for the customers.
On the business side, companies need to focus on several different factors such as customer relationship, supply chain, warehouse management and marketing.
Considering that each of these department requires specific care and large amount of analysis with workforce, AI helps companies to figure out and organize all the processes with different applications such as robotics and customer targeting.
Handling the performance of the website is another subject that should be taken care of carefully since the websites are visited by millions of visitors each day and require special treatment for each issue that they come across such as review supervision and marketplace moderation.
2.4.1. Voice Assistants
Voice assistants on our mobile phones or separate voice assistant units such as Alexa or Google Assistant have been in our live since the beginnings of 2010s. They help people to organize their daily activities such as setting alarm, arranging reminders, set calendars, etc. Because of the complexity of daily routines and tasks, people might not have enough time to even take their phones into their hands, or sit at a computer and spend time for shopping. That is where voice assistant integrated e-commerce websites take a step forwards because this application helps a lot for customers to buy or check a product easily.
Figure 2.3 Percentage of smart speaker owners in the US who have used the device to do the following
Source: https://www.statista.com/chart/9579/smart-speaker-use-cases/ Date Received: 10th December, 2020
In the graph above, it can be seen that %30 of owners of smart speakers such as Alexa purchased products via the voice assistant applications on e-commerce websites.
In addition, respectable amount of people used voice assistants for streaming services and %13 of people ordered food and services, which gives us a brief example how voice assistants are important and takes place in e-commerce and AI world.
2.4.2. Searching Products with Visuals
Previously, we mentioned that, with the deep learning process of the it, AI can differentiate visuals from others. For example, if we show thousands of cat photos to the AI and let it analyze the cats; and after a while show to the AI a photo of a cat and ask whether it is a cat or not, the AI can recognize what a cat looks like and give us a positive feedback.
If the customers do not know the name of the product they are looking for, or if, for example, the customer is looking for a product in foreign counties with different languages, this application gives a great advantage to the e-commerce website to keep the customer on the website and turn the situation into profit.
Figure 2.4 Alibaba’s visual search feature
Source: https://www.alibaba.com/ Date Received: 10th December, 2020
For example; Alibaba, which is one of the biggest e-commerce websites not only in Asia but also in the world, has this feature and makes it easier for users to find the right product that they are looking for to be manufactured.
Figure 2.5 Visual search result on Alibaba’s website
Source: https://www.alibaba.com/ Date Received 10th December, 2020
Let’s say that a customer has found a product on Amazon US and wants to sell the same product with different brand name in different regions. In our example, we are looking for a thermos with a screen showing the temperature of the liquid in it. As it can be seen above, only thing the customer needs to do is to download the image of the product that he/she found in Amazon and upload it on Alibaba. In the second picture above, we can see that Alibaba gives the customer positive results and shows different manufacturers.
2.4.3. Recommendation System
Recommendation system that most e-commerce websites provide is a great tool for both the customers and the company. The main purpose of the system is that websites try to sell more product according to the customer’s interests. As the customer spends time on the website, AI collects more and more data about their search history and interests. For example, if the customer has bought a decoration for his/her home, the customer will realize that when they visit the website next time, they will see related products or equipments to the previous purchase. Let’s say that the customer bought a vase for the living room, they might see pictures for wall decoration as recommendations later.
Netflix, which is the biggest movie/tv shows streaming platform in the world, is one of the greatest examples on how AI can trace and trick the customer to make movies be watched. UX Planet, which is an online magazine that focuses on user interface and experience processes of applications, states that Netflix uses an algorithm that not only focuses on the type of movies or shows that the customer has watched, but
also it can trick the customer with visuals to get attention. According to Blattmann (2018) , who is the writer of the article, Netflix uses such an algorithm that if the member of Netflix has watched movies featuring Uma Thurman before, when the system recommends the movie Pulp Fiction, the member will see a thumbnail of Pulp Fiction mostly focusing on Uma Thurman in the photo. On the other hand, let’s say that the member has watched movies featuring John Travolta before, when Netflix recommends Pulp Fiction, it will probably show a thumbnail with Travolta in it.
2.4.4. Personalized Experience
Personalization of content on e-commerce websites is well known experience for every e-commerce user and it can be said that this concept works in collaboration with the concept of recommendation system sometimes because after the AI collects data from the user, it can recommend new opportunities based on the choices of the individual.
One of the most common features of this concept is seasonal content that is presented based on your location. If the user is located in an area which is experiencing cold months of the year, the e-commerce website possibly suggests winter related products such as coats, boots, anti-freezer, etc. With such information, the concept of personalization can send e-mails individually to every customer to draw their attention on certain products.
If we give another example for personalized experience; almost every e- commerce website designs its homepage based on the products that the customer has taken a look before. This is also one of the examples that personalization and recommendation systems work together. With the data that the customer has provided before, the website designs the homepage with products that could draw the attention of
the customer. Plus, it might show previous products’ pages that the customer has checked and spend time before to convince the customer to buy the specific product.
Figure 2.6 Perceived benefits of using AI for marketing personalization worldwide in 2018
Source: https://www.forbes.com/sites/louiscolumbus/2019/07/07/10-charts-that-will-change-your-perspective-of-ai- in-marketing/?sh=6b9ad0862d03 Date Received: 10th December, 2020
According to the graphic chart that Forbes presented in 2019, we can say that marketing personalization on e-commerce websites for the purpose of targeting potential buyers and turn viewers into profit work perfectly.
First of all, improving customer experience is leading factor of why companies try to focus on personalized marketing tactics. When the contents are personalized, as we mentioned before, users can find the right product more easily and buy it thanks to the way that the product is recommend to them. With the correct methods, it becomes inevitable for companies to increase their performance metrics, analytics, KPIs, productivity, etc. That is why, only %1 of companies only states that they see no benefit using AI in personalization.
2.4.5. Pricing Optimization
The AI of an e-commerce website can decide what is the best price for certain products or for certain companies with using its machine learning application. On some e-commerce websites, owner of the virtual store can set a percentage for increasing or decreasing the price for the AI and the AI can set a value for the specific price according to the competition among the other same or similar products in the market.
Let’s say that “company A” sells a product for 20$ and tells the AI that it can increase and decrease the price %10 depending on the competition with the other products. If
“company B” puts a discount on its similar product in the same category, let’s say they made the price 19$, the AI can make the company A’s price 18,90$ to compete with the other companies. Of course, this option is, most of the time, depends on the company owner’s choice.
Other great example for pricing optimization is Udemy (www.udemy.com), which is a website that sells educational materials and courses to its members. When
most people visit Udemy, they will realize that almost every time, there is a discount on the website in general. The website does this on purpose so that they can create a rush moment for the customer and persuade them to buy a course. In addition, if there is no discount generally on the website, after users spend some time on specific course page, they will realize that Udemy recommends special discount on the specific course, which also draws the customer’s attention and calls the customer to action for buying the specific course.
2.4.6. Filtering Fake Reviews and Products
When the website gets millions of views each day and creates maybe millions of sales in a year, it is certain that there will be fake reviews on products to create traffic and there will be possibility for counterfeit products. In this case, the AI of the website could support the website with its filtering methods. When there are a lot of fake reviews on certain products, AI can detect the difference between human-made comments and machine-made comments. Also, it can detect the difference when there is massive number of comments written suddenly.
One of the great examples of this filtering is system is, again, Udemy. When an instructor creates a new course and wants his/her course to appear on the first page of the search engine, they may sometimes give the course to people they know for free and make the comment and review the course. In this case, the AI of Udemy can understand if the reviews and comments are fake or not by looking at the names/surnames, the method that they attended to the course or how fast the reviewed and left a comment on the course.
In addition, it is inevitable for websites such as Amazon to come across with fake products and services. When a seller tries to list a product, Amazon requires a real invoice from a real retailer, reseller or manufacturer. It is a well-known fact that the whole system is controlled by an algorithm and depending on the given documentation, the AI decides if the product is reliable or not.
Other services such as Entrupy provides a service to differentiate an original product from a fake one. They especially work with luxury brands and according to
their website, they use AI and machine learning algorithms to understand the difference.
Brands send microscopic images of leathers, logos, materials, etc. to the company and the AI works out the rest of the process with %99.1 success rate.
2.4.7. Chatbots
When it comes to customer service, it is one of the most frustrating things for companies to keep quick and creative communication with the customers. Because there are millions of people doing shopping online every day, and because there is inadequacy to respond too many demands from customers, chatbots that are programmed and utilized with AI can make a difference for companies to cope with such an issue.
Chatbots can be a great alternative for quick solutions for companies. When customers want quick responses to basic questions and concerns, a chatbot can respond quickly with very well design conversation styles. Even though they are very effective to save time, there might be need for help in complicated situations, yet it is still one of the best solutions for customer problems.
We can give chatbot of WHO as one of the popular chatbots nowadays. Since Covid-19 is a worldwide pandemic and it is a concern of billions of people, WHO created a chatbot that people can communicate with a messenger application called WhatsApp. The system works as follows; individuals text “hi” to specific number that is provided by WHO via WhatsApp and a chatbot replies with a menu. The menu includes several frequently asked questions and gives number to each one of them. As the individual text the number of the question, the chatbot answers the questions, and gives fast and detailed answers to each of the questions.
2.4.8. Robotics
Using robot that are controlled by AI seems as they are only required in manufacturing items but in e-commerce world, it also gets more and more popular. As the leader of the sector, according to Jason Del Rey (2019) from Vox, Amazon has more than 200.000 robots working in their warehouses.
Figure 2.7 Robots at Amazon warehouse
Source: https://www.aboutamazon.com/news/innovation-at-amazon/bots-by-the-numbers-facts-and-figures-about- robotics-at-amazon Date Received: 11h December, 2020
These robots mostly help to package goods and deliver them to correct place in the warehouses. Because they do not require any interaction from outside, they do their tasks without any human help and speed up the process for delivering products to the customers. Amazon, with its Prime service, can deliver product really fast to the customers, sometimes on the same day; and these robots can be seen as the main reason why the system works perfectly.
In addition, Amazon is trying to improve their robotics capabilities even further.
They have got robots that deliver products directly to the customers’ home or given address. These robots called Scouts. Even though they are not as popular as the ones in warehouses, they are used in some regions, especially in Southern regions of the USA.
Figure 2.8 Picture of Amazon Scout on a street
Source: https://www.aboutamazon.com/news/transportation/whats-next-for-amazon-scout Date Received: 11th December, 2020
As we have seen in the previous chapters of the thesis, e-commerce and technology has been improving simultaneously and keep changing as the time goes by.
Though it seems like e-commerce only includes improvements on website, other services to support the capabilities of the e-commerce website might increase the effectiveness as well.
CHAPTER 3. INSIDER
As we have mentioned before, while technology and the internet keep growing and improving, it is inevitable for tech companies to develop and use better opportunities for their own benefits. Considering that automation has its own importance in the world, it also has massive importance in today’s internet world with the shape of artificial intelligence. When people use the internet in their daily lives, they come across with AI functions almost every day. For example, the biggest websites such as Instagram, Facebook, Twitter, etc. use AI to run their websites and applications, and it can be said that most of the processes that are being done are run by AI itself.
Even though some companies develop their own AI systems, there are also companies that give this feature as a service such as Insider. Insider is an Istanbul based company that was founded by six co-founders to create better marketing experience for companies. With the platform that they call Growth Management Platform (GMP), they provide a system that draws the attention of potentials customer and turns them into profit for the company that is associated the with Insider.
With channels such as web, messaging, mobile app, mobile web and advertising, Insider directs its AI to find the best possible solution to get to the customer and make the purchase. Even though sectors that they are working in can be different from each other, because the main purpose is to call the customer to action, we can say that their methods can be applied to every company concept that is related with the internet, including e-commerce.
If we give an example for their usage of channels; traditionally, companies, when they have a discount on their websites or stores, send a pre-prepared e-mail to their customers to announce that they are having a discount on their platform. What Insider does is that their AI system creates a personalized experience for the customer and sends a notification via their browser to show that the product that they were previously interested in has discount on it. In this case, we can say that the AI keeps tracing the steps of the potential buyer and notify him/her when there is discount. The system not only gives a personal experience to the customer, but also helps the company to make profit by targeting specific products for specific customers.
Apart from the channels that they use for drawing attention of the customer, they use four-step solution plan to have companies make profit; Acquisition, Activation, Revenue and Retention.
In acquisition phase, what basically Insider aims is that they want to focus on the right target group without wasting any extra cost. As Insider states on their website, many companies take a poorly targeted shotgun approach when it comes to acquisition of customer, so, their understanding of acquisition is exact opposite. Insider narrows customer groups down to the best possibility of revenue.
After the acquisition phase, there is activation. With the activation phase, Insider tries to evoke the customer to take an action on the website or buy a product that he/she is interested in. This means that Insider drives the customer to engage and increase the conversion rate on the website or the app.
After calling the customer to action, there is revenue phase. As it can be understood from its name, in the revenue phase, they make specifically targeted customer to buy the specific product and in return, get the revenue.
And lastly, after creating satisfied customer, Insider focuses on royalty and a good relationship with the customer. They call this phase as retention. In retention phase, even though the customer has already made the purchase, the company starts to have a good relationship with the customer and tries to make profit from him/her in next periods.
Although all these processes might seem traditional, what Insider uses is AI.
During the time period that the customer is experiencing on the website, AI has a major role. From the beginning of acquisition phase and the end of retention, AI automatically takes action with inspecting the customer’s previous behaviors and predicted decisions.
To get the maximum benefit, Insider uses several different communication methods and channels, which people come across most of the time when they use the internet even for basic internet actions.
3.1. Channels That Insider Uses with Examples
As it was mentioned before, there are several different channels that Insider uses with its AI: Web, Mobile Web, Messaging, Mobile App and Advertising. With all these channels, Insider takes action with different methods to draw the attention of a customer.
3.1.1. Web
With its AI, Insider presents several different options to use for their customer.
We all know that AI collects as much data as possible about the user on the web and tries to come up with a plan for given purpose. In this case, for e-commerce websites, there are multiple alternatives that e-commerce websites can use via Insider.
First of all, notification centers are now one of the most popular features that most websites and e-commerce websites use. With a notification center, an e-commerce website can announce their promotions, discounts, coupon codes, etc. and all of these can be seen on the customer’s screen easily since most of the time notification centers are located on the top of every page of a website.
What is important about Insider’s notification center system is that it track how customers react to certain notifications such as how many clicks they have got or what the conversion rate is. As these numbers are followed and calculated automatically by the AI, the person or the company who is using Insider’s GMP can easily shape their promotional options and can get the best result from it. In addition, with the help of AI, customers can get personalized discounts and offers via the notification center.
We mentioned personalized banners before and it is one of the key features that Insider offers. With the personalized banner management system, they provide the best experience for the GMP users so that they can make profit by engaging the customers.
As we all know, banners are one of the most important graphical aspects of a website since it is an image file with big resolution and size and it is one of the first things that people come across when they visit a website. Let’s say that a customer searched a wallet when they first used the e-commerce website. When they visit the website for the second time, since now the AI knows what they have searched before, the banner will show a picture of a wallet model that can draw the customer’s attention. If the customer
has a habit of buying discounted products, the banner will show a wallet with a discount. If not, it will show type of a wallet that customer search for before.
In the picture below we can see an example from the website of Insider. In the picture, one of the customers probably bought discounted products before. So, in the first yellow graphic, the customer sees a personalized banner that focuses on the discount. But, in the other two graphics, we have a customer that is into sport shoes. In this case, the AI shows the pictures of two different sport shoes twice so that the customer can take action and add one of the products to the cart.
Figure 3.1 Example of banner management system
Source: https://useinsider.com/web/ Date Received: 5th January, 2021
In addition to banner management system, there is also a layout optimization system that Insider offers to websites. Most of the time, e-commerce websites or other websites have standard website layout that is seen the same to every customer even if they are new visitor or previous users. With the AI system of Insider, new visitors see the standard layout for the first time when they visit the website, but in the later visits, they might see that menu and its categories has changed due to their search history on the website. For example, if there is a category called “pants” in the menu of the e-
commerce website, it might change to “watches” after the customer searches for a walled with the search bar. Just like the banner system, personal layout system also helps customers to have a personalized experience during their journey on the website and help companies to make profit.
One of the other features that Insider offers is social proof. With the social proof feature, the customer sees how many people is looking at the certain product. This situation leads the customer to give quicker decisions about buying the product since it creates an urgency. It also shows the customer that the product they are looking for is popular and available at that moment. Creating that rush feeling for the customer can make them buy the specific product as soon as possible.
Figure 3.2 Example of social proof feature
Source: https://useinsider.com/web/ Date Received: 5th January, 2021
Lastly on the web side, if the customer has looked at several different products and the AI has learned what the customer is interested in, the AI will offer the similar product when the customer is looking at a product so that they can compare the one on the screen with the others. Let’s say that the customer has looked at five different jeans before and is looking at another one at that moment. With the help of Insider’s AI and
its data collection capability, it might show similar products to compare them. This situation again can convince the customer to make the purchase more quickly by showing the comparison and giving the best option.
3.1.2. Mobile Web
It is a well-known fact that in today’s world, we are all living with a small computer in our pockets, which means that all websites including e-commerce websites must optimize their system to mobile versions. That’s why companies like Insider pays attention to mobile side, too. As some of the features are the same with the web features, we will focus on different features on the mobile.
First of all, what Insider tries is to give better experience to the customer while helping the company to make profit. That’s why they create some sort of a game concept to keep the customer on the mobile website and make it fun for them. For example, features such as Scratch to Win Coupon, Wheel of Fortune or progress bars evoke the customer to buy products while having fun. With “Win Coupon” feature, the customer comes across with a scratch to win coupon and as he/she scratches it, they will get discounts or specific offers. In this case, the customer feels special and the situation triggers them to buy products.
The same thing happens with the “Wheel of Fortune.” The customer spins a wheel, gets a specific offer and feels special. As the customer starts to add some products to the cart, they might see a progress bar with a title indicating how much money they should spend more to make the delivery free. All these things might seem a little fun game to the customer and can create a special connection with the company
and the customer.
