Talk:Issues in Interdisciplinarity 2020-21/Evidence in Amazons consumer profiling

First meeting 20/11/2020
In our first set up meeting for the Wikibooks chapter, unfortunately only three of us could attend, as one of us had to catch his flight back home. We started off by discussing the disciplines we want to play a role in our group project and continued with the planning of an interdisciplinary approach to them. Furthermore, we came up with a first type of research question, which was important to us in order to narrow down the research we have to do and get us a better idea of what the actual issue between these disciplines is. We came up with the research question (Is consumer profiling through data mining, the most accurate way to predict economical behaviours?) and chose evidence as the issue to be discussed.

Second meeting 24/11/2020
This week all of us attended the meeting, therefore, we started off with a quick recap of week 1 to explain the organization and the plan with our research question to our fourth member and made our approach overall more clear. Lastly, the whole meeting was set up with the goal to break our research question down into four areas, which we then assigned to each of us in order to have a clear and easy idea of who needs to do what type of research and in which area.

The areas were the following:

For Jade: How data mining is being done? How is the data collected, processed, stored? What types of models are being used? Of what consists the quantitative evidence?

For Shervin: Behavioural Economics (How does it work, and what actually is consumer behaviour, what do firms do with it? Still with the aspect of data mining and anthropology in mind!)

For Simon: An anthropological approach to consumer profiling (what form of evidence is being used? How accurate is this in general? (in surveys, for example, people do not always say accurate statements about themselves, so how can that be accounted for, etc.)

For Niklas: How is the data that has been mined used for consumer profiling? In what sectors does this have a lot of applications? How effective is it from a business perspective?

Third meeting 27/11/2020
Everyone present at the meeting. After having discussed the findings of everyone's research (see below for a few extracts) we decided that our research question was too broad for a 1,200-word chapter. Therefore we took the decision to focus on a particular case study: Amazon's approach to consumer profiling as it is one of the biggest companies that use data-mining. This decision followed after a long debate, where we discussed the effectiveness of data-mining, its limits and contrasted it to anthropological market research.

Extracts of the research done by all members:

Jade:

From Chapter I: Introduction to Data Mining Osmar R. Zaïane, 1999 CMPUT690 Principles of Knowledge Discovery in Databases The concern in data mining are noisy data, missing values, static data, sparse data, dynamic data, relevance, interestingness, heterogeneity, algorithm efficiency, size and complexity of data. The data we have is often vast, and noisy, meaning that it’s imprecise and the data structure is complex. The Knowledge Discovery in Databases process comprises of a few steps leading from raw data collections to some form of new knowledge. The iterative process consists of the following steps:

1. Data cleaning: It is also known as data cleansing, it is a phase in which noise data and irrelevant data are removed from the collection.

2. Data integration: In this stage, multiple data sources, often heterogeneous, may be combined in a common source.

3. Data selection: At this step, the data relevant to the analysis is decided on and retrieved from the data collection.

4. Data transformation: also known as data consolidation, it is a phase in which the selected data is transformed into forms appropriate for the mining procedure.

5. Data mining: it is the crucial step in which clever techniques are applied to extract patterns potentially useful.

6. Pattern evaluation: in this step, strictly interesting patterns representing knowledge are identified based on given measures.

7. Knowledge representation: is the final phase in which the discovered knowledge is visually represented to the user. This essential step uses visualization techniques to help users understand and interpret the data mining results.

AMAZON.com Recommendat° Item to Item Collaborative Filtering

find° a set of customers whose purchased and rated items overlap the user’s purchased and rated item → problem : people doing raking → pay for it → unique algorithm aggregates items form these similar customers and eliminates recommands remaning both algorithms used collaborative filtering and cluster models

Shervin:

→ The study of economic decision making through a psychological lense, gives firms, organizations or just us as the society the chance to predict human behaviour and to understand the daily economic actions. Richard Thaler, the father of behavioural economics, came up with the whole concept of behaviour economics at first. https://hbr.org/2017/10/the-rise-of-behavioral-economics-and-its-influence-on-organizations

→ Came up with the idea of the nudge theory (positive reinforcement and indirect suggestions to lead consumer behaviour)

-https://www.ciim.ac.cy/how-to-use-nudge-theory-for-business-success/

the way we phrase things influences our thinking and forms our consumer behaviour 2008 Richard Thaler and Cass Sunstein

Prospect Theory: Kahneman and Tversky (1979)

Consumer Behaviour related to data mining:

https://arxiv.org/pdf/1109.1202.pdf

every single step we take if it is in the supermarket or on Amazon is considered a byte of data Consumer behaviour: Consumer behavior means the study of individuals, groups or organizations about their process of selecting, securing, using and disposing the products, services, experiences or ideas to satisfy needs and the impact of these processes on the consumer and the society. Application of consumer behaviour: 1. Making of marketing strategy, 2. Public Policy, 3. Consumer Common Sense (Purchasing Patterns etc.) Endowment effect: endowing someone with a good almost instantaneously makes him/her value it more highly. This can be thought of as a combination of reference dependence and loss aversion.

Simon:

consumer studies and anthropology - why and how consumers make their purchase decisions anthropological approach = more subjective and qualitative methods Abrams 2000: some cases quantitative analysis might not help decision makers to truly understand consumers WHILE “DESCRIPTIVE ANTHROPOLOGY” (qualitative and observational) provides revealing insights Thompson and Hirshman 1995: applied classic anthropology theories to study the consumers' self-conception of body images and self care practices in the modern urban society to help the marketers understand the relationship between consumer “socialized body” and consumption behavior. Mc Farlane: observes that when consumer reaction to a new product needs to be determined, companies traditionally refer to the qualitative method

Psychographic approach: define market segment through customer lifestyle lifestyle and demographics: age, location and gender activities, interest and opinions values, attitudes and social class: buying power / purchase orientation

Consumer Typology approach: consumers motivations and mindsets loyal consumers: rare but valuable → they promote brands through word of mouth discount consumers: buy if there is an “opportunity” impulsive consumers: emotionally driven rather than logically driven need-based consumers: fulfil a need

Consumer characteristics approach: traits that influence buying decisions convenience driven: need to be fast, simple and easy connectivity driven: feel connected through buying the sale product personalization driven: prefer a fully customized experience

Niklas:

“Data for the people” (2016) Andreas Weigend: describes the flaws of the current system of data collection and privacy but also discusses their potential. Advocates for more transparency of what is being done with our data, for people to become “data-literate” and for people to gain access to their data profiles and give them the possibility to change/blur data about them. The more data a company collects about an individual, the more targeted advertising can it bring to them. For example Alphabet collects an enormous amount of personal information through Google Maps (where you shop, work, live, what places you often visit) and this information is on a database, also accessible by Google (search) that can then give you targeted advertising according to the collected information.

source: https://books.google.co.uk/books?id=pKtVDgAAQBAJ&printsec=frontcover&dq=data+for+the+people&hl=fr&sa=X&ved=2ahUKEwjT99uh16LtAhVSXMAKHeNbCtYQ6AEwAHoECAYQAg#v=onepage&q=amazon&f=false “The Internet Encyclopedia” (2004) Hossein Bidgoli: Consumer profiling based on data mining is divided into two main categories: factual and rule-based. The factual profile is the information given by a customer by actively giving the information to the company (address, name, gender, etc) while rule-based profiles are designed after tracking and then analyzing the activity of the user on the Internet.

Looking further at the rule-based profiling, there are two steps in determining such a profile: rule-discovery and rule validation. → Agrawal et al. (1998): multidimensional indexing structure and algorithm for mining profile association rules. consumer-profiling is very closely linked to recommendation systems developed by the companies. Two common techniques in developing recommendation systems are item-to-item correlation (recommending similar items) and customer-to-customer correlation (recommending products that other customers who bought the same product, bought after that).

--Poccy3 (discuss • contribs) 17:28, 27 November 2020 (UTC)

Fourth meeting 02/12/2020
In this meeting, Simon and Jade discussed the main features of Evidence of Amazon’s consumer profiling. The main purpose of this meeting was to come together with the key point to address in the Wikibook chapter. The subject seemed again very broad and it was difficult for them to find any problem with Amazon’s algorithms. They mainly talked about how dangerous it can be for society to read the same books or buy the same products. Amazon algorithms are using a ranking system to recommend product and an uniformization of ideas could be a great way to generate profit. They realized that maybe it was better to address the issue of Amazon through the lens of Power rather than Evidence.

Fifth meeting 04/12/2020
On Friday, the whole team met on Zoom during almost 3 hours. They discussed the change of issue suggested draws their attention on Amazon monopoly in e-commerce. They come together with the new following question and main points:

To what extent should politics interfere with Amazon’s power in e-commerce?

Introduction: (200 words) Shervin

Define power of firms in economics and show why Amazon is a firm with a lot of economic power. How politics can regulate the economic market, in regards to the research question.

First part: (400 words) SIMON and SHERVIN

Show how Amazon has an unfair advantage compared to their competition. Argue that this gives Amazon a quasi-monopoly over e-commerce. Explain why monopolies are bad for economics. Define Monopoly and what a monopoly means for the economy/politics.

Second part: (400 words) NIKLAS and Jade

Discuss how wealth can be redistributed and Amazon’s economic power limited through laws and regulations. Power conflict between Regulators and Amazon. Lobbyists will try to maintain Amazons power.