Question

!1) Describe how you would determine if a review concerning an e-commerce startup is legitimate or a fraud.

2) Discuss how you would handle the following cases: fraudulent bad review, legitimate bad review, and fraudulent good review

Answer 1

1) Description of how I would determine if a review concerning an e-commerce startup is legitimate or a fraud:
There are different factors and subjects one has to consider from every possible perspective.
One would have to do research on the concerning an e-commerce startup, its products manufactured or sold, its services, features, options, any prior reputation, check if it is genuine, its business, its existence, etc.
On the other hand, one should also check and research on the person who has submitted a review about the concerning e-commerce startup. Check if it is a real person, not a bot, software, or some kind of Artificial Intelligence (AI) Machine Learning (ML) system, check on the person's existence, role, if he is genuine, his Internet usage behaviour, his online purchase and even on the concerning e-commerce startup website, his earlier reviews about this company or any product he/she bought on this website, his review activities, surfing etc, check for any plagiarism, check if he is getting paid by anyone for writing the review or if he is doing it for free.
One has to check if he has submitted a review about the company on another person's purchase or his own purchase.
One determining this may have to use the service, product himself, or anything about the concerning e-commerce startup is selling or providing service.
One has to check for the user or the customer's use case and see if the product or service he/she purchased fits his/her use case.
The same has to be done with other customers so one can compare, determine, find differences, and distinguish between the reviews submitted by different customers.
One has to research on other users' product review about the company.
One has to check if the user who is naturally a human is biased in any way.
One should see the customer's relationship and his business with the concerning e-commerce startup.
One has to consider the ratings, feedback, comments, complaints, suggestions, advice, and the customer's purchase history.
One could use an Artificial Intelligence (AI) software to automate the determination process to see if the review is legitimate or fraud, which the software works based on algorithms, as humans suffer from “truth bias”.
Whatever the person determining the review to be legitimate or fraud should not believe it to be true just by reading someone's review unless it is proven with considerable and believable evidence.
AI is accurate (90%) at detecting legitimate or a fraud reviews accordingly.
* From the observation, fraudsters, in their reviews use more verbs, whereas legitimate reviews would contain more nouns submitted by honest people.
* Fraudsters, in their fake reviews include more first-person pronouns such as I, we, us, our, ours, my, myself, me, etc., to cover-up or over-compensate their review's illegitimacy to look real.
* Legitimate reviews would have concrete words specifically emphasizing the concerning company's products, services, features, options, specifications, and other details with facts and figures, including proofs for the same.
* Fraud or fake reviews would have vague words and descriptions and talk about their experience in general more.
* AI, when used is less expensive, consumes fewer resources, and provides speed.
* One can integrate developed features, such as for a product- the number of times it was viewed, purchased, number of people reviewed (positively, negatively, and neutrally), commented, recommend, liked or disliked it, rating, etc.
* Manually or automatically using AI, to determine and weight verified and newer customers' purchase reviews and if their reviews are helpful and true determined by other users through crowdsourcing.
One could use logo detection technology to protect the legitimate brands selling products on their official websites against fraudsters and counterfeiters.
* Fraud reviews would not normally and mostly have the reviewer’s name or in the company's customer database, as they may not have purchased any product or goods from the company.
* The fraud reviewer would have submitted the same or similar fake reviews for other businesses and companies and their products.
* Fraud reviewers mistake for the identity of the company to another brand or company for the products or services he/she bought from a different provider.
* Fraud or fake reviewers do not comment, comment less, or do not describe the product and elaborate about their experience.
* Fraud reviewers review would have false or incorrect details, facts, figures, and information.

2) Discussion about how I would handle the following cases:

* Fraudulent bad review: These reviews from mostly unsatisfied and angry customers could be false and malicious in nature.

* Legitimate bad review: This can be found out in one way, when the person determining it himself/herself uses the product or service with his experience, and with many other people's bad reviews taken together. These when they are directly and really interacted they can be legitimate based on one's experience.

* Fraudulent good review: Check if the reviewer is connected, an employee, a relative, or someone familiar to the company's key person/s. Check if he is paid for writing good reviews by the company. Check how he is related or associated to the company. Check how came across the product or service which the company offers. Check if he indeed used the product how his experience is in real.