Antonyan Artem

Analysis

Our team made a stakeholder map to identify those groups of people with whom we were to interview to research user experience and conduct high-quality, in-depth interviews. Our stakeholder map covered essentially any ordinary user, whether it is an artist or a manager. Then, we invited two respondents for an in-depth interview, during which we tried to identify the main problems of Avito and further compile a Customer Journey Map.

During the interviews, we learned about the following possible situations our respondents experienced:

• The seller asks to send an advance payment for the postage, and then disappears;
• The buyer asks for the bank card security code from the trustworthy seller;
• Rob during transferring goods;

Thus, this stage results were the identification of the main difficulties and the preparation of the Customer Journey Map.

DESIGN

The target user group has already been defined at this point. Based on the last stage results, we were able to identify some difficulties that our users faced. Further, by conducting a business audit and applying the CANVAS method to combine the user needs and business objectives that were set before us during the brief, we were able to identify three problems that we were going to solve, and which, in our opinion, were the most important at the moment:

• Security of trade transactions;
• Trust of users to each other;
• Fraudsters;

This allowed us to create an offer that solves the problems of the target audience. Thus, having a Brief available, one of the moments of which was Avito’s expectation of using machine learning and the problems we identified, we created features such as, for example, that could, in our opinion, make the user’s life easier and solve the company’s problems:

• Implementation of the onboarding system
• Implementation of a chatbot
• Use of Machine Learning (what we teach the car, what indicators we look at the compliance with the price of the goods; early actions associated with this phone and mail; the environment related to the user with this number and mail; communication with profiles in social networks; GPS tracking; support calls related to this user and/or similar user, etc.)
• Tracking numbers by their geolocation
• Using Token password

At a meeting with Avito representatives, we voiced our ideas and presented our vision for MVP. The management approved some; some were removed as they did not correspond to the business.

For example, Avito did not want to be responsible for conducting money transactions and delivering goods.”

The product we created had to use Big Data. The solution should be simple, with a minimum of clicks. Using the company’s budget to apply Token password was also not included in Avito’s plans.

Based on the corrections made by Avito during the meeting, we removed the Use of the Token password, the ability to conduct monetary transactions. We also prioritized Machine learning and continued to think about how to solve our problems using neural networks.

We have narrowed our offer down to a mobile-only solution.

Next, our team created several stories and compiled a Product Backlog. We have chosen one, in our opinion, to study the story. Then we made the Story Mapping of the following story: as a buyer, I want to be aware of the seller’s reliability so as not to lose money and extracted MVP from it:

PROTOTYPING

At the moment of prototyping, based on the MVP, each team member created his own paper prototype in a limited time. Having three paper prototypes from each team member, we did BRAINSTORMING, corrected our version of the prototype, held a presentation, and corrected our solution again.

The main difficulty of this work consisted in working with data visualization, namely, in what way we will show the AVITO user that it is not worth contacting this or that buyer or seller, while not changing the interface of the already existing Avito Marketplace and not creating a large number of buttons and clicks.

Our team’s visual solution was the use of a chatbot, which, in case of risk, intervened in the correspondence. In our application, the bot learned to decide to show warnings or blocking messages in the event of a threat.
The bot could also recognize addresses, names of locations and give recommendations in case of suspicions.

We also made a sound solution that signalled danger in the event of a call from a potential fraudster. Thus the machine identified the danger through the application’s dialer.

During the conversation, the machine analyzed the characteristic voice features to compile a database, which is necessary for recognizing fraudsters if they change their phone, catching keywords and phrases in the conversation.
Thus, in the event of danger, the machine put the conversation on hold and issued a threat message.

After confirming the paper version of the version, we proceeded to create a digital layout. The layout was done in Sketch.

Also, our team faced the need for an animation to demonstrate how our solution works. We used Principle.

PRESENTATION

We presented our application in Readymag. The main difficulty with this presentation was that it was not a selling landing page for a specific application, but a technology presentation. Here, the use of animated interfaces to demonstrate the capabilities of the technologies that we applied in our solution helped us a lot.

TASK

As part of the training assignment, Avito shared with us the problems of their trading platform. The main difficulties of the platform were:

• Safety;

• The trust of users in each other;

• Fraudsters;

• Search and selection of goods.

It was also suggested not to dwell on this and, if another problem is identified, work on it.

PRESENTATION

PRESS HERE