Alexei ChernobrovovConsultant on Analytics and Data Monetization

Видео

Magic conversion of ML into money

How to measure the necessity of Machine Learning in business and to know the result?

We will try to analyze ML- and business-metrics, will speak about pluses and minuses of using ML in business and about possible alternatives.

How to choose an attribution model taking in concern the tasks of business

Attribution is an explanation of the traffic source that helps to make conversion (targeted interventions and sales).

- What models of attribution can be and what model is better?

- Why do they work and how to make administrative decisions?

We will discuss that questions and demonstrate different examples.

How to build a Customer Journey Map

Today we would discuss Customer Journey Map and also:

  • for what reason we use CJM;
  • when do we need it and when do not;
  • how to build a CJM;
  • how to improve business using that technic.

Proxi-metrics and estimating of advertising campaign

What instruments should we use when there is a lack of data and we want to estimate the advertising campaign.

We will discuss for what reason we use proxy-metrics and what benefits we can get from using them.

Discount personalization for the online-store

How to predict the probability of buying according to discount and how to give a right discount in order to earn money.

Website personalization for users beside on Customer Journey Map

That lecture will tell you about the customer journey map (CJM) and what is it.

The main approaches to the automated allocation of patterns of user behavior using modern methods of machine learning, as well as methods of building personalization based on CJM are considered.

Cases: app for the Bank, website of the well-known cartrader and the large online-store.

Marketing management based on user segmentation by CJM.

The presentation about marketing management was presented at the conference Locomotiv based on an example of Pult.ru. There were discussed different segments of users and models of consumer behaviour. And also strategies of working with customers.

How to set the effective prices on products using machine

A huge review on different ways of pricing.

How do the prices depend in online stores and what should you do when the demand on products is inelastic?

How can pricing help in increasing earnings and margin?

A base way of building a system of pricing would be discussed and a successful case of the store from TOP-100 runet'а would be showed.

Automatization of ML problem solving on the example of HackerEarth ML

The report reviews the way of automatization ML problem solving using the example of the HackerEarth Machine Learning Challenge 1.

What would you know:

  • The review of the existing implementation, its pros and cons.
  • What scalability architecture should you choose
  • What datas and results of working have to be stored
  • Difficulties during the process of implementation
  • How the system has showed itself while solving a contest

The analysis of prediction in granting the loan (Tinkoff Data Science Challenge)

What would you know:

  • How to approach the solution of the problem analytically without delving into the training of models
  • What model is usually used when granting a loan
  • What methods of processing noisy categorial signs could be used
  • Where could you get the additional information
  • What additional signs could help in the task of credit scoring.

Contacts