Hey Habr! On June 30 at 18:00 Moscow time we will hold an online meetup for analysts. The speakers will talk about regional A/B tests, managing the issuance of goods in an online store, predicting profit from new features, and data science in the delivery of goods.
Under the cut, as always, are the abstracts of the reports and all the necessary links.
Reports
Regional A/B tests. Why are they needed and how are they arranged - Igor Krasovsky, Avito
What if the test group in an A/B test is not exactly known, and the user is exposed to offline, for example, in a regional TV ad? How to form a control group for a biased test group? How to measure the effect and distinguish it from a random error? I'll tell you how we answered these questions in Avito and what problems we encountered.
About the speaker: I've been with Avito for a little over 2 years, before that I worked in eCommerce and IT consulting. Now I work in the Core Analytics team, which is responsible for such areas as Data Management, Strategic Analytics, Core Analytics Platform, Key Account Analytics.
The best data products are born in the fields β Marina Kalabina, Leroy Merlin
Most of our online orders are collected from the sales floor of stores, not from warehouses. This leads to errors between what is displayed on the site and what we can actually collect.
Due to the high turnover rate of goods in stores and the complexity of stock management systems, errors occur that can be detected automatically. Based on our knowledge of systems and using social engineering, we proposed a solution that would automatically find problematic products and adjust their stock before publishing on the site.
About speaker: 9 years of experience at Leroy Merlin. First I opened stores, then I worked in them, and now I'm putting things in order in commodity stocks. Gathered a team and launched a data product in 6 weeks.
Growth model β predicting profit from features for prioritization β Pavel Mikhailov, Ostrovok.ru
- We are building a growth model β a framework based on cohorts and key metrics that models revenue in the medium term.
- We translate product and business metrics into money using a model.
- We evaluate the potential profit from features using examples.
About speaker: Head of growth at Emerging Travel Group (Ostrovok.ru) with an analytical background. I generate, develop and test growth hypotheses.
How data science helped Avito Delivery β Dima Sergeev, Avito
... Or a story about how to stop offering users to buy a Kamaz Cabin with Delivery. There are already more than 60 million goods on Avito. It is not easy for every one of them to determine whether the seller can put it in a 120x80x50 box and send it to the buyer in another city.
From time to time we make such mistakes: we offer delivery where it obviously should not be, and vice versa. I'll tell you a little about how we cope with this problem and what results we managed to achieve.
About the speaker: For the last year I have been doing analytics at Avito Delivery. Prior to that, he worked as an analyst at OZON for three years.
Ask questions in the broadcast chat - we will answer the most interesting ones on the air. After each report, it will be possible to communicate with the speaker separately.
Passwords and attendances
Broadcast starts on Tuesday 30 June at 18:00. We plan to finish by 20:40. On the broadcast, you can immediately click the "remind" button so that you do not miss anything.
If you want to receive a reminder with a link to the broadcast by e-mail, you can .
See you online!
Source: habr.com