Lamoda Tech Analytics Party: meeting materials

Every month we conduct hundreds of tests and experiments to understand how to make online shopping easy and inspiring for the user.

On June 27, we held an analytical party where we shared experiences, talked about the specifics of geo-tests, approaches to accelerating A/B tests, and discussed the evaluation of complex content products.

Videos and presentations of reports

Approaches to Accelerate A/B Testing for Data Products

Angelica Shakhurina, Team Lead of Product analytics at Lamoda Tech

When evaluating our Data products, such as recommendation systems, it is not enough to look at adding products to the cart. It is important for us to see the actual profit. At the same time, one of the features of the Fashion E-commerce segment is a long chain of user interaction with the product. Testing new features can take months, but there are several approaches to speeding them up. Angelica talked about them.

Slides

Geo-tests: instructions for use

Ivan Uvarov, Product analyst at Lamoda Tech

To test changes in some products, it is necessary to divide not users but entire cities into control and experimental groups. In our case, geo A/B tests were needed to analyze the effect of changing delivery conditions. How did we approach the task of designing geo-tests? Vanya talked about this in his report.

Slides

How to evaluate the effectiveness of a content product if you are afraid to conduct an A/B test?

Sima Gizzatullina, Product analyst at Lamoda Tech

Discovery is a content product that inspires users to create their own style. The effectiveness of such products is very difficult to measure. They do not directly generate profit, which raises the question: “Do they make any sense?”

In her report, Sima shared her experience of how we solved the problem of evaluating the Discovery product, what approaches we tried, and how we conducted the research.

Slides