November 30, 2023
Digital
NAF
Share:
Ekaterina Makarenkova, Getintent: How Getintent developed domestic dynamic retargeting technology
Today, more and more people buy things online, and the competition between brands is growing. According to Data Insight forecast, the e-commerce segment will grow by 30% in 2023. This means that promoting product offers and competing for market share has never been more important for e-commerce players.
After the departure of major foreign dynamic remarketing services such as Google Ads, Criteo and RTB House, the multi-billion demand was left untapped. The domestic market required worthy alternatives in the performance segment, where the main tool is Dynamic Retargeting Technology (DCO).
In Getintent (a part of the Gazprom-Media Sales House, which in turn is a part of the Gazprom-Media Holding), we clearly saw the need to replace the departed players in order to cover the remaining demand in the market. However, the creation of the DCO is a costly affair, both in terms of development and infrastructure. Therefore, it was impossible to offer a quick solution. The product had to fully address the challenges of attracting users to the website through retargeting with dynamic creatives and employ various effective ML (machine learning) models to achieve set goals. It took us a year to optimize all the features and present an efficient solution to the market.
An important prerequisite was to teach the system to display those items that are of real interest to the user. We conducted a market needs study to determine which tools would be able to meet the needs of our clients. Based on the survey results, the number of parameters analyzed in the ML model of the DCO doubled during development (up to 34). At present time, we analyze over 5 traffic parameters, 6 user device parameters, 3 geo parameters and more than 15 behavioral criteria.
Products appear in a personalized banner based on the following mechanics:
- 4 behavioral factors for a given user: views, added to cart, converted, categories viewed
- best-selling product offers
- ntermediate goals (adding a product to favorites)
- prioritization of products and product categories, stop lists
- new visual solutions for creatives
This amount of data makes it possible to predict the likelihood of a user being interested in a product on the banner.
In addition, we paid special attention to the automation of item feed registration in ADO (Advertising Data Operator) systems. This feature makes it easier for advertisers to submit reports on their own, complying with amendments to the Federal Law “On Advertising.”
A new visual editor Creative Studio allows to simplify work with dynamic creatives. It helps you to set up the appearance of creatives and the mobile Software Development Kit (SDK), which makes it possible to launch the DCO not only for websites, but also for apps.
The launch of DCO 2.0 in the Russian market enables clients with article feeds on their websites to promote their product through dynamic retargeting. However, it is important to remember the following: DCO is a long game. To understand how effective the tool is, one has to plan at least a 2-month advertising campaign and add a pixel to the site a month prior to the campaign start. In this way, ML will be able to train on your own data and show the best result in terms of conversions and ROI.
Do you want to advertise?
Send us a message right now, and our managers will get back to you right away!
What type of advertising are you after?