In Search, YAN and Campaign Master we promoted hot queries:
We have collected a common semantic core.
It was divided into segments by type, form of training targeted industry database and programming language. A segment with related areas and brand requests was separately identified.
Examples of queries we used: "programming schools", "programming courses", "programming for schoolchildren", "teaching programming for children at home", "Python for children", "Scratch for children", "mental arithmetic", "robotics".
We also launched auto-targeting, look-a-like and advertising campaigns customized according to the interests of visitors to competitors' websites. The best results were achieved by auto-targeting with fine-tuning of campaigns combined with other types of targeting.
Analytics
The customer already had a configured analytics system. We only slightly improved it. We added the following to the UTM tags:
1. Fixing placement for campaigns in YAN to track the quality of incoming traffic and remove spam sites . It is important to understand that fraud in YAN will always be there, but experienced marketers know how to reduce its amount.
In this project, we cut fraudulent traffic at the campaign level:
recorded placement and monitored traffic quality;
turned off impressions on “.com” sites that were taking up traffic;
we ran repeated tests until we received traffic from Zen, Pikabu, VKontakte, Avito and other high-quality sites;
analyzed data on high-quality sites and turned off those that were less profitable.
