Keyword price prediction for digital marketing campaigns profitability
Summary of the entity:
JOT Internet Media is part of Cube Ventures, one of the leading digital groups in Europe. Based in Madrid and with offices in Germany, Mexico, Brazil and Italy and more than 280 employees, the group is focused on lead generation, services monetization, digital marketing, media and investments.
Summary of the challenge:
The goal is to predict the optimized price when launching marketing campaigns, reducing the optimization time and generate campaign impact and profits from day1.
The profitability and impact of the marketing campaigns depends on how the ads are positioned in the ad platform rankings. There are two main factors: i) landing page quality and ii) maximum price per click.
Today, digital account managers spend a lot of time (up to one month) until they adjust the price of a keyword. Consequently, from one side, the keyword generates relevant amount of traffic and, from the other side, the keyword revenue is higher than the cost. Managers use to increase the CPC on a daily basis to avoid over pricing, which generates impressions with no clicks, so the campaign does no generate conversions.
By the prediction of this optimized price when launching the campaign, JOT will reduce the optimization time and generate campaign impact and profits from day 1.
There are many technical challenges to be solved to offer this prediction as a Service model, such as keyword data aggregation, definition of optimized price, data volume, data variability and so on.
In this challenge data set will be related to mobile campaigns, where there is a unique correlation between the spent budget and the incomes at keyword level.
- Keyword statistics in mobile campaigns
- Keyword list (mobile)
Expected results is a SaaS prediction service enabling:
- Reduction from 1 month to 1 week the time needed to get a campaign with positive Return of Investment
- Predict the optimized CPC in a batch mode of a list containing 500,000 keywords
- Model and Predict the CPC with a R2>0,75
- Train the model with a complete year of campaign data