Ad Inventory and Audience Forecasting for Telecom

Client Company

Maxima Telecom, fast-growing telecommunication company with about 3 billion impressions per month and a monopoly Wi-Fi provider in Moscow and St. Petersburg’s subway, aero-express trains, airports and surface transport.

Challenge

70% of MaximaTelecom's income comes from advertising revenue. The company’s business model is based on mobile advertising, and the unique user data is their key advantage. For this reason, advertising campaign optimization is crucial for them. Additionally, MaximaTelecom had 25 people in their sales team, and it was difficult to synchronize their work.

Maxima Telecom was looking for an ad tech solution that would:

  1. Improve advertising campaign efficiency.

  2. Enhance the mechanism for targeting and audience segmentation.

  3. Minimize ad campaign delays and increase revenue.

  4. Synchronize work of their sales and ad operations teams.

Solution Description

Ad inventory forecasting platform is a solution that helps to predict the performance of both planned and live ad campaigns with any custom targeting with an <b>accuracy of 85-90%</b>.

  • Inventale integrated with Yandex Adfox ad server via API to get raw data on the advertising network.
  • The platform processed terabytes of data daily and used AI and Big Data techniques to analyze audience data, build behavior models, develop user segments, and apply targeting parameters, frequency capping, and seasonality.
Ad Inventory and Audience ForecastingAd Inventory and Audience Forecasting

Technology background

Inventale’s solution allowed MaximaTelecom to get a high-quality forecast for an ad-hoc targeting selection/ several targetings’ intersection.

At the part with multivariate time-series forecast Inventale used the following approaches:

  • Adaptive sampling;
  • Post-stratification;
  • Time-series forecasting;
  • Multidimensional forecast;
  • Secondary metrics models.

The second part is simulation, the success of which depends on several features:

  • Simulation model;
  • Model parameters learning;
  • Statistical quality analysis;
  • Query execution.

Results and Current Status

Inventale met all the requirements set by the client, successfully passed all the quality checks. Apart from the integration and the backend, Inventale provided MaximaTelecom with a web interface used by several client teams on a daily basis.

Inventale and MaximaTelecom worked together for 5+ years. The solution processed hundreds of forecast requests from tens of MaximaTelecom users in real time.

KEY RESULTS:
  • +25

    %

    of employees' time savings

  • 15-17

    %

    increase in revenue

  • 85-90

    %

    forecast accuracy achieved

projects.maxima-telecom.results.image-alt

Next Projects

Brand Affinity: Platform for Brand Compliance Monitoring, Analytics and Audit Automation

See Project
Brand Affinity: Platform for Brand Compliance Monitoring, Analytics and Audit Automation thumbnail
If you want to learn more about this project, leave an email, and we will send you our presentation.
Medium size orange graphic cube imageOrange graphic cube image