Back

White-Label Forecasting for Premium Sales Platform

  • The client company
    an American media corporation
  • Ad network of 48 websites and platforms
  • 15 billion
    impressions per month

Challenge

The company has developed its own Agency Trading Desk. To help partners plan ad campaigns, the client has integrated the inventory forecasting functionality in the ad campaign booking and management tool.

The system calculated the total amount of traffic, but not the amount of available inventory. And on narrow targeting, forecasts deviated from actual results by a factor of 2-3. Our team has enhanced the forecasting tool and integrated it into the client platform interface.

The Client’s Goals

  1. Develop a forecasting system that
  2. Increase customers’ loyalty
  3. Increase revenue

Solution

    We developed a forecasting tool and integrated it into the client platform via API. Advertisers can now get an accurate ad campaign delivery forecast on the selected inventory slice in several seconds.

    Processed Data

    Before:

    • the tool builds a forecast based on the ad server reports from the previous week

    After:

    the system analyses the logs of the client’s ad server, DSP, and SSP:

    • determines traffic volume trends and seasonal fluctuations
    • monitors atypical traffic spikes and smooths out their impact on the forecast
    • includes external events that
    • influence the traffic input by traffic managers into the platform in advance
    • determines the traffic characteristics of each website
    • identifies overlapping campaigns, competing for specific audiences and inventory
    • calculates the future impact of booked campaigns on the amount of available inventory
    • takes into account current changes in the network

    Supported Functionality

    Before:

    • forecast of the total volume of inventory

    After:

    • forecast of the total volume of inventory
    • forecast of available inventory
    • forecast on unique users
    • inclusion of frequency capping
    • forecast on the overlapping of any targeting parameters
    • monitoring live ad campaign delivery

    Results

    • The raw data processing time has reduced to zero. Traffickers can concentrate on other priorities

    • The system proactively alerts traffic managers to potential problems

    • STR and the revenue grew by 5%

    • Searching for problematic campaigns has become automated

    • The number of delayed ad campaigns decreased by 2.5 times