Back

Yield-Optimisation of Waterfalling Increases Online Advertising Revenue by 15% in One Month

  • The client company
    a large Russian media platform
  • 48 million
    monthly users
  • 1 billion
    ad impressions per month

Challenge

The platform would sell ad impressions through a daisy chain. Twenty external monetisers, along with guaranteed campaigns, were set up for a waterfall auction, one for the entire ad network. The company wanted to ensure its sales model’s best performance and chose Inventale Yield Management to optimise the waterfall.

The Client’s Goals

  1. Automate sales data analysis
  2. Set an optimal waterfall model
  3. Increase the revenue and eCPM
  4. Reduce the number of lost impressions, increase fill-rate

Solution

We analysed the customer’s sales model and found out that the waterfall order conflicted with the demand to particular inventory segments. Despite the billion impressions sold per month, the revenue was falling short from what had been anticipated: the inventory was sold cheaper than it could have been because it was sold to the wrong advertisers.

To fix this situation, we built a model that performed

  • Inventory and demand forecasting
  • eCPM, fill-rate, the share of unfilled impressions, guaranteed campaigns and external monetisers yield analysis
  • Floor price optimisation for dynamic monetisation management
  • Inventory segmentation and waterfall configuration for each of the segments. The system placed advertisers in optimal order, taking into account their bids and fill-rate on the given segment
  • Sales forecasting for the configured daisy chains
  • Self-learning based on the latest ad server logs and reports
  • Weekly inventory re-segmentation and the waterfall adjustment

Results

In the first month after applying the Inventale Platform’s recommendations

  • The revenue has grown by 15%

  • eCPM has increased by 10%

  • STR has grown by 5%

  • Individual partners and segments' fill-rate has increased by 8-10 times