Skip to content
You are here:

How does Algorithmic Optimization work?

Algorithmic Optimization

At the heart of AdLibertas is our algorithmically optimized waterfalls. Much like the bulk of modern stock-trading – we’ve found algorithmic management is much more effective than manual optimization. Through programmatic integrations we can optimize your inventory in a more granular way, bringing forward more competition, pushing your prices higher and increasing the yield of your inventory.

How does algorithmic optimization help?

In short optimization is a perfect problem for computer’s assistance. It combines many layers data (data consolidation), applied against many of decision-points (algorithms) to come up discrete decisions with a single goal (increase earnings). While it’s too in-depth to review all the algorithms we employ here, we’ve included some specific rule-sets in our cascading algorithms in way of example:

  • Automatic line item creation & updates: having programmatic abilities to create and manage line items allow us the luxury of not being weighted down by a very complex optimization configuration. Complexity– managed properly– can greatly increase performance.
  • Discrepancy Monitoring: Ad serving can generate serving discrepancies that can cause problems. Monitoring this over time is important to ensure you’re not losing impressions (& therefore revenue).
  • Graduated allocation ramping: Moving line item-order is risky. Just because a network performed highly, doesn’t mean they deserve all traffic. We gradually ramp in and out line items to reduce the risk of incorrect “point-in-time” metrics.
  • Fill rate efficiency monitoring: Low fill rates cause inefficiencies, hurting performance. And while weighed against revenue-contribution we are constantly monitoring efficiency of our serving placements.
  • Harmonic Oscillation Production: Low amounts of traffic can skew results. Just because a line item performed well on 2% of your traffic does not mean they deserve first-look on 100%. We place them and slowly ramp up their traffic to reduce risk.
  • Market Price Testing: The market fluctuates, so should your pricing strategy. It’s important to dynamically adjust your market pricing to be inline with market appetites.

Most of the time, the results on algorithmic optimization is clear quickly. Just ask Narwhal for reddit – where revenue was doubled in a week. Or Taptalk, who’s revenue increased 75%.

Table of Contents