# 8.4.7 R8. Google AdWords - Video 6: A Greedy Approach | Summary and Q&A

526 views
December 13, 2018
by
MIT OpenCourseWare

## TL;DR

This video explains a greedy approach to allocating ads to queries, prioritizing combinations based on average price per display.

## Install to Summarize YouTube Videos and Get Transcripts

### Q: What is the difference between the greedy approach and linear optimization in ad allocation?

The greedy approach prioritizes combinations based on average price per display and allocates ads sequentially. Linear optimization, on the other hand, finds the optimal allocation by considering all combinations simultaneously to maximize revenue. Greedy approach may not yield the same revenue as optimization.

### Q: How are budget limits and query limits calculated in the greedy approach?

Budget limits are determined by dividing the remaining budget of an advertiser by the average price per display. Query limits are based on the number of remaining unallocated displays for a specific query. The smaller of the two limits determines the number of displays allocated.

### Q: Can multiple advertisers have the same average price per display for a specific query?

Yes, multiple advertisers can have the same average price per display for a query. In such cases, any of those advertisers' combinations can be chosen for allocation. The choice may depend on other factors or be arbitrary.

### Q: How does the greedy approach handle scenarios where unallocated displays are exhausted?

Once all displays for a query have been allocated to advertisers, the greedy approach moves on to the next stage of allocation. The remaining cells for that query in the average price per display table are highlighted to indicate they can no longer be used.

## Summary & Key Takeaways

• The video discusses using a greedy approach to allocate ads to queries, sequentially prioritizing combinations based on average price per display.

• The allocation process uses a spreadsheet to track budget limits and unallocated displays for each combination of advertiser and query.

• The process involves selecting the combination with the highest average price per display, calculating budget and query limits, and determining the number of displays to allocate.