Tuesday, July 18, 2006
Yahoo advertising targeting presentation July 18
Presented by Madhu Vudali as part of BI Sig:
Yahoo advertising:
Adv A:
- 100M impressions
- Y! Music
- US-IP, 18-34
- Aug1 – Sep 30
- Guaranteed (pay according to number of impressions only)
Adv B:
- 100M impressions
- Y! “run of network”
- “Autos behavior”
- Aug1- Sep 30
- Guaranteed
These two adv are competing for same pool potentially..
Classic example of airlines conflicts
Yield mgmt:
Right customer, right price etc…
- Perishable entry
- Reservation based
- Price discrimination: willingness to pay
- Prediction models : when to expect what traffic…
- Control of resource: lies with the seller here (serve in whatever order we want)
Pricing, Yield mgmt group for 4 years:
Pricing, inventory allocation etc.
Class 1 (guaranteed)
Class 2 (non guaranteed) : floor = min price to guarantee
Inventory not used by class 1 goes to class 2
Prices may change dynamically, but individual sales contracts are not affected / goodwill etc.
Impressions vs CPM (1 : 10)
Advertisers typically tell the total amount of dollars… number of impressions would be less or more
Segmentation
Endemic vs non endemic: e.g. movie sellers on yahoo movie: more efficient, cheaper than going on yahoo mail which could be provided
Brand marketers vs direct marketers : direct ones more price insensitive
Pricing:
= time of purchase
- length of campaign
- rich media
- delivery period : e.g. thanksgiving to Christmas period has higher base
Rapt application for pricing
Supply forecasting for:
- Seasonality
- Volatility
- Attributes of contents…..
Inventory pooling problem:
Probabilistic solutions to overlaps
Look at overbooking option:
Spoilage cost, overbooking cost
Ben Vinw or someone from Sabre… for pricing diagrams
Exception processing:
- Set a floor price: walkaway
o Sales reps can do anything above that, incentives for higher prices
Maximize ad campaign delivery as it is progressing; changing the placement etc.
- Prioritization:
o If probability of targeted consumer is low (rare), give that ad a priority over generic ones…
Yahoo advertising:
Adv A:
- 100M impressions
- Y! Music
- US-IP, 18-34
- Aug1 – Sep 30
- Guaranteed (pay according to number of impressions only)
Adv B:
- 100M impressions
- Y! “run of network”
- “Autos behavior”
- Aug1- Sep 30
- Guaranteed
These two adv are competing for same pool potentially..
Classic example of airlines conflicts
Yield mgmt:
Right customer, right price etc…
- Perishable entry
- Reservation based
- Price discrimination: willingness to pay
- Prediction models : when to expect what traffic…
- Control of resource: lies with the seller here (serve in whatever order we want)
Pricing, Yield mgmt group for 4 years:
Pricing, inventory allocation etc.
Class 1 (guaranteed)
Class 2 (non guaranteed) : floor = min price to guarantee
Inventory not used by class 1 goes to class 2
Prices may change dynamically, but individual sales contracts are not affected / goodwill etc.
Impressions vs CPM (1 : 10)
Advertisers typically tell the total amount of dollars… number of impressions would be less or more
Segmentation
Endemic vs non endemic: e.g. movie sellers on yahoo movie: more efficient, cheaper than going on yahoo mail which could be provided
Brand marketers vs direct marketers : direct ones more price insensitive
Pricing:
= time of purchase
- length of campaign
- rich media
- delivery period : e.g. thanksgiving to Christmas period has higher base
Rapt application for pricing
Supply forecasting for:
- Seasonality
- Volatility
- Attributes of contents…..
Inventory pooling problem:
Probabilistic solutions to overlaps
Look at overbooking option:
Spoilage cost, overbooking cost
Ben Vinw or someone from Sabre… for pricing diagrams
Exception processing:
- Set a floor price: walkaway
o Sales reps can do anything above that, incentives for higher prices
Maximize ad campaign delivery as it is progressing; changing the placement etc.
- Prioritization:
o If probability of targeted consumer is low (rare), give that ad a priority over generic ones…