adspace measurement and metrics — arjun jayaram

13
PANELISTS: Arjun Jayaram, VP of Engineering, Become.com Jack Mardack, Director of Marketing, EventBrite Shelley Ellis, CEO, Shelley Ellis Consulting Measurement and Metrics

Upload: adtechfan

Post on 15-May-2015

451 views

Category:

Technology


0 download

TRANSCRIPT

Page 1: ADSPACE Measurement and Metrics — Arjun Jayaram

PANELISTS:Arjun Jayaram, VP of Engineering, Become.com

Jack Mardack, Director of Marketing, EventBrite

Shelley Ellis, CEO, Shelley Ellis Consulting

Measurement and Metrics

Page 2: ADSPACE Measurement and Metrics — Arjun Jayaram

PANELIST:Arjun Jayaram, VP of Engineering Become.com

Measurement and Metrics: Scalability, Optimization and Measurement for your content advertising

Page 3: ADSPACE Measurement and Metrics — Arjun Jayaram

Size of SEM Portfolio at Become.com• Developed the SEM platform internally. Initially for keyword search and then extended for content

• Active portfolio of over 8 million keywords. Keyword corpus of over 50 million keywords

• Built the Keyword Generation and analysis

• Bid Optimization at keyword, account and campaign levels

• Tracking, Reporting and analytics by channel and by session

• Had to retool our processes for content networks

Page 4: ADSPACE Measurement and Metrics — Arjun Jayaram

Different strategies for keyword and content ?• Unlike search, the user is not typing in something in a content network

• There is an order of magnitude more web sites that use Google content ads over search ads

• In search, if a keyword is underperforming, either your match type is sub optimal or the keyword is bad (small variance for position and network)

• In content, the same ad can perform very differently across domains

• The ad generation, tracking, and optimization is very different in keywords versus content.

Page 5: ADSPACE Measurement and Metrics — Arjun Jayaram

Ad Generation strategy for content• In content match, your ad gets displayed (not your keyword)

• You need to help the search engine determine when to display your advertisement. This is the ‘related keywords’ for Google

• Researching on the ‘Related keywords’ is very important for the success of your ad. Semantically related keywords are likely to improve the performance

• If you have very similar ads, they can compete with each other – Evening dresses, cocktail dresses, prom dresses etc.

Page 6: ADSPACE Measurement and Metrics — Arjun Jayaram

Keyword Lifecycle in content

• The first couple of days of an ad , the traffic is low• It ramps up very quickly • It shows the day of the week cyclicality • There is more volatility in traffic for content ads

Page 7: ADSPACE Measurement and Metrics — Arjun Jayaram

Keyword death spiral for content keywords• Unlike Search, you need to be with in 5 positions for your ad to be shown on a page

• Google tries to show an 5 ads for ever page for each publisher. So a near identical page is likely to have different set of 5 ads

• If you lower the CPC rates, your ads are likely to show up on lower quality networks.

• Your ROI may actually hurt instead of improve by dropping CPC rates

• Don’t try to low ball your bid initially. Also be aggressive in trimming ads instead of lowering CPC aggressively.

Page 8: ADSPACE Measurement and Metrics — Arjun Jayaram

Keyword death spiral for content keywords

Page 9: ADSPACE Measurement and Metrics — Arjun Jayaram

Optimal keyword life cycle

Page 10: ADSPACE Measurement and Metrics — Arjun Jayaram

Bidding strategies for content

• Most publisher pages display three content ads per page

• If you are below third position, your ad is likely to show up on ‘lower quality’ sites

•Determine the optimal traffic quality that is required. There are distinct tiers for Quality/Price ratio

• Content ads display good bid compression. So your cost is likely to go down over time

Page 11: ADSPACE Measurement and Metrics — Arjun Jayaram

Bid Dynamics over time

This is for one ad

Page 12: ADSPACE Measurement and Metrics — Arjun Jayaram

URL tracking and analytics

• Tracking performance of content is different than search.

• In Search you look at performance of each ad individually

• In content, each ad should be further broken up by its performance per URL/Domain

• Quality is judged on performance of the Ad to a URL

• Conversion should be tracked on a URL basis

• Treat each URL as an individual listing rather than the ad group as a whole

Page 13: ADSPACE Measurement and Metrics — Arjun Jayaram

Conclusions

•Contextual ads is a good and reliable source of traffic

• Your keyword generation strategy needs to be optimized

•Measurement metrics – KPIs measure

• Conversions measured at Ad/URL level

• CTR of existing ads when you publish new ads

• CPC, conversion trend over time