Amazon mechanical turk intro to bpo's v3
4 likes1,950 views
The document presents an overview of Amazon Web Services' Mechanical Turk (MTurk) as a crowdsourcing platform to improve business processes. It discusses the benefits of MTurk, including elastic capacity, on-demand availability, and cost efficiency, along with common use cases like data management and content annotation. Additionally, it highlights the features of MTurk, such as access to a global workforce and a pay-as-you-go model.
![_ Presented by
John Hoskins, hoskins @ amazon.com
Mechanical Turk, Amazon
[ Before we start
]
2](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-2-320.jpg)
![How does it
High Crowdsourcing work Examples
Level
Questions
Detail What is MTurk? How is it used
[ Our plan for today ]
3](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-3-320.jpg)
![[ ]
What is Crowdsourcing?
4](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-4-320.jpg)
![[ Crowd Labor Trends ]
5](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-5-320.jpg)
![[ Crowdsourcing:
Key Advantages ]
7](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-7-320.jpg)
![[ 1. Elastic Capacity ]
8](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-8-320.jpg)
![On and Off Fast Growth
Variable peaks Predictable peaks
[ Staffing Usage Patterns ]
9](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-9-320.jpg)
![On and Off WASTE Fast Growth
Missed
Variable peaks Opportunity Predictable peaks
[ Staffing Usage Patterns ]
10](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-10-320.jpg)
![On and Off Fast Growth
Variable peaks Predictable peaks
[ Usage Patterns:
Crowdsourcing ]
11](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-11-320.jpg)
![[ 2. On Demand Availability ]
12](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-12-320.jpg)
![[ ]
3. Pay as you go, for what you use
13](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-13-320.jpg)
![$ $
Lower costs Capital efficiency
Improve productivity Focus on your business
[ Business Benefits
of Crowdsourcing ]
14](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-14-320.jpg)
![Marketplace for Work
_ Access to global workforce, on demand
_ Quality controls and workflow “building blocks”
_ Pay as you go, only when satisfied with results
_ Programmatic Access (API)
[ What is Mechanical Turk? ]
16](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-16-320.jpg)
![_ 500,000 Workers
_ 190+ Countries
[ Global Workforce
]
17 (as of Apr 19th, 2012)](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-17-320.jpg)
![[ How it works ]
19](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-19-320.jpg)
![Highest
Performing
Average
MASTERS
Unknown
[ Mturk Workforce ]
20](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-20-320.jpg)
![Highest
Performing
Average
MASTERS Unknown
Your Custom
Workforce
[ Custom Workforce ]
21](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-21-320.jpg)
![Speed
Price
Accuracy
[ Balances & Levers
]
22](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-22-320.jpg)
![Data Management Analysis/Research
_ Verification _ Sentiment
_ Entry & Collection _ Relevance
_ De-dupe _ Online Research
_ Algorithm Training
Content Management Annotation
_ Moderation (Photos, Content) _ Tagging
_ Transcription _ Classification
_ Localization/Translation
[ Common Use Cases ]
24](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-24-320.jpg)
![How can we assess relevance of
tweets in real-time?
How can we research and categorize
merchants in our database?
How can we verify accuracy of low
confidence OCR scans?
How do we build a dictionary to
categorize ads on the internet?
[ Examples ]
25](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-25-320.jpg)
![Categorization
Web Research
Transcription
Categorization and DeDupe
[ Examples ]
26](https://image.slidesharecdn.com/amazonmechanicalturkintrotobposv3-130208105902-phpapp02/85/Amazon-mechanical-turk-intro-to-bpo-s-v3-26-320.jpg)
Amazon mechanical turk intro to bpo's v3
- 1. amazon web services Crowdsourcing for improving Business Process October, 2012 - Optimized for 1280x960 - © Amazon Web Services
- 2. _ Presented by John Hoskins, hoskins @ amazon.com Mechanical Turk, Amazon [ Before we start ] 2
- 3. How does it High Crowdsourcing work Examples Level Questions Detail What is MTurk? How is it used [ Our plan for today ] 3
- 4. [ ] What is Crowdsourcing? 4
- 5. [ Crowd Labor Trends ] 5
- 6. 1. Crowdsourcing 6
- 7. [ Crowdsourcing: Key Advantages ] 7
- 8. [ 1. Elastic Capacity ] 8
- 9. On and Off Fast Growth Variable peaks Predictable peaks [ Staffing Usage Patterns ] 9
- 10. On and Off WASTE Fast Growth Missed Variable peaks Opportunity Predictable peaks [ Staffing Usage Patterns ] 10
- 11. On and Off Fast Growth Variable peaks Predictable peaks [ Usage Patterns: Crowdsourcing ] 11
- 12. [ 2. On Demand Availability ] 12
- 13. [ ] 3. Pay as you go, for what you use 13
- 14. $ $ Lower costs Capital efficiency Improve productivity Focus on your business [ Business Benefits of Crowdsourcing ] 14
- 15. 2. What is MTurk? 15
- 16. Marketplace for Work _ Access to global workforce, on demand _ Quality controls and workflow “building blocks” _ Pay as you go, only when satisfied with results _ Programmatic Access (API) [ What is Mechanical Turk? ] 16
- 17. _ 500,000 Workers _ 190+ Countries [ Global Workforce ] 17 (as of Apr 19th, 2012)
- 18. 3. How it works 18
- 19. [ How it works ] 19
- 20. Highest Performing Average MASTERS Unknown [ Mturk Workforce ] 20
- 21. Highest Performing Average MASTERS Unknown Your Custom Workforce [ Custom Workforce ] 21
- 22. Speed Price Accuracy [ Balances & Levers ] 22
- 23. 4. How is it used? 23
- 24. Data Management Analysis/Research _ Verification _ Sentiment _ Entry & Collection _ Relevance _ De-dupe _ Online Research _ Algorithm Training Content Management Annotation _ Moderation (Photos, Content) _ Tagging _ Transcription _ Classification _ Localization/Translation [ Common Use Cases ] 24
- 25. How can we assess relevance of tweets in real-time? How can we research and categorize merchants in our database? How can we verify accuracy of low confidence OCR scans? How do we build a dictionary to categorize ads on the internet? [ Examples ] 25
- 26. Categorization Web Research Transcription Categorization and DeDupe [ Examples ] 26
- 27. amazon web services http://www.mturk.com John Hoskins, Mechanical Turk hoskins@amazon.com 27