SEM campaign builder
Product Requirements
Version 4.0
Last Saved on 09-30-2011
Author: Ruth Shen
Status: In Progress
Page 1 of 22
1 Document Background
1.1 Document Owner & Scope
This document defines the Search Engine Marketing (SEM) Campaign builder for MerchantCircle.com ARB/SMB project, a too that will build and populate SEM campaign with multiple categories and cities. This structure can used to create a baseline for future SEM initiative and forecast.
1.2 Document Approvers List with Due Date
1.3 Document Reviewers List
1.4 Revision History
2 Table of Contents
1 Document Background 2
1.1 Document Owner & Scope 2
1.2 Document Approvers List with Due Date 2
1.3 Document Reviewers List 2
1.4 Revision History 3
2 Table of Contents 4
3 Overview 5
3.1 Executive Summary 5
3.2 Key Concepts, Terms & Definitions 5
3.3 Product Objective 5
3.4 Key Value Proposition 5
3.5 Problem Assessment 5
3.6 Competitive Space Assessment (Recommended) 6
3.7 Business Cost Benefit Analysis (Optional) 6
4 Summary of Systems Impacted (Optional) 7
5 Functional Requirements 8
6 Technical Requirements 10
7 Security & Privacy Requirements 11
8 Data Requirements 12
9 International / Globalization Requirements 13
10 Other Applicable Requirements (Optional) 14
11 Legal and Contractual Issues 15
12 Use Cases 16
12.1 Use Case 1: Inconsistent Date Ranges 16
12.2 Use Case 2: Random External Event 16
12.3 Use Case 3: Incomplete Picture 16https://docs.google.com/Doc?docid=0ATGnxOx82NLdZGQ1a2RmY3pfMTJ4OTVxajRmag&hl=en
12.4 Use Case 4: Analysis Error 17
12.5 Use Case 5: One Bad Subdomain Spoils the Whole Bunch? 17
13 Key Dependencies 18
14 Risks and Mitigation Plan (Recommended) 19
15 Future Considerations 20
16 Open Issues 21
17 Appendix (Optional) 22
17.1 Research/Prototyping Data 22
3 Overview
3.1 Executive Summary
We will design a SEM campaign to deliver traffic to MerchantCircle.com listing pages that are currently receiving CG payout.
The SEM campaign builder creates 5 cities and 5 categories in Google and Yahoo/Bing SEM accounts . It will insert 25 (5 cities x 5 categories) the adgroup structure and category name as first keyword. The campaign will have both Geo targeted and national targeted. This tool is designed to support more than 16,000 adgroups. (414 cities x top 40 categories = 16,560)
3.2 Key Concepts, Terms & Definitions
- SEM – Search Engine Marketing (Google Adwords, Yahoo/Bing Adcenter)
- Adgroup – 3nd level structure in Adwords and Adcenter where to keywords, ads, geo targeting, destination URL are inserted
- Campaign – 2nd level structures (limited to 100 for Google) In our case, campaigns will be US states.
- Account Level – 1st level
- CG – CityGrid – advertising platform for CitySearch
3.3 Product Objective
The SEM campaign builder will:
- Populate Google & Yahoo/Bing campaign with more than 16,000 adgroups.
- Populate each adgroups with 1 keyword (keyword = category name)
- Create a baseline and prepare use to forecast search demand that may be direct to traffic demand (CityGrid’s merchants).
3.4 Key Value Proposition
The Ad Delivery Report will provide visibility into domain performance, enabling better decision-making when using the Domain Blocking feature.
3.5 Problem Assessment
Advertisers need to know where their ads are being displayed to monitor partner quality. At present, advertisers who have Full Analytics are trying to use the existing Traffic Referrer Report to determine which domains to block. However, this report tells users the performance metrics associated with the referring URL. The referring URL is not necessarily the same thing as the URL that the ad was actually delivered/displayed. Examples of problem use cases include:
- Example 1: Content Match with iFrames. Content Match inserts overture.com in the referring URL. This is misleading for advertisers, as the iFrame is actually appearing on a completely different site.
- Example 2: Partners using a 3rd party redirect for capturing traffic metrics. The URL of the redirect site will appear in the traffic referrer report rather than the URL of the site on which the ad was actually displayed.
- Example 3: Inaccuracy and/or gaming. The value in the domain field can be inaccurately populated by accident or intentionally populated with inaccurate information by the partner. The accuracy of this information is not checked in an automated fashion.
- Example 4: 3rd Party Affiliate Sites. Some partners may syndicate our ads to various affiliate sites. The URL displayed belongs to the primary partner, not to the affiliate sites. If the advertiser blocks the partner due to bad traffic from a few different affiliate sites, he/she may also be cutting off good traffic from affiliate sites.
- Example 5: Email channel. Some URLs are unknown, as the ads may be appearing in an email.
We are currently advising advertisers to read their own web logs to get a sense of where their ads are being displayed. Advertisers want YSM to provide information to help them understand where their ads are being displayed.
A new, corresponding report for EWS would also be required; however, this report may not be available at the same time.
3.6 Competitive Space Assessment (Recommended)
Google currently provides a Domain Blocking feature. Please refer to the Domain Blocking PRD at http://twiki.corp.yahoo.com/view/Panama/DomainBlocking to view a comparison between Google and YSM’s Domain Blocking capabilities.
3.7 Business Cost Benefit Analysis (Optional)
The Domain Blocking feature enables advertisers to block their ads from appearing on various domains, which may have a negative impact on revenue for both Yahoo! and the publisher.
4 Summary of Systems Impacted (Optional)
5 Functional Requirements
6 Technical Requirements
7 Security & Privacy Requirements
8 Data Requirements
9 International / Globalization Requirements
10 Other Applicable Requirements (Optional)
11 Legal and Contractual Issues
12 Use Cases
12.1 Use Case 1: Inconsistent Date Ranges
- Advertiser has a conversion cycle of 30 days.
- Advertiser runs the Ad Delivery Report and selects a date range spanning one week (11/1 through 11/7).
- Certain partner URLs appear to be low performing that week, though over the course of the month, they perform quite well.
- The advertiser blocks those URLs, unaware that the time period selected does not represent the performance of the domain.
- Issue: The advertiser should look at a 30 day date range to evaluate the effectiveness of the URLs over an entire conversion cycle.
12.2 Use Case 2: Random External Event
- Jensense.com, a popular blogger/publisher, goes on vacation for 2 weeks and leaves a note on her blog that she’ll be back on a certain future date.
- 1 week later, an advertiser runs the Ad Delivery Report.
- Jensense.com, usually a source of great quality traffic, appears with 0 conversions for the time range.
- The advertiser decides to block that domain based on the report metrics.
- Issue: The advertiser may not realize that Jensense.com is actually a great source of conversions. With the author out on vacation and the conversion data low, the advertiser may incorrectly assume that the low conversions are the norm rather than the exception.
12.3 Use Case 3: Incomplete Picture
- Advertiser owns a bakery and sells wedding cakes.
- Advertiser runs the Ad Delivery Report.
- Advertiser sorts by Clicks, then takes the publishers with the lowest clicks and blocks their domains.
- Advertiser has turned on analytics, but has not correctly tagged his web site.
- As a result of the tagging issue, conversion data is not captured for each URL.
- The advertiser cannot see conversions in the Ad Delivery Report, so the advertiser blocks several domains that have a low CTR but a high conversion rate.
- Issue: The advertiser is making decisions based on incomplete data.
12.4 Use Case 4: Analysis Error
- Advertiser owns a web site and sells baby items.
- Advertiser runs the Ad Delivery Report.
- All of the data is displayed in the report; however, the advertiser decides to block all publishers based solely on CTR.
- The advertiser allows ads to be displayed on a certain publisher site with a high CTR. However, that publisher’s domain had a cost and avg. cpc that was 10x higher than the other publishers, and with very low conversions.
- Issue: Even with all of the data available, advertisers may fail to look at more than one variable, causing them to fail to draw the right conclusion even when all of the data is presented to them in the report. This in turn hurts both the advertiser and the publishers he blocks as well as Yahoo!.
12.5 Use Case 5: One Bad Subdomain Spoils the Whole Bunch?
- Advertiser has done everything right: implemented analytics, tagged his web site, displayed all columns in the Ad Delivery Report.
- To determine which domains to block, advertiser compares all domains vs. each other.
- Advertiser uses all of the data provided in the Ad Delivery Report and blocks the 250 domains that have lowest conversions, highest cost, lowest CTR, etc.
- Issue: If an advertiser only looks at the domain level, he may not be blocking “properly” or “optimally,” as one bad subdomain can bring down the average for the entire domain. Advertisers must look at both the domain level and the URL level to understand a specific subdomain’s impact on the overall quality of the domain. Users cannot tell when one back subdomain spoils the average of an otherwise top performing domain.
13 Key Dependencies
14 Risks and Mitigation Plan (Recommended)
[Building from the Risks and Mitigation Plan from the MRD, present an assessment of the risks surrounding this project and the strategies for mitigating those risks.]
15 Future Considerations
16 Open Issues
17 Appendix (Optional)
17.1 Research/Prototyping Data
[List of data sources, methodologies, and raw data used to prepare this PRD.]
17.2 References / Links
Example: Google AdWords Placement Report
Confidential and proprietary | For Internal Yahoo! Use Only | Not to be Distributed Outside of Yahoo!
Page 5 of 22
No comments:
Post a Comment
If you have any question, please contact me at rs.ucla.zot @gmail.com