samgjohn / experimental-pipeline Goto Github PK

1. Start Date
2. Objective
3. Hypothesis
4. Experiment Design
5. Expected Results
6. Expected Resources
7. Gut Feeling
8. Results
9. Learning
10. Action Items
11. End Date

1/17

1. Start Date
2. Objective
3. Hypothesis
4. Experiment Design
5. Expected Results
6. Expected Resources
7. Gut Feeling
8. Results
9. Learning
10. Action Items
11. End Date

1. Start Date: 1/11
2. Objective: Rapidly close gap between last quarter's peak and current conversions
3. Hypothesis: Either search with display select or non modified broad match keywords will increase overall conversions drastically
4. Experiment Design: Created a Search with Displaoy SElect campaign and loaded it with all converting keywords from last year
5. Expected Results: At least double daily conversions
6. Expected Resources: Ad Spend, SEM
7. Gut Feeling: Should work
8. Results
9. Learning
10. Action Items
11. End Date

1. Start Date: 1/10
2. Objective: Increase Impressions / Run percent of ad
3. Hypothesis: Taking out the word "BitBucket" from ads will allow them to run much more frequently
4. Experiment Design: Campaign - Search Signup Non Branded. AdGroup - CI/CD
5. Expected Results: 1000% increase in impressions
6. Expected Resources: Low
7. Gut Feeling
8. Results
9. Learning
10. Action Items
11. End Date

1. Start Date: 1/11
2. Objective: Increase CTR
3. Hypothesis: Responsive ads will have a better CTR than Text ads
4. Experiment Design: Created a responsive ad with same text asPhantomJS specific ad #10.
5. Expected Results: Increase CTR by 50%
6. Expected Resources: Low
7. Gut Feeling
8. Results
9. Learning
10. Action Items
11. End Date

1. Objective: Increase the number of impressions and ultimately conversions for branded keywords containing only terms like circleci.
2. Hypothesis: Increased impressions will create more conversions
3. Experiment Design: compare "+circleci", "+circle-ci", "+circle ci" with "circleci", "circle-ci", "circle ci"
4. Expected Results: 50% increase in impressions 25% increase in conversions
5. Expected Resources: 1hr of marketing time
6. Gut Feeling: broad match is better for modified broad when it comes to branded
7. Results
8. Learning
9. Action Items

1/18

KW: +Heroku continuous integration

1/17

1/18

Bring down cpc

1. Start Date: 11/17
2. Objective: Increase Clicks / Conversions
3. Hypothesis: A more specific ad, actually including the word "PhantomJS" will double CTR
4. Experiment Design: Create a new ad in the same ad group
5. Expected Results: 100% increase in CTR
6. Expected Resources: SEM
7. Gut Feeling: May not work
8. Results
9. Learning
10. Action Items
11. End Date