Compare Proposal

Nothing to compare.

BME9175 - A/B testing

  • Posted at : 4 months ago
  • Post Similar Project
2000

Budget
2
Proposals
92
Views
Awarded
Status
Skills Required

Posted By -

QT

5.0
Projects Posted : 517
Projects Paid : 258
Services Purchased : 0
Total Spent :
9734
Feedbacks : 91 %

Project Details show (+) hide (-)

Word- 1000
Deadline- 7th Dec 


The aim of the assessment is for you to think through and plan a design for an A/B experiment. In order to do this, please follow these steps:

  • 1. Choose a webpage. Find a company you like and choose a page on the company’s website where you would like to run your A/B experiment. This can be the landing page, a product page, a payment page, etc. Importantly, you should choose only one webpage. Tell us why you find this webpage interesting from a business and personal interest perspective. 
  • 2. Identify goals.  What is your objective and why is this important to the company? (To increase sales revenues? Profits? Customer volume? Email signups? Click through rates? etc)
  • 3. Track outcomes.  What outcomes will you need to track in this context? 
  • 4. Generate hypotheses.  What A/B ideas do you have and why do you think they will help you attain your objective? Rank the ideas in order of effectiveness and ease of implementation.
  • 5. Define your intervention(s)/treatment(s). Decide precisely which ideas to test and design prototypes. You have the status quo (what is currently on the webpage). In addition, you (literally) have prototypes for 3 additional treatments (i.e. an A/B/C/D design). What do you expect the effect of the treatments (relative to the status quo) to be?
  • 6. Select your unit of randomization. What is your unit of randomization? Webpage visit? Individual user? Session? etc.)
  • 7. Consider threats to validity. What threats to internal validity are you worried about and how do you intend to deal with them?
Wordcount how much is needed i would say 1000 words Plus design prototype I need to select a single page in a website ideal related to boost sale and increase conversion rate With 4 treatment groups different from the control group