Zen Sport is a training service dedicated to making young athletes stronger, faster and bolder via scientifically developed training methods, they are based in New Jersey.
Phase 1: Fulfilled all aspects of complex migration for Zen Sport including planning and operations to Amazon Web Services Cloud.
Phase 2: Developed a Machine Learning model using historical data to predict future outcomes on AWS for Zen Sport which increased membership conversion.
I was the Creative Project Manager and reported to the Founder.
Worked on the project independently and developed a detailed project plan to monitor and track progress by using the right tools and methodology.
Created comprehensive project documentation for my client, including a scope document, budget, timeline diagrams and report on phases.
Successfully performed the migration within scope and budget.
Developed a machine learning model using historical data to improve membership conversion.
When you move scaling into the cloud, you experience an enormous amount of flexibility that saves both time and money for your business.
With a managed cloud provider, it is quick and easy to get the resources you need to increase your business revenue and improve your data management.
When I explained this to the owner of Zen Sport, he was interested to hear more about the process and our next moves for his growing business.
Zen Sport's Wordpress website was growing rapidly in size and complexity and the booming business needed a platform with the infrastructure that could stay ahead of the competition. So I recommended AWS Cloud.
I planned to migrate the entire Wordpress website to AWS Cloud and presented the products that are going to be used in this process to my client including features of the products, the time that we needed for each process and the costs.
To store the business data, we needed a scalable and secure cloud storage. Amazon Simple Storage Service (S3) offers all the features that we needed including data availability and performance within the budget.
To host the Wordpress website we needed a virtual environment to configure and scale our high demand of compute capacity, so we used Amazon Elastic Compute Cloud (EC2).
EC2 allows you to pay for the volume you use rather than a set fee, which made this an effective cost containment feature of this project.
Amazon Route 53 (R53) connects user requests to our choices of running infrastructure including EC2 and S3 and highly reliable domain name system for our end users, so it was an obvious choice.
I used Auto Scaling service to ensure that the Wordpress website is optimized for availability with no additional charge.
First step was to create a timeline and present the products, costs and methodology to the stakeholders.
The plan was to deliver a human-centered approach within an agile workflow to perform the process of migration and design.
I analyzed the current state of the website, and noticed a few issues that had to be resolved before starting the migration.
The most notable structure problem was the poorly designed information architecture and navigation systems. I recommended a series of categories that each linked to their own landing pages to make it clear for end users. We ran a few card sorting sessions with the stakeholders to create the new categories for products and services.
It also helped Search Engine Optimization because categories are the most prominent landing places when users search for a type of product, service or information.
Finally, before moving on to the migration phase, I created a prototype of the Wordpress dashboard to show the new web architecture to the web managers and trained them to maintain it.
To start the migration, I had to get the AWS environment ready by launching the EC2 instance to run our Wordpress website. Then installed the Wordpress from the Marketplace and transferred the domain name to Route 53.
The next step was to migrate our existing Wordpress website to AWS Cloud. I used WP Migration Plugin to transfer the website.
Performed additional data evaluation and migration to create a resourceful database for further business purposes including email marketing and staff training. Last but not least, we practiced secure governance of the cloud environment and billing.
MACHINE LEARNING MODEL
After migrating the Wordpress website successfully to the AWS Cloud environment, I planned and created a Machine Learning model to increase membership conversion for Zen Sport.
My client wanted to perform a direct marketing campaign to get a response from his open house guests. 300 people attended the open house which only 100 of them expected to respond. He didn't have a budget to reach out to the 300 attendees, therefore to minimize the cost, I created a marketing campaign to reach out to the smallest number of attendees.
Collected information on current and prospective clients into a data set and built a datasource to predict whether a client who attended the open house event would purchase a membership.
After adding the datasource to our storage in AWS Cloud, we launched the Machine Learning standard setup on the console. This was a binary classification model with the actual output of a prediction score.
Any records above the cut-off number predicted as "1" which meant the person who attended the open house was interested to purchase a membership, and if it was "0" the opposite. After testing the model successfully, we launched the marketing campaign.
BINARY CLASSIFICATION MODEL
The best score on a classification problem is 100% accuracy and almost impossible to achieve given the stochastic nature of data and algorithms.
However, we managed to maximize profits by identifying patterns among potential clients and brought new members to Zen Sport. The model was successful comparing the results to the previous year. My client had a 30% increase in membership conversion after using the model.
With the pressure of a tight budget and deadlines it was difficult to redirect the stakeholders to focus from the next feature and onto the end users. If I had a prior training session for them to practice feature desirability of end users, the project would have had a better flow.