portfolio

Segement your customer base by their preference and behavior beyond monetary spend - 

The project helped the client to understand their customer profile by conducting unsupervised machine learning given the customers' purchase data and personal differentiators. The project results helped answer how the client's customers could be grouped into different personas/profiles, given their interaction with the brand.

With this analysis and the consolidated customer lists with these persona labels, the informed client was able to conduct a marketing strategy, retail operation process, and customer relationship strategy that targeted specific groups given their engagement pattern and preferences.

Customized automation/semi-automation solutions for your business KPIs - 

Created customized CRM reporting for brands providing insights on Customer Overview - Insiders, Acquisition Channel, Repeat; Customer Profile - Geo, Gender, Payment, Age Range, Device; CRM Metrics - new vs existing, value tiers, retention, full price vs sale, data capture rate and data quality; List Growth: email subscribers, list health.

The reports provided a holistic view of the customer base and the effectiveness of marketing and operations approaches that impact business decisions, leading to strategic focuses that improve profits and margins.

Visualization on Email content effectiveness - 

My solution ended up with 200+ hours saved per year for a single task that generates and distributes more than 21k screenshots of email heatmaps to 20+ brands and countries.

We started with organizing and listing every step needed to take to run the heatmaps weekly. We solved the challenges by dividing and conquering - starting with the most painful repetitive steps, then connecting the dots and stitching all separated pieces into a whole automated project.

Utilizing Machine Learning to maximize your Email Open Rate - 

Starting with conducting a regression model on a few key features of the email subject lines (e.g., have emoji, all caps, word count, punctuation, etc.), we identified the correlations between Email performance metrics (e.g., open rate, click rate, conversion rate, revenue, etc.) and these features, then conducted consecutive A/B tests for causations analysis. The results informed the client with data-supported best practices in drafting email subject lines, which increased the metrics by 1-2% when they were already above the industrial average.

Re-activate and Re-direct your subscriber base to their most engaging communication platforms -

For this project, the challenge was to reengage and reactivate lapsed or inactive customers. To reach customers with a more personalized approach, we created a segment for different business purchases: for reactivation – Lapsed customers given their purchasing frequency, i.e., if the customer’s expected next order date in the past 30 to 60 days; for retention improvement – customers haven’t purchased for a year. 

Translate data into easy-to-digest visualization - 

This Project is a real-time heatmap visualization on the data featuring all shooting incidents reported to the New York City Police Department (NYPD) since 2006.

The map is built and customized with Mapbox API. The data is shown in two forms: Heat Map and by Cluster. Both forms allows interaction where users can zoom in and out, and hover on one single incident for more details (time, the definition of the crime, the victim's age, etc.). The visual allows user to move around and see which areas are more crime concentrated.

Understand how your cusomters engage with your brand to build more and enhance the relationships - 

This project helped our client unlock the narrative of your customer's experience with our bespoke Customer Journey Mapping - how the customers interact with the brand pre-purchase, during-purchase, and post-purchase. With detailed visual storytelling that charts every touchpoint and emotion your customers encounter, we provide insights that drive impactful strategies for engagement and growth. We also created examples based on the customer's persona to understand further and strategize customer loyalty and upgrading.

Developt the RFM model that talks to your specific business - 

The RFM (Recency, Frequency, Monetary) model is a data-driven marketing analysis we tailored for some luxury brands to enhance customer engagement strategies, particularly in sectors with distinctive consumer behavior patterns. The model segments customers based on the recency of their last purchase, the frequency of their transactions, and the monetary value they bring. The unique application of the RFM model in this context ensures that marketing efforts are more effective and efficient by focusing on the most lucrative customer segments.