Reflektion (acquired by Sitecore in 2021) employed machine learning and developed a next-generation personalization platform that raised conversion and revenue for leading digital retailers.
I was a full end-to-end designer for the platform for many years until it was acquired. I did everything from research to delivering spec. We also served companies of varying revenue, from <50M USD -> 10B+ USD. So when I was designing, I always had scalability in mind.
Provided high fidelity designs of e-commerce company's end user websites with Reflektion widgets.
I inspected their site, learned the style, and designed our widgets in their pages to match. This helped increase sales drastically for existing and potential customers.
A state-of-the-art site analytics reporting module.
Over several years of work, I designed a full analytics platform tailored to e-commerce.
A greater insight into products.
Users got a full analytics performance experience for their product catalog that enabled them to manage how they were promoted.
Understand widget performance.
Our customers loved our analytics because they could further understand visitor behavior and create greater promotional opportunities.
A/B testing that identified better performance.
When designing this report, I heavily researched what a data analyst wants to see when determining statistical significance.
Extensive analytics on your stores .
Before I designed this, we interviewed existing customers. From those discussions, I focused on top stores, opportunities for BOPIS (Buy Online Pickup In Store), and which departments were selling the most. The idea was to give customers the big picture with their website and physical stores.
A greater understanding of inventory.
Customers could see which products were doing well, not selling, and completely selling out.
Locale reporting for global brands.
Customers had locale reports that gave them reports across global regions.
Triage any issues.
Engineers could understand any issues with their widgets on their website and fix them.
An extensive data ingestion system.
Customers could manage their data feeds, create and map attributes to their existing data catalog, and perform data transformations. The idea was to make a generic system, where any attribute can be used anywhere.
A catalog for customers to view their products, fitments, articles, and fitments.
When designing this, I followed industry-standard practices of catalog management: (a) list page (b) toggles & filters (c) tabbed detail pages
Be notified of anything.
Notifications helped keep the users of the platform appraised if there were any feed issues, widget issues, or when certain metrics past a given threshold.
Full click and view event monitoring.
In order to ensure all events were firing correctly, I designed an event monitoring interface. A user inputs their UUID, then browses and clicks around on their website, and view/click widget events are recorded.
Diagnose any widget.
Additionally, I designed a diagnostic feature for our search widgets to allow clients to diagnose the widget. This ensured all products were appearing in their correct position.
A visual data modeler to create algorithms.
A user dragged and dropped node functions on a canvas, linked them together, and edited the parameters of each node. These algorithms determined what products were displayed in the Reflektion widgets on the customer’s website.
When researching this, I looked at node editors like Microsoft Visio or ITSM platforms. Typically, you’d see a visual representation of nodes, have a way to connect them together, and some way to edit the properties.
Host your page templates.
Users could preview the template in rendered via HTML format, as well as revert back to a previous template, or configuration via version history.
Manage your contacts.
We created the ability to manage contacts in the platform. The UX pattern was to create a list page, and then an edit page where a user could edit basic details and upload a profile photo.
Promote products in email.
I designed the ability for customers to create a promotional email campaign by writing rules and previewing the outcome. This widget would display products in a promotional email.
