FreshDirect

The Process

We first evaluated the entire site via site audit and came to the conclusion that every core action in the modification flow revolves around a mode since the site logic treats regular shopping and modifying a completed order in the same manner page to page. At first, we explored the thought of having a “contained” modify experience on one screen but tech constraints didn’t allow this possibility without re-platforming the entire site (ouch). However, we moved forward with our hypothesis which was if we could create a clear indication of “mode” for modifying an upcoming order, modification completion percentage would drastically improve. Since carts with modification have a roughly $30 higher AOS, higher completion of order modifications would lead to a substantial revenue spike.

We broke the project apart into three separate phases based on the perceived level of effort-

1. Optimizing existing elements on the page that already exists via coloring, labeling and copy enhancements and removing any unnecessary features 

2. Adding new elements to existing pages to further clarify the modify mode,

3. Re-designing new key pages including Order History, Shopping Cart and Order confirmation pages. My focus was mostly involved refining the cart page to accommodate both modified order and standing order (B2B recurring order) functionality and cleaning up the UX for the order history page.

Each phase would have its own rounds of designing, user testing, spec writing, and development sprints.

Concept Design:

Testing:

Refining:

Documenting:

The Result

Without even releasing the final phase of the project, FreshDirect is on pace for an additional 700-800 modifications daily after these released changes. If the pattern based on a month of tracking GA data holds true, then the company should gain roughly $2-4 million in additional revenue from these enhancements. We have sustained 45% increase on order mods (roughly 700 per day) with a $26 increase in average order.

.

The Opportunity

Unfortunately, we were in the process of migrating our analytics to GA (great timing) so asking for data wasn’t as easy as we hoped and we needed to tag all elements of the order modification flow in order to understand how many users complete the process. Once we spent a month properly tracking order modification, we realized the market opportunity was quite immense - One out of every four users who selected the modify order CTA did not even complete the process 🙃. This is even worse when you consider that the average order size goes from $126 to $152 when a modification is completed. In other words, we were basically telling users that we didn’t want their money because of the classic scenario of building a rushed MVP that was never iterated upon for years (damn you waterfall).

The Problem

FreshDirect is the original online e-grocer with a rich 20 history and a unique business model. Since the company started prior to the great e-commerce wave pioneered by Amazon, its business internally controls all aspects of their supply chain - from sourcing directly from the farm, to order packing, to delivering orders straight to customers’ doorsteps. Due to this self-contained model (outsourcing, pshhhh who needs it), the FD online store has unique features that its competitors don’t have the control to offer but some of them have not been refined for years and needed the cobwebs dusted off. Our UX Director pinpointed the potential market value of our order modification feature, which allows users to add, subtract from or edit the delivery details of their order up to 48 hours before it’s delivered. He knew that the main customer use for this feature was adding more products to the cart, so we all dug a little deeper to figure out the market value of optimizing order modification.