top of page

YOUR VIRTUAL PERSONAL SHOPPER

PROJECT OVERVIEW

Reviewing and redesigning the live FINDMINE merchandisers’ admin interface page which allows merchants to add, view, and bulk edit hundreds of products at a time, while considering who the end user is, and the tools they will need to complete their tasks. 

ABOUT THE COMPANY

FindMine is bringing artificial intelligence to a $500B segment of commerce. Their SaaS (Software as a Service) technology automates and personalizes merchandising for online retailers and shoppers. FindMine is innovating the manual process of curating window displays, endcaps, and online lookbooks. FindMine products include recommendations on retailers' ecommerce sites (currently available), automated style guides for email campaigns, and even dynamically ­styled retargeted ads and in­store styling for shoppers and store associates. 

Problem: Retailers struggle to convert incremental sales: 50% of the top 500 retailers' products go without a single click each day. Retailers default to window displays, curated end-caps, and online "lookbooks" to service these needs, but each of these efforts is manual, static, and cannot scale without technology.

 

Opportunity: Improve the FindMine admin dashboard user experience keeping in mind that retail merchandisers and FindMine administrator/support staff are primary users.

MEASURES OF SUCCESS

Time to create and review outfits (any decrease over the current amount of time, which needs to be measured, is good)

USER RESEARCH

We interviewed four merchandisers, one product manager, two digital merchants and one operations manager to gain insight on how the retail world is run. Our key takeaways were:

 

  • The act of 'curating outfits' happens before anything gets to merchandisers.

  • Merchandisers are just grabbing lookbooks and inputting information onto the site.

  • To get outright onto a site is an extra task, whether it's for the buyer/stylist/merchandiser or Findmine Staff.

PERSONA CREATION

After interviewing the participants extensively, we looked for trends, motivations, habits and pain points to develop our four personas.

RESEARCH RECOMMENDATIONS

Can FindMine convince retail execs to give their employees new job responsibilities?

 

Short Term:

Focus on FindMine admin page
& create 24-hour zen desk
 

Potential Pain Reliever: Improve scraper for more accurate data results within FindMine

 

Long term:

Stylists/merchandisers to
create outfits with FindMine
 

Potential Pain Reliever: Introduce a FindMine plugin
for retailer CMS

SITE MAP
CURRENT - BROAD & SHALLOW

The current flat hierarchy is good for discoverability, but doesn’t make sense for FindMine because
(1) its purpose is not enable users to browse

(2) there are 8 major categories, half of which

do nothing is directly accessed on the main page

SITE MAP
PROPOSED - SHALLOW & DEEP

The proposed deep hierarchy isn’t good for discoverability, but
(1) will not confuse users with

unnecessary options
(2) allows for users to directly

do their intended task

USABILITY TESTING

 

Admin Dashboard

As a baseline, we tested 5 users for the below essential tasks on the current Admin Dashboard:

 

(1) Search for item
(2) Create an outfit
(3) Review outfit and edit

(4) Select items to bulk edit

 

We found that users (1) expected natural language input, (2) were unsure how to add items to outfits and, therefore, unable to complete the task, (3) did not know where to find previously- saved outfits and (4) did not understand which items were mislabeled and could not find the ‘bulk edit’ option.

 

Recommendation: Redesign the Admin Dashboard to simplify and (1) increase search path options, (2) make the ‘Create Outfit’ option prominent, (3) create a page for outfit management and (4) surface bulk edit options.

Current Admin Dashboad

The client's main concern and KPI was creating and reviewing outfits. While we demonstrated all four tasks to the client, I will show TWO below.

BASELINE USABILTY TEST

Task: Create an outfit and publish it to the Mango Outlet Site

*5/5 participants created an outfit in an avg. time of 1:44, but all clicked “save” to publish to the site rather than switching the enable button on. With a simple label change for enabled/disabled we anticipate at least a 40% increase in task completion from baseline and an improvement of 0:04 seconds to create and publish an outfit.

v1 Prototype Enhancements:

  1. Multiple item select

  2. “Add to outfit” button

  3. Persistent outfit canvas for less scrolling between product inventory and outfit creation canvas 4. Display enable/disable options within outfit canvas

 

RELEVANT KPI RESULTS:

 

60% Decrease in completion rate

--:-- average time to compelete task

TASK FLOW: "Creating an outfit"

CURRENT vs. PROPOSED

BASELINE USABILTY TEST

Task: Bulk edit an item(s) that has been mislabeled by the algorithm

v1 Prototype Enhancements:

  1. Flag items in need of review

  2. New bulk editing options display

 

RELEVANT KPI RESULTS:

 

20% Decrease in completion rate

--:-- average time to compelete task

*4/5 participants found our item that had been flagged for review with a red outline, and could edit categories individually, but people struggled to understand the bulk editing task

TASK FLOW: "Bulk editing items"

CURRENT vs. PROPOSED

WIREFRAMING

After all of the baseline usability testing and task flow creation we ran a design studio to develop the look of the Admin Dashboard. Combining and collaborating our ideas together we created the paper prototype and began usability testing.

HIGH FIDELITY ANNOTATED PROTOTYPE

Following the wireframing and usability testing on the paper prototype we were able to determine the behaviors of the users, what they liked, what they didn't like and iterate on that and develop the high fidelity prototype.

NEXT STEPS...

If there was more time, the next step would be to usability test on the second version of the prototype in addition to a look book type of integration for when this software expands to retailers.

After the first phase of the project was completed I was hired again to finsh the pages
that were not designed and iterate on the existing pages.

DASHBOARD UPDATE

CREATE OUTFIT

ITERATION 1

CREATE OUTFIT UPDATE

ITERATION 2

 

1

2

3

4

ITERATION #2

 

  1. The side bar has been collaped allowing for additional surface area. This opens up the page adding a 6th column for additional outfit items. When magnifiying glass is clicked the side bar pops out with filter and search options as in Iteration #1.

  2. Scrolling arrows have been added in the outfit canvas area so that more than 5 items can be added to an outfit.

  3. An area for the item name, edit and delete button has been added to each item for quick accessable updates.

  4. When the "item inventory" is filled with a lot of items the back to top arrow will appear for quick seamless navigation to the top of the page without having to use the infinate scroll.

ITEM PANEL

 

Two Item Panels. One for the FINDMINE associate to be able to update and edit. The other for the Retailer/Brand to be able to see and update their items separately.

ITEM PANEL - EDIT

FINDMINE ASSOCIATE VIEW

ITEM PANEL - VIEW DETAILS

RETAILER/BRAND VIEW

MANAGE OUTFIT

ITERATION 1

MANAGE OUTFIT UPDATE

ITERATION 2

 

MANAGE OUTFIT

ITERATION 1

MANAGE OUTFIT UPDATE

ITERATION 2

bottom of page