Making product uploads faster and more accurate for jewelry sellers

50%

Less time to upload

40%

Increase in seller satisfaction

85%

Reduction in upload errors

ABOUT BRAND

Kearlit is a jewelry e-commerce platform I designed from the ground up. The founder came to me with a vision to help jewelry sellers manage inventory and sell online more efficiently.

DURATION

4 Weeks

MY ROLE

My role was end-to-end:

Research the market,

Understand seller pain points,

Design the product, and deliver a system that could scale from day one.

TEAM

Chinmay Yadav - Full-stack engineer

Shivam Agrawal - Founder

SETTING THE CONTEXT

For an online jewelry seller, the typical journey starts with:

USER RESEARCH

Interviews revealed that in the current market jewelry sellers were spending hours on manual uploads and error checking

Fewer Uploads

Online uploads are tedious and time-consuming resulting in delays and errors, impacting the efficiency of their uploading process

Fewer Uploads

Online uploads are tedious and time-consuming resulting in delays and errors, impacting the efficiency of their uploading process

Manual Checking Errors

Reliance on manual checking for mismatch entries during bulk uploads can be time-consuming and prone to errors

Manual Checking Errors

Reliance on manual checking for mismatch entries during bulk uploads can be time-consuming and prone to errors

Inventory Inaccuracy

Keeping track of inventory updates manually resulted in inaccurate stock levels, making it hard to know the actual available stock at any given time

Inventory Inaccuracy

Keeping track of inventory updates manually resulted in inaccurate stock levels, making it hard to know the actual available stock at any given time

OVERARCHING GOAL

How might we create a platform for sellers to effectively upload their products in less time, and avoid manual mismatch checks, thereby boosting sales and operational efficiency?

CHALLENGES

Due to limited engineering resources, instead of building the full in-product system, I utilized Pendo, Braze and Sigma to solve this at low cost.

Complex Design Brief

The brief was to design a simple e-commerce site, but its complexity became clear later.

Fluctuating Deadlines

The project timeline was initially set for 2 months but was cut to 4 weeks due to client demands.

Stakeholder Management

The client insisted on closely mirroring numerous references, making it challenging to align their preferences.

SOLUTIONS

Goal 1: Reduce upload time and errors for single uploads

From a list of ideas this proved to be high impact low effort idea

DIRECTION 1

Scrollable Form

Reduce errors by breaking down of steps

No Progress Indication

Overwhelming for Users

DIRECTION 2

Expanding Accordions

Divided content into collapsible sections for a less overwhelming experience.

Improved focus on individual sections, reducing errors.

Risk of overlooking unexpanded sections, causing incomplete submissions.

Constant opening and closing of sections can be cumbersome.

DIRECTION 3

Navigating through tabs

Tabs provided a clearer division of steps, improving navigation.

Users could focus on one tab at a time, reducing errors.

Tabs lacked inherent progress indicators.

Multiple tabs caused confusion and potential accidental slips.

FINAL DECISION

Progress Stepper

The progress stepper clearly indicates completed and remaining steps.

Breaking down the process reduces cognitive load and errors.

Guides users through a logical sequence, ensuring all necessary information is entered.

Our user testing results proved that new sellers required

‍50% less time to upload products, thanks to our intuitive and straightforward stepper interface

SOLUTIONS

Goal 2: Reduce manual mismatch check for multiple uploads

Interviewees found the manual process of uploading and reviewing product details tedious and error-prone. Competitive analysis helped me redesign it into four streamlined steps for efficiency.

STEP 1

Bulk .csv Upload

Sellers begin by uploading their .csv files containing all product details.

STEP 2

Column Matching & Validation

To ensure the system reads the data correctly, the seller needs to match their columns with Kearlit's field names. This step is required only for the first upload, and the system remembers these settings for future uploads

This is done via a dropdown menu that shows a match percentage next to each option, helping them make quicker selections

After selecting column names, the system generates a sample value from the .csv file

A "tick" indicates a match; a "cross" shows a mismatch. Unmatched columns are corrected with a replace option

STEP 3

Image Upload

Next, sellers upload product images, with the system checking if the image titles match the corresponding .csv titles

The preview of the uploaded files is visible for the user's reference with options to view or delete the file.

STEP 4

Image Verification

The system flags titles as "Matched" or "Mismatched" and suggests the correct title name

It provides an option to replace incorrect titles or delete any image altogether

FINAL STEP

Track Progress & Schedule Listings

A progress stepper, allows sellers to track their upload process, go back and review entries, and save listings as drafts to complete later.

They can also schedule their listing, with a design inspired by Gmail's scheduling feature.

Automating multiple product uploads ensured seamless accuracy, eliminating mismatches by 30%, and creating a smooth seller experience.

"Less Manual More Auto" - Anonymous User

CHALLENGES

Due to limited engineering resources, instead of building the full in-product system, I utilized Pendo, Braze and Sigma to solve this at low cost.

Complex Design Brief

The brief was to design a simple e-commerce site, but its complexity became clear later.

Fluctuating Deadlines

The project timeline was initially set for 2 months but was cut to 4 weeks due to client demands.

Stakeholder Management

The client insisted on closely mirroring numerous references, making it challenging to align their preferences.