Order Collaboration

Impact & results

Problem

Most of the time, brands negotiate with their supplier on orders through emails or over the phone. As you can imagine, this gets messy when they start to handle multiple Purchase Orders, each with several to hundreds of line items requested. A lot of time is wasted sifting through decentralized information that could be spent scheduling work or addressing issues on the production floor. That’s why Nulogy wanted to build Order Collaboration.

+ more context

Solution

Order Collaboration removes the reliance on manual tracking and updating of Purchase Order information over disjointed documents and spreadsheets by allowing brands and suppliers to collaborate on one platform.

Surface the most urgent information first

The table is by default sorted by status, so planners can focus on the Purchase Orders that require the most immediate action first.

PO table
Make it easy to collaborate

Suppliers can easily make proposals on requests from the Brand.

Supplier Details Supplier Making a Proposal
Account for varied volume

Line Items can be accepted or cancelled individually, or in bulk. This scales whether you are managing 2 line items or 100 line items.

Brand Available Actions
Highlight incremental changes

The most recent changes are emphasized on both the brand and supplier side. The user can quickly identify what has been modified since they last sent the Purchase Order.

Proposal Highlights
Design for auditability

The history log keeps a record of all changes to the Purchase Order, making it easy to trace actions back to a user.

History Log

My role

I was the sole product designer a team comprising of a Product Manager, 4 Developers, and 1 Agile Tester.

Tools

Sketch Abstract Whimsical Invision

Approach

Build a shared understanding by collaborating early in the process

One of the first things we did was run an example mapping session with the team. This created early team-wide alignment and the multidisciplinary perspective allowed us to identify behaviours and use cases that if missed, would have led to re-work further into the design process.

Validate quickly, and early on

I spent two weeks with a design co-op sketching out our ideas and validating them with former Brand Planners that now worked at our company. We spent several hours each day together sorting out the workflow and bouncing ideas off each other. At the end of the two weeks, we showed them our sketches and made adjustments based on the feedback.

Phase 1: Define

Who are the key players and what are they trying to accomplish?

I interviewed 3 planners on the brand side, and 5 customer service representatives (CSRs) from the supplier side. They were each asked similar questions, with two versions of the script for either the brand or supplier perspective.

+ view questions

Using responses from the interviews, I generated personas by narrowing down the findings into three categories using the job-to-be-done, pains, and gains value proposition.

Brand Planner Persona Supplier CSR Persona
How can we help our users accomplish their goals?

After defining the brand and supplier personas, I created a workflow to capture how both supplier and brand would collaborate together.

Workflow Workflow

Phase 2: Ideate

Information Density

When we ran our initial validation with the paper prototypes, we experimented with Purchase Order details in a modal and in a slide out side panel.

Modal Sketch Side Panel Sketch

However, our requirements changed partway through the design process – users needed to specify multiple line items on a Purchase Order, instead of just one. I opted to move the information to a new page, to allow for more real estate to manage this information.

Full Page Sketch
Actionable Language

Capturing the status of a Purchase Order was key for users to understand which party needed to action on a Purchase Order. We initially approached the naming of statuses from by describing the actions taken on the Purchase Order, using terms like “modified by supplier” and “modified by brand”.

Statuses Sketch

However, after validation with the Brand Planners, we discovered the naming was not resonating with them. Planners have an endless to-do list and are therefore constantly under pressure to ensure deadlines are met. The passive language used did not convey urgency to act on the Purchase Orders. We refined the language to encourage action and rather tell the planners what needed to be done. We settled on the following statuses listed in order of highest priority first: Requires Response, Pending Response, Accepted, Canceled.

Phase 3: Prototype and Test

Task Walkthrough

To test the overall usability with suppliers, I designed a task walkthrough that simulated the supplier receiving a Purchase Order from the brand, and asked the supplier to complete a series of tasks to successfully respond to the Purchase Order. One of the most notable findings occurred in 80% of the tests. The suppliers would immediately go to reject the Purchase Order, instead of responding to the Purchase Order and making proposals.

+ view tasks

Reject Button User Test

From the testing, I found 3 major takeaways:

1 - Blanket “Reject” button was drawing too much attention

One of the most notable findings occurred in 80% of the tests. The suppliers would immediately go to reject the Purchase Order, instead of responding to the Purchase Order and making proposals. Upon further investigation, I discovered that suppliers will almost never reject a Purchase Order, instead working with the Brand to agree on terms. I removed the reject button in later tests and noticed that the success rate for the suppliers to find and make proposals jumped to 100%.

2 - Planners and CSRs are responding to line items on an individual basis

I had initially designed the feature to have a "Make proposals" button that would make the whole line item table editable, under the assumption that the planners are likely to quickly make changes to each line item if they are making proposals. During testing, it became obvious that the users were distracted by the visual noise that was created in the table when all of the fields suddenly became editable. I dug into this a bit more during testing and learned that users are often reading the content for each line item, and then deciding whether or not to make changes to the proposal or accept the line item. With the visual noise on the table in “edit” mode, they were not able to decipher the information as easily. I opted to have the option to make proposals on a single line item, available as a row-level action in the table, with the edits happening in a side panel. So when users needed to make changes, they could still read the information in the table without being obstructed by unnecessary input fields.

3 - Having the ability to accept the entire Purchase Order and accept individual line items should be handled consistently

In the design I tested, users were able to accept the entire Purchase Order through the “Accept Purchase Order” button. They also had the option once they clicked “Make Proposals” to accept line items individually or in bulk. With this design there was a lot of confusion around what accepting the entire Purchase Order meant versus using bulk accept on the line items. Introducing two different ways to achieve the same outcome was not making it easier or more efficient for users to accomplish what they needed, but rather complicating the interaction on the page. I modified the design to always expose the checkboxes and allowing the user to accept individually on the row-level, as well as in bulk by using the select all checkbox. By removing the extra “Accept Purchase Order” button, users are able to still accept the entire Purchase Order, and the confusion of two pathways to the same outcome was removed.

Here's what the revised designs looked like:

Improved Wireframe View Request on Improved Wireframe