Verifying user suggestions

Improving the admin’s workflow for crowdsourced data collection and verification.

Overview

Food Oasis is a web application that provides information about food pantries and meal programs for food seekers in Los Angeles County.

The site users (tipsters) can submit suggestion for the food pantry’s data correction from the web form. This info needs to be verified before being reflected on the website. The verification process is conducted by an admin who manages volunteers, and a volunteer who confirms the suggested data.

Currently, admin’s tasks during the verification process have a number of manual steps. This project is to improve the admin's workflow.

YEAR

2021, 2 months


ROLE

UX/UI

TEAM

Seiko (self), Bryan (Lead Designer), John (Lead Engineer)


TOOLS

Figma

What are we trying to solve?

Our task was to improve the admin’s workflow during the verification process. As we went through the details of the admin’s workflow, we found out it requires a lot of manual work for each task.

Understanding the workflow

To understand the workflow, I created the cross functional user flow and took screenshots of the interface we were currently using. This diagram shows each user’s task during the verification process. Later I added the problems and the things we need to clarify in this chart. This helped the team to discuss the problems in each stage.

cross functional flow chart

Info verification process

This is the high level overview of the info verification process.

  1. The tipster sends a suggestion through the web form. (This suggestion is saved to the server but needs to be verified before it can appear on the website.)
  2. Admin views the suggested correction
  3. Admin assigns this to a volunteer
  4. A volunteer will contact the food pantry to confirm the proposed correction by phone or email and leave a note for the admin.
  5. Once the correction is confirmed, an admin can approve it.
  6. When the admin approves the info, it will be automatically reflected on the website.

verification process

Admin's current workflow and the challenges

Now I’d like to take you each process and admin’s challenges.

Step 1. Tipster sends info

verification process step 1

When the food pantry's info is incorrect, the web user ( tipster ) submits the correct information to Food Oasis, because it is crowdsourced, we need a way to vet information that people suggest.

This is similar to Google map's correction. For example, when you go to the store, you find out the store hours are incorrect on Google map's data. You can submit correction data from a web form.

send correction on webform

Step 2. Admin reviews info

verification process step 2

This stage the admin checks the information from a tipster. The image left is the admin view, which is a table of suggestions from tipsters. When a tipster sends a suggestion, the data displays here. When you click a row, it shows the detail of the suggestion, image right.

table of suggestions

Then, the admin switches the screen to the existing food pantry's data to compare tipster data before assigning it to a volunteer because if the suggestion is spam or already correct, then the admin doesn't need to assign a volunteer to verify it.

Challenge 1: Admin has no way of comparing the suggested data with the current food pantry's data. Admin needs to switch screens to check both data.

food pantry's overview

Challenge 2: If there are a lot of suggestions for the same pantry, the admin has to look through all the suggestions individually.

The image shows the same food pantry but the suggestion is submitted twice. In this case, the admin checks each suggestion and detail twice.

table design verification

Step 3. Admin assigns info to a volunteer

step 3

Admin assigns the info verification to a volunteer from this Food pantry data overview.

table design verification

Admin also needs to send an email to the volunteer to share the suggestion from the tipster. Currently, the admin doesn't have a way to share the tipster's suggestion to a volunteer through the site.

Challenge 3: Admin doesn’t have a way to share the tipster's suggestion info to volunteer through the site so the admin sends an email manually.

send email to volunteer

Step 4. Volunteer confirms the info

step 3

After the admin assigned the info verification to a volunteer, the volunteer can see the food pantry's data like the image. This screen tells the food pantry's data like name, address and business hours.

The volunteer contacts the food pantry by phone or email and confirms the info.

After confirmation is done, the volunteer marks the checkboxes. As you can see the critical fields have marked orange color.

table design verification

Step 5. Admin reviews the info

step:5 admin review info

Diagram below shows the admin's workflow after the volunteer's confirmation.

  1. Check the food pantry's info that is confirmed by a volunteer and approve.
  2. Go back to the Suggestion Administration screen and click the approved food pantry.
  3. Change the status to Closed.

These are extra steps for the admin and if admin forget to change the status, it may happen that the already verified suggestion from the tipster is again sent to the verification.

Challenge 4: Admin needs to come back to the Suggestion Admin screen and change the suggestion's status to Closed.

Summary of admin challenges

  1. Switch screens to compare tipster suggestions and the current food pantry's data.
  2. Check the details multiple times if there are multiple suggestions for a food pantry.
  3. Send an email to a volunteer to share the tipster's suggestion.
  4. Close the suggestion after approving a volunteer's confirmation.

Major updates to solve challenges

Update #1: Adding Suggestion column

After identifying the challenges, we came up with this solution. We combined the suggestion from the tipster table and the food pantries' data overview. The new design, suggestion column is added.

Let's zoom into the new design. This column shows a number of suggestions from tipsters. The number is highlighted because these need to be verified. After verification is done and approved, this number is back to 0 and the highlight will be cleared. Now the admin can immediately see how many pending suggestions exist for each food pantry. This solves challenge #2: redundancy check because the admin can now check by food pantry instead of by suggestion.

Update #2: Adding suggestions to food pantry data

Here is another design update. We combine the suggestion detail page and the food pantry's data. The new food pantry's data includes tipster's suggestion data with buttons that communicate status.

Iterations

The final design, we changed the button to a radio button to show the suggestion status. We also added the note section. If something comes up during the confirmation process, the volunteer can write it down to communicate to the admin. The background color is updated from red to blue. Red is used for warning and we'd like to save the color for more critical situations.

Let's look at more detail with the image below.

1. Show the suggestion data on the top of the food pantry's data page.
This way both admins and volunteers can share the information from the tipster. Now the admin doesn't need to send email to a volunteer to share the information. In addition, admin doesn't switch screens to compare with the tipster suggestion and current food pantry's data.

2. Have the New, Confirmed, Incorrect status.
Volunteer updates this status during the confirmation process and admin checks the status when reviewing the verification. Now, the admin doesn't need to go back to the Suggestion admin page to change the status.

Demo of the prototype

Lessons and learned

1. Think holistically instead of focusing on each screen design.

At first, I tried to improve the suggestion table screen but this didn't work well because I didn't look at the whole admin user flow. Once I looked at the workflow and found the problems, we came up with better solutions like merging together the Suggestion table and the food pantry's overview screen.

2.Don't stick to the current model if there is a better solution.

We had a breakthrough when we brainstormed and decided not to use the current suggestion table. When the feature already exists, we try to improve existing features by adding more features. However, there might be a better way.