Data Valuation

Client
Gulp Data
Team
Sticky Ventures LLC
My Role
Jr. Product Designer
Project Type
Client Side App
Project Timeline
4 Months
Release
2023

Background

the client

GulpData is a startup specializing in evaluating data assets and providing a lending service that enables businesses to borrow money by using their data as collateral.

The Needs

GulpData, as a rapidly growing startup, witnessed an exponential rise in both clients and tasks. One of their core services is data evaluation, and in response to the escalating demands of their clients, they aimed to create an application.

The requirement was for an app that operates automatically, eliminating the necessity for the GulpData team to manually advance the valuation process for each client.

Project Overview

Data Valuation is an application utilized by potential Gulp Data customers to upload their company datasets and receive ongoing updates on the data valuation process until its completion.

We had to speed up the process and release an initial version of the app that allowed GulpData's team to work with their first clients, as the client had an urgent need for it. After launching the initial version, we were able to gather valuable feedback from their clients, which enabled us to continuously improve the product until we reached its third and final version.

Understanding

Research

As an initial step, we conducted an internal investigation within the Gulp team to comprehend the actions their clients would have to undertake to finalize the data valuation.

The process essentially divided into two flows. In the first, the customer had to fill out an informative form and submit their datasets ("Uploader"). In the second, these datasets were evaluated by the GulpData team ("Data Evaluation").
The main purpose of this application was to automate these two flows as much as possible.

Map Project Flow

During the design phase, our client urgently requested prioritization of the initial segment of the project (Uploader), necessitating expedited processes. As a result, the first version was launched within a span of just three days.

Quick Version

In response to the client's urgent request for the immediate implementation of the automated flow (Uploader) and considering the simplicity and intuitiveness of the UI, we opted to forego the wireframing phase and proceeded directly to UI creation on Figma.

Map Of The Flow

This quick version essentially outlines the most crucial steps users need to follow during the uploader process, culminating in the manual evaluation of their data by the GulpData team.

01 verity your information

To enhance clarity, we have structured this process into two steps, accompanied by a guide on the right side for each step.
In step one, users are required to verify the information they provided during the pre-qualification process.

02 ready to dive in

In the second step of the process, users encounter two sections: Dataset Type and Upload Sample.

In the Dataset Type section, users are prompted to choose between two options: Sample or Full Dataset. Opting for "Sample" mandates providing details about their Full Dataset.

Within the Upload Sample section, users must upload the sample dataset. Moreover, if the schema isn't embedded in the sample file, users must furnish it separately.

This step guarantees users the ability to advance in analysis by utilizing either a representative sample or the complete dataset, while also ensuring the provision of all essential data and schema details for precise analysis.
(Click to enlarge it)

03 thanks message

Once the user has uploaded the data sample or the full dataset, they were prompted with the message "Thanks, swimmer" to express gratitude for submitting their data sample and to inform them that the Gulp Data team would commence the evaluation process as soon as possible.

Considerations & Feedback

We expedited the process because our client needed to quickly develop the feature. They already had clients waiting and could no longer manage uploading data manually for each one.  Expediting this process provided a quick win for them.

Following the launch of this workflow, valuable feedback was gathered from GulpData's clients. This feedback facilitated further product enhancements, leading to the subsequent release of a second version. Common feedback included:
  • Users disregarded the guide, resulting in incorrect field completions.
  • Users reported insufficient information in the guide, causing difficulty in understanding requested information.
  • Users prematurely closed the application without completing the upload process, despite successful file downloads.
After launching the 'quick version' of the Uploader, we had the time to start designing the foundations of what would later become the final application, with the first part involving dataset upload and the second part concerning the data evaluation process by the GulpData team.

First Version Full App

Layout

As previously mentioned, the app is divided into two main flows: dataset upload and evaluation. To optimize focus on content and maximize space, we chose a collapsible menu.  

Additionally, we enhanced the user experience by designing a layout that highlights the guide and content during each step, along with implementing a one-page layout allowing users to preview the next step before finalizing the current one.

Part I - Uploader

In response to client feedback from the initial release and in collaboration with the GulpData team, we have implemented several changes to the Uploader flow:
  • Revised the "Contact Info" step to allow users to quickly verify provided data accuracy, eliminating the need for form completion.
  • Added a step titled "Primary Keys In The Data" providing technical information for subsequent dataset assessment by clients.
  • Included a final confirmation step before upload completion, requiring users to confirm their understanding and compliance with all requirements.

Part II - Data Valuation Tracker

Once datasets are uploaded, users reach the critical phase of the app, involving evaluation of their data and potential loan opportunities. Instead of email updates, we proposed an in-app tracker for users to monitor evaluation status. This feature is also available in the mobile version.
For the layout, we preferred to maintain a similar approach to that of the uploader. Therefore, on the left, there is a descriptive section, which in this case represents a summary of the upload performed by the user. On the right, we have the actual tracker. At the first step, the tracker remains on "Valuation" until the GulpData team advances the evaluation to the next step.
In this phase, users are primarily in "view only" mode. They can either wait for the data evaluation process to finish, whether it's positive or negative, or start a new evaluation by clicking "New Evaluation." This feature allows users to retry evaluating the data if they uploaded the wrong datasets or don't want to wait for the final result, especially if it's expected to be negative.

Data Valuation Tracker (Mobile)

Since this part of the application is mostly in "view only" mode, we have designed a mobile version to allow users to conveniently track the data valuation from their smartphones.

users feedback

After the project release, we received feedback from users, mostly positive. However, some critical points emerged:
  • Too much text in the uploader guide.
  • Display of subsequent steps that may distract from the current one.
  • Implementation of the final step as an overlay, causing discomfort for users in reviewing previous content.

Latest App

In response to most recent feedback, we finalized the application with the latest modifications. Given minimal alterations to the flow for this final stage, I opted to provide a comparison between the previous and current versions, elucidating the rationale behind each specific change.

Menu

After user experience evaluations, we removed the duplicate menu for uploader and data evaluation, keeping only the latter. This decision ensures a seamless flow, as the application's primary purpose is data evaluation. Additionally, displaying both menus, where one excludes the other, was deemed illogical.

contanct info

We changed the style of the guide to be clearer in subsequent screens. We chose a light background with dark text to improve readability, especially when there is a lot to read. Additionally, we added icons to give substance to the content, and placed the call to action on the right to indicate progress.

Uploader

Below, I present the new concept of the uploader. The main changes are:
  • Only the current step displays the guide, preventing reader distraction caused by too much information shown at once.
  • Subsequent steps to the current one are presented as closed cards, which open only upon completion of the previous step, ensuring no incomplete steps at the end of the process.
  • Lastly, the "before completion" section is shown as the final step to complete before proceeding to the data evaluation tracker.

Data Evaluation Tracker

The tracker has remained almost unchanged, with only two updates:
  • We relocated the portion devoted to the phases of the Data Evaluation Process to the left, as it's the most crucial section of the page.
  • We changed the background from dark to light, like the other guides in the uploader.

Data Evaluation Tracker mobile

  • We relocated the portion devoted to the phases of the Data Evaluation Process to the left, as it's the most crucial section of the page.
  • We changed the background from dark to light, like the other guides in the uploader.

final consideration

Addressing user feedback to ensure smooth completion of the Uploader flow proved to be the most challenging aspect of this project. It was quite frustrating, as we aimed to develop a UI as simple as possible. Persistent feedback guided us to iterate towards the final version and resolve the issue.

Regarding the mobile version, while I initially intended to do more, technical user feedback indicated a preference for desktop usage. Consequently, we focused on creating a mobile version for viewing only. Although limited, it's better than nothing.

What I've learned from this project is that even if something seems extremely simple to me, it's not always the case for users. Therefore, I made an effort to understand where the problem was and how to, in a sense, guide the user along the right path. Sometimes, showing less is better than showing everything.
NEXT

Allyfe