dab | My Site 1

Project Overview

Project: DAB Data Extraction Tool – UX Redesign
Company: DAB (Data Analytics & Business)
Role: UX Researcher & Designer
Timeline: 3 months
Objective: Improve the usability, clarity, and error handling of DAB’s internal data extraction tool to support both technical and non-technical users through a more guided and user-friendly experience

The Challenge

The primary challenge was designing a tool powerful enough for technical administrators but simple enough for analysts or business users. Since the app lacked visual guidance, smart validation, and progressive disclosure, users often:

Misconfigured inputs
Failed to authenticate
Struggled with interpreting results

There was no support for gradual onboarding or user-friendly error recovery.

My Process

I applied a human-centered design process, following the Double Diamond model: Discover → Define → Develop → Design → Deliver

Final prototype :

Discover Phase

To understand the problem and user pain points, I used a combination of qualitative and quantitative research:

5 stakeholder interviews
Interviewed 2 internal developers, 1 product owner, and 2 support engineers to understand system limitations and common user complaints. the insights were that: “Most users fail during the connection setup—not because the system fails, but because they don’t know what to enter.”

6 usability observation sessions
Observed 3 expert users and 3 non-technical users while they performed a full extraction flow.
- Non-technical users took 4–7 minutes longer to complete the form.
- 100% of novice users asked for help with terms like "RPC configuration" and "Auth Mode."
Support ticket analysis (last 3 months)
- 47% of tickets were about connection/authentication failures.
- 31% were about extraction jobs that “ran forever” with no visual progress or confirmation.
- 22% were about how to interpret the extraction progress page.
Heuristic Evaluation
- Identified major UX violations in feedback, consistency, and error prevention.
- Form did not support inline validation or field-specific help.
- UI mixed advanced and basic options with no visual hierarchy.

Define Phase

User Journey Mapping

I mapped the end-to-end user journey to visualize touchpoints, emotions, and blockers across the workflow:

Start – Open the extraction tool (user feels neutral)
Step 1 – Enter connection details (confusion, fear of making a mistake)
Step 2 – Choose authentication type (overwhelmed by jargon)
Step 3 – Configure advanced options (uncertainty, trial and error)
Step 4 – Run the extraction job (unsure if it’s working)
Step 5 – Monitor progress (information overload, unclear indicators)
End – Receive result or error (relief or frustration depending on outcome)

Personas for distinct user groups :

Develop Phase

Here is the user flow of the redesign of the dab application :

1. Welcome

Screen: Welcome page with onboarding or intro message
Action: User clicks "Start" or "Begin Extraction"

2. Basic Info Input

Screens: Multiple basic data entry forms
Steps:
1. Enter target system or project info
2. Fill in metadata (e.g., name, ID, data type)
3. Continue to next section

3. Auto Configuration

Screens: Several configuration steps
Steps:
1. Set up connection parameters (host, protocol, port)
2. Choose authentication method (Username/Password, SSH, etc.)
3. Input credentials
4. Configure advanced options (RPC, filters, etc.)
5. Submit setup

4. RFC Check

Screens: RFC configuration and validation step
Purpose: Pre-validation before running the main job
Action: Confirm system compatibility or correct format

5. System Check

Screens: System check loading screens
Purpose: Backend tests connection, credentials, and form logic
Feedback:
- Loading spinner or “checking…” status shown

6. Status Feedback

Screens:
- Success screen (green)
- Failure screen (red)
Outcome:
- If success → user proceeds to data table
- If failure → retry or fix input errors

7. Data Overview

Screens: Several versions of data tables
Steps:
- Show list of extracted records
- Indicate status, duration, errors (if any)
- Provide filters and sorting