Problem

Current Driver License scanning has high dropoff and only accepts limited forms of identity.

Understand the Problem

Why does the product have those problems?

Today, we have a basic US-style driver’s license scan using Google Vision. We did some user testing and data analysis, and found out that this approach has a number of problems as it does not:

• Create a seamless UX because of the quality of scanning tech and the user flow.
• Work for people without a US-style DL
• Enable all the government need.

Design Goal

What do we want to achieve through the ID scanning?

• Seamless UX. 
• Meet governments’ needs. We fulfill base requirements for all governments today and in the next 3 months
• Scalable solution. We create base ID scan modules that work across all Lime products.

Proposed Solution

How are we going to solve the problem?

We will build two in-app modules:

• Photo-based ID scanning for multiple IDs powered by Microblink
• Manual ID entry.

High level user flow

Based on the technology constraints and the design goal, we would like to invest design resources in the photo scanning screen and to see how we can craft a smooth UX.

1. Testing the scanning flow

The main purpose of this user testing is to find out if:

1. Users can pass the ID scanning smoothly.
2. Users can discover the entry of other ID types.

Key takeaways:

1. Dropdown list oftentimes works but needs to be more obvious.
2. Horizontal scroll menu and slideshow diminishes the discoverability.
3. Using a watermark to educate users works but need to be more specific.

Comments from engineers:

1. The technology can work seamlessly without specifying ID types.
2. Have a separate ID list would be a scalable solution on eng side.

Results:

1. Have a dedicated area on the screen for scanning.
2. The watermark is placed according to the best scanning distance and size. 
3. Highlight the part needs to be scanned by text and symbol.
4. A more obvious drop-down list entry.

2. Testing the manual entry flow

The main purpose of this user testing is to find out:

1. how users are going to react when the scan failed.
2. if users can discover the manual entry option.

Key takeaways:

1. Users will be actively looking for other options if the scan failed.
2. a single CTA button has the highest discoverability.

Comments engineers and the GR team:

1. Manual entry is not always a legit option because of the technology we will be using.

Results:

1. Place friction in the UX for discovering manual entry intentionally.
2. Still, make it an obvious option when the scan fails to cut down drop-offs.

Final Design

Result

What’s the outcome of the design?

Target metric outcome

We would now like to start quantifying the cost of our gov-tech features, i.e. the impact on trips (via user drop off) this feature will have. Thus, we also need to gather the data to show the  impact our feature will have on the market.

• ID scan works for 70% of users
• We manage to convert 5% through the funnel for Lime-S