Motor Quote Engine Enhancement (Split Test)

Project Date

Jan-Jun, 2021

Company

DirectAsia (Thailand)

Technology

Angular JS, Google Optimize, SessionCam (heatmap)

Who's in the party?
User Experience Team
  • UX Manager (my role)
      • Oversee project and manage resource/timeline
      • User research
      • Split test brief: Hypothesis and success criteria
      • Wireframe, UX Design
      • Data analytics for conclusion
  • UI Designer
  • Webmaster – Set GTM tracking, ensure data connection to CRM.
  • Web Developer
Performance Marketing Team
  • Monitor CPS/CPL based on improved funnel conversion.
Business Analyst/Apps Team
  • Ensure all api connection to core calculation is working.
chuttersnap-gts_Eh4g1lk-unsplash
Upon 6-month after we launched online motor quote, there was a few people who purchased on the website. I joined DirectAsia in late 2019 to tackle this challenge.

Buying motor insurance should be hassle-free but what happened during the journey?

User reseach

I began my research by trying to complete the online purchase myself, wrote down my notes, Then, I went to Customer Services team, who normally handled customer complaint from website to get some feedback on our current quote engine.

 

Number-wise, I looked into my sources of data like Google Analytics, SessionCam (Session Recording & heatmap tool) along with raw customer data from CRM.

 

From both qualitative and quantitative research, here’s some interesting findings:

  1. 71% users drop from the first page due to 13 required underwriting pre-screen questions
  2. Among the first page drop offs, 32% was caused by missing one of 13 required fields; for example Date of Birth.
  3. Some of the dropoffs called in and mentioned that they fail to complete all questions in time. Resulted in ‘session expiry’ popup.
  4. 46% drop-offs during step 3 (Your information#1) and 4 (Your information#2). CTR% for checkout page is 2%.
  5. For checkout, half of the total users decided to pay offline as it offers installment option, while online didn’t. Learned more about this from running an exit-intent survey for specific users who left checkout page (Case study coming soon).

Current

As-is interface is a straightforward form with 13 fields to complete prior getting an initial quote.
Funnel Conversion
(Mockup data)

To buy online, you need to complete a total of 64 questions. Some of Underwriting questions are in Page 3 (of 7). This means if you happen to fail it, you simply waste prior 5-10 mins.

Solution Discovery

Hypothesis
  • By taking all UW questions to Step 1, it will lift up drop off in Step 3. However, the number of questions on Step 1 will increase.
  • By shortening the journey and eliminating Step 4, it will lift the number of user who reach checkout
  • By redesigning new UI and reduce questions, it will lift up people who successfully buy online

Simplify Steps

After a long discussion among Business requirements (Comprise of Underwiting, Pricing), we were able to remove and reorder some questions. This resulted in 1 page shorter!

Removed Questions

  • Delivery Address-Country / ประเทศที่อยู่จัดส่ง
    Reason: Only one choice available: Thailand
  • Method of delivery / ช่องทางการส่งกรมธรรม์ (Might need to come back from OIC/PDPA regulation)
    Reason: Only one choice available: Mailing
  • Register as a company or not? / รถจดทะเบียนในนามบริษัทหรือไม่?
    Reason: There is no ability for us currently to administer these types of policies online. Given the volumes (20 per month) we would prefer to remove this question.
  • Chassis/Engine number
    Reason: It is compulsory at this stage. Users did not know this but as it was marked as optional they skipped over it. There was an expectation that they would need to give them later.

After

As mentioned earlier, session recording research showed 32% of first page drop-offs was caused by missing one of 13 required fields. Also, from Google Analytics, it shows that 70% is mobile user.

 

 

For these reasons, the below is the major tweaks for this version of design:

First, adjusted the input fields to fit in mobile view and made the most important CTA button sticky at all time so users can get validation and know what they missed.

Second, for users who switched back and forth between pages to recheck and tweak some of the coverage and leave, this new UX imitates air ticket/hotel purchase behaviour, where users can:

  1. Ensure their information in correct
  2. Adjust their information on the fly.
  3. For mobile, they can swipe up to see selected coverage plan all the time.
On desktop, users can see their profile on the left to ensure the correctness and adjust their coverage plan the right. Re-calculation is triggered when users adjust their profile.

Outcome

Funnel Conversion
  • The A/B test ran for 5 month, there was 13.6% growth in online quote.
  • Average time to complete Step 1 shortened from 5 mins to 2.5 mins.
  • Compared to 2020, we managed to generate 5x online policies in 2021.
  • Read more about this here (content in Thai).

Next Step

  • Upgrade tech stack for better and faster quoting performance
  • Add more financial options
  • Lift conversion during step 3-4
  • Apply predictive modelling to quote engine