Conversion rate is an essential measure of the performance of any online-based business. Along with churn rate[1], it gives the most accurate indication of a business’ growth. Conversion rates are highly affected by the experience the users have to go through during signing up for a service or a product. Conversion rates are typically around 5%, but for unoptimized-for-conversion experiences, conversion rates drop to a value between 2.5% to 3.1% in online markets[2][3]. I had the chance to work on a conversion rate optimization in a live environment of a subscription-based app in the United States market with ~35,000 paid customers (paying $49 to $69 per month).
The mobile web version of the app had a conversion problem: users drop off during onboarding, and exit rate from the checkout page was high compared to the iOS app and the desktop version. On analyzing the onboarding flow, I noticed two things:
- Exit links (top navigation bar) appear in the checkout page, which comes after the user adds his email (becomes registered unpaid user)
- The checkout page itself is long as the user has to enter all information in only one page
On seeing this, I went on a quest to analyze and check common patterns for both checkout page and navigation bars during checkout. As expected, exit links are never shown during the checkout or signing up process. This eliminates any possibility of the user to get distracted and exit the process, as well as drawing focus and support to the primary task of checking out (if you are interested, this is part of the persuasive system design model proposed by Oinas-Kukkonen et al.[4]) .
Common patterns for checkout eliminates exit links
Another finding was that, most services divide the checkout step into two or three steps; first is usually personal information, second is payment information, and finally a confirmation page. I decided to go ahead and design a whole new checkout page, with no exit links, and with two steps checkout. First step was personal/shipping information, and second page was for payment info (with a small area of highlights of personal info and edit CTA). We ran – along with the growth team – two separate A/B tests for the new flow vs. the old one. The first test removed the exit links, and the second test divided the checkout page to two. Each test ran for 20 days. Optimizely platform was used to deploy tests and monitor results.
First test (removing exit links in header) had 30,000 unique visitors during the test period. The control design resulted in 1,120 (7.16%) unique conversions, while the variation design resulted in 1,210 (7.83%) unique conversions. Using the variant design improved conversions by 9.3%. Second test (splitting checkout page) had 16,484 unique visitors during the test period. The control design resulted in 723 (8.85%) unique conversions, while the variation design resulted in 805 (9.68%) unique conversions. Using the variant design improved conversions by 9.1%.
It was very interesting and impressive to see how small changes in the experience can actually lead to a significant improvement in growth and overall business goals.
References
[1] Kim, S. Y., Jung, T. S., Suh, E. H., & Hwang, H. S. (2006). Customer segmentation and strategy development based on customer lifetime value: A case study. Expert Systems with Applications, 31, 101–107.
[2] Moe, W. W., & Fader, P. S. (2003). Capturing Evolving Visit Behavior in Clickstream Data. Journal of Interactive Marketing, 18(1), 5-19. http://dx.doi.org/10.1002/dir.10074
[3] King, A. B. (2008). Website Optimization: Speed, Search Engine & Conversion Rate Secrets. O’Reilly. Sebastopol, CA
[4] Oinas-Kukkonen, H. and Harjumaa, M. 2009. Persuasive Systems Design: Key Issues, Process Model, and System Features. The Communications of the Association for Information Systems.
Engineering The Design 