Quickbooks.com homepage experiments

nov 2014

As part of our mission to improve the end-to-end first use experience of QuickBooks, we wanted to look at the very top of the funnel. Most new user traffic was coming directly to QB.com, but bouncing right away. Our team did a 4 week design intensive to generate and test hypotheses on the live site to try and improve the number of customers considering our solution.

The problem

The landing page of the QuickBooks marketing site had a bounce rate of 96% - most prospects were landing on the homepage and leaving immediately.


Our hypotheses

There were many potential explanations for this behavior. For example, there's quite a bit of brand confusion between the consumer financial management software Quicken and the small business financial management software QuickBooks. But our team had a few key hypotheses to test that focused specifically on our target audience.

Hypothesis 1: Prospects will self-identify in hopes of getting a right-for-me QuickBooks recommendation.

Hypothesis 2: Prospects are simply trying to "get in fast" and will stay engaged if we get them to the product quickly.

Success metrics

Since this test would be the landing page only, our primary success metrics were bounce rate, engagement rate, and click-through rate. We kept an eye on, but did not measure success, on conversion to sign up.


This problem took a village to solve. Designers from across the organization partnered closely with our lead web marketing team to distill insights, brainstorm concepts, generate hypothesis and crafts hi-fidelity mock ups to test. We also reviewed each design iteration with the leadership team for feedback.


After a two-week design intensive and a quick round of qualitative testing, we landed on four concepts to test around hypotheses. Concept A tested self-identification by allowing prospects to tell us a quick story about their business before getting a recommendation. Concept B tested self-identification by asking prospects to choose between one of 5 business types. Concept C tested self-identification by allowing prospects to choose a specific task that they wanted QuickBooks to help them with today. And finally, Concept D tested Hypothesis 2 by allowing prospects to get into the product as quickly as possible by simply creating an account (skipping the process of SKU selection altogether).