LucidChart provided our team with a large amount of internal research. Most of which were the results of A/B Tests, surveys, and typical user flows.
Our analysis of the provided research led us to design for users inability to find the appropriate template for the content type they are working with.
M.S. Human Computer Interaction Design
Solution: Smart Template suggestion
In our solution, we proposed that the user flow gets modified slightly. Instead of picking a template before creating any part of a chart, we found a case for the option to select a template after the user had started from scratch as well.
We proposed rather than showing the user all of the templates, they are shown templates that are relevant to the first few pieces of information that they insert into the chart.
To create the relationships between inserted information and the relevant templates, we thought that LucidChart should record what information users are currently inputting into their templates. From there, LucidChart would be able to create a set of keywords and determine the strength of the information to template relationship.
As the user new user is creating their new chart, LucidChart could intelligently suggest relevant templates.
As part of the project brief, LucidChart requested that we provided a theoretics A/B test for our solution as well as what the success criteria would be.
In the data LucidChart provided us, our group picked out that
- 60.2% of customers start from a new blank document
- When using a template, 41.5% of people use the diagram as a starting point, keeping and editing the shapes they need
- Of the 58 samples of customer feedback we were shown, 25 of them were some version of "no/wrong template"
Any decrease in these statistics would indicate that our solutions is working as intended.