
Decanter AI
Responsibility
Research, Strategy
Role
UX Researcher
Team
User Experience and Usability Evaluation Course Team & Mobagel
Period
2018.3 - 2018.6
OVERVIEW
OVERVIEW
We cooperate with Mobagel to help them evaluate their product, Decanter, which is a cloud-based data prediction service with strong processing capacity. First we made an alignment with their business strategy and product roadmap, defining the target users and started to conduct UX research. We are focusing on the group of students, which leads to the market of education. From qualitative and quantitive method, we help Mobagel clarify their next step.
INTRODUCTION
Machine learning and big data analysis have been the hottest issue for a while. While there are a lot of scientists and engineers building models and algorithms with code, the same needs from business, marketing, or entry level learner have never been disappear. Learning coding is a big challenge, and tools like business intelligence solutions are less powerful.
Decanter is a whole new solution for this kind of issue. Decanter AI is a fully automated machine learning (AutoML) engine that requires minimal skills and background to achieve powerful machine learning models, and It also provides models for business context and the API for companies to integrate the existing resources.
We as a third party spent times on identifying their customer segmentation, and decided focusing on the learner in data analysis field. After research, we generally found that these people found it hard to get used to coding, and they had hardship in figuring out the parameter when building models and choosing or designing algorithms. However, they also feels the power of the flexibility provided by code so that in some kind of circumstances, they want to see the records of how the system work and allowed them to make the system be adapted by themselves.
CHALLENGE
-
The taget population's scale is extremely large, and we need to help them converge and stay focus on specific groups.
-
It's hard to reach their clients and their users, even if they help.
Research Report Highlight
-
We help Mobagle decide to targeted the education market, and established two primary personas, Tracy and Hank.
-
Identified 2 level-four and 23 level-three usability issues through heuristic evaluation to provide design recommendations.
-
Conducted usability test with 5 participants, measuring SUS at 68, and produced a usability report.