UI / UX / Web Design
Big Data Platform
Project Type:
SF Motors Company Project
New Product, Internal Tool
Role:
UI / UX Design, IA, Style Guide, Wireframe, Prototype
Duration:
11 months from October 2016
Overview
The project was to create a cloud-based platform from scratch to optimize business operations. The request came in when I first joined SF Motors in 2016 as a creative designer. The company, as a start-up, inherits sorts of innovative projects from its mother group - Sokon Automotive in China, and the goal of having such software was to better help the OEM make data-driven decisions.
Without much exposure on data intelligence, I was closely working together with project managers, data analysts, and engineers for 11 months, during which I was mainly working on style guides, IA, wireframe drafting, prototyping, demo, user experience, and user interface design.
What's the problem?
Like many other big companies, the OEM we are serving does not know how to use structured primary data in making intelligent decisions. More than 70% of its employees have access to data they should not, and 80% of analysts’ time is spent manually discovering and preparing data.
The assumption is, by introducing the highly automated big data solution to Sokon, the cloud-based internal tool with easy access to operational data and marketing intelligence will help the business improve resource efficiency and ultimately drive up the revenue.
Examine the assumption
The Assumption provided a guideline to conduct user interviews with decision-makers in Sokon Group. 16 employees in Managers, Directors, and VP positions were interviewed. The most challenging part was, without exposure to data science in their careers, the interviewees needed to be educated to understand how could data drive the BI in their work; which may lead them to give biased feedback based on the hypothetical guideline we provided. Thus, we interviewed 8 different teams in Sokon as listed below, to collect inputs from different expertises.
After 2 rounds of interviews, we get to learn more about the routine, KPI and needs of Sokon marketing and operation teams, as well as how can we access their database and what kind of data can we get.
Based on our conversation, we discovered the problems as:
-
Sokon Group generated a great amount of operation data, scattered across different departments and affiliated businesses, making it difficult to track and gather.
-
As so many departments are involved in the data creation, none of them had enough resources or experience to gather the data, let alone manage or analyze it.
-
Without a recognized data administrator, Sokon lacked a convergent data-driven strategy. In other words, no one knows where to start.
Dive into the problems
In consideration of the client's input, our feasibility and resource prioritization per communication with analysts and engineers in my team, the solutions as follows were pitched to Sokon for approval:
-
Our team will track and gather data from all internal teams, the interrelated market, customers, dealers, OEMs, service, logistics, and channels.
-
Our team will design and develop a could-based platform to implement, administrate, prep, analyze, and visualize all the data collected.
-
The platform will have applications that support BI projects in SOP (Sales and Operation Planning).
-
In consideration of the demands, feasibility, and resource, the development of the platform will start with OTD (Order to Delivery) planning.
Once the scope of the project is confirmed by Sokon, we conducted another round of user interviews to dive into the OTD operations. It allowed us to focus on defining the problems for OTD only; we discovered that the problems can be broken down into the following 5 stages:
1 Demand Acceptance
Dealers sent orders to Sokon
-
Accuracy of monthly production planning was 60%, the industry average was from 85% to 90%.
-
The accuracy of the regions that had higher sales was lower than the all-region average.
Duration (days):
Sokon Group / Industry Average
2 2
2 Production Planning
Sokon processed the order, then planned for the production
-
Production planning could only be done 2 days ahead, below the industry average of 7 days
-
Supplier and logistic were not preplanned for the incoming orders, it took too long to be ready for production.
Duration (days):
Sokon Group / Industry Average
8-12 4
3 Production
The vehicles were produced and stocked
-
After the vehicles were produced, it takes 1-2 days to stock them in the warehouse.
Duration (days):
Sokon Group / Industry Average
4.5 2.5-3
4 Inventory
The logistics were scheduled, the vehicles were prepared to ship
-
Most of the time, dealers won't execute the full amount of their order, which left extra units produced in stock.
-
The shipment distribution in transportation was not well planned, delayed transportation would put extra pressure on warehouses.
Duration (days):
Sokon Group / Industry Average
36 12-15
5 Delivery
Vehicles got shipped and delivered to dealers
-
When the demand was low, the destinations of a delivery truck would be located across a large area, which took longer to deliver.
Duration (days):
Sokon Group / Industry Average
8-11 4.5
With the insights collected, we would be able to develop the use cases, to define the scope and the key interactions of the platform. They helped our team to convert the pain points into flows and functions.
Information Architecture
Based on the use cases, with many rounds of internal discussion, we were able to build up the information architecture for the OTD planning, it helped us assign tasks to each of our team members.
Design Guidelines
While the Analysts were working on the data visualization, I created the style guides to ensure the charts they produced are visually consistent. By the time when we upload charts from different Analysts to the platform, they could be easily fit in.
The theme of the color scheme came from the color of the Sokon logo. The gradient from red to blue reminded me of the sunrise on the mountain top, so I collected reference photos to study the color variations during the sunrise. It allowed me to add orange and purple to the scheme, as my original color group to create sample charts for Analysts.
Due to limited resources, the platform is designed in the resolution of standard HD screen, which is 1920 by 1080 pixels. We have confirmed with our client that most of managers in their team are using displays with the same resolution. That allows me to sketch the basic layout of the platform to get a basic understanding of the size of charts.
I created sample charts in 3 different sizes for Analyst to apply in their data visualization tools, with the selected fonts and color combinations.
I explained the color scheme to Analysts, went through the samples I created, talked about the flexibility they could have applying the guidelines and collected their feedback about feasibility. The most critical feedback I had was, there were not enough colors in the original group to match their data, so I added another set of colors to the scheme, as the additional group.
Wireframe
Meanwhile, I was working on the wireframe based on the sitemap, constantly having 1-on-1 meetings with Analysts to construct the basic structure of the platform to accommodate all incoming charts. I kept updating the wireframe as the engineer started to build the platform.
UI Design
I worked on the UX and UI Design of the draft of the platform, spent most of hours working with Analysts and Engineer to understand their thoughts regarding data visualization and presentation. It's challenging to organize so many different charts and data into reasonable order and layers, but I enjoyed the process since I gained a lot of data acknowledge out of it.
Here are some screenshots of the platform version 1.0, for internal demo sessions.
Design Iterations
There were more than 3 rounds of internal demo for us to identify defects and deliver the beta version for UAT. UAT allows us to further polish the platform with refinements in user experience; followed by the soft launch. Please email me at river.floating@gmail.com to get more details about the latest platform update I worked on.
