Research always leads design.
First of all, I collected the insights from research team. Based on their report, before SAP Edge Cloud was introduced, to let those services run on Edge Nodes(Server), users had to deploy services at each Edge Node with Edge On Premise Edition. Therefore, customer want to have life cycle management on the services efficiently.
Then I discussed with Head of Product Subha to understand the business strategy, it would help me to understand the position of this new product. Based on her, we will mainly target big deal customers that have to manage a lot of Edge Nodes in critical business process.
User Journey is crucial to understand how users use, think, feel and interact with a product. It will help me to understand users, verify the pain point and find design opportunities. Even though Research Team had provided the insight but they didn’t provide a complete user journey to demonstrate the user need. So first I discussed with Product Owner to complete the whole use journey.
As there were too many use cases, I chose a typical one to start analysis. The customer was a Greek Shipping Company. When vessels were offshore, there was no connection to Cloud. Only when vessels were inshore, the target user could deploy services on each Edge Node, which is not efficient.
By analyzing the real use case, I found some important facts which may effect the design scopes.
1. Target user actually tagged Edge Nodes with Device Type, for examples, Streaming Service will run on Nodes with Pump Type.
2. As the inshore time is limited, target user is likely to make mistakes when executing her work.
3. Sometimes, target user may update the service on a few Edge Nodes, and then apply the services to all corresponding Edge Nodes.
4. Target user works independently
Also, I defined the Design Questions and Success Measurements
Mass Deployment: How to make those Services run on the right Edge Nodes correctly and effortlessly?
Number of new customers > 10 after Q4
Customer satisfaction rate > 8
In some sketching time, there were some crazy ideas like smart automatic deployment. But after analyzing the possibility and value, the main ideas were scoped to two main one. One is Clustering. Another is Tagging.
For Clustering, user can cluster Edge Nodes first, the cluster share attributes like Device Type, which relates to Services. Then user deploys services on the right cluster.
For Tagging, user can assign the Edge Nodes with multiple tags. In bulk deployment, services can be deployed to Edge Nodes with certain tag. Tag is like filter, can filter out the right Edge Nodes for deployment.
So, which direction is better?
To figure out which solution is more intuitive to users, I should ask users. Quick Interview is a good way to understand how user perceive a thing. In the activity, the tester is asked to perform the tasks with the low fi mockup. By observing the way testers acts, I can think in their shoes. Below are key findings.
Comparatively, testers would like the tagging way a little bit better, not very much, it is more flexible, but testers also mention the cluster one is more organized.
As user doesn't show strong preference, I have to make some decision based on the big picture. In the Cluster approach, the clusters may share the same Node, it is not exactly like folder concept in laptop. It may cause confusion in long term, therefore, we moved toward Tagging approach.
Then I refined the user flow and scheduled meetings with Dev and Product Owner to illustrate my design. Below link explains how I designed the Bulk Deployment feature on Edge Node List.
Bulk Action Design