Data, Information and Actions

Updated: Sep 10, 2021

Around March 2021, I was asked to help redesign another internal-facing product that surfaces cloud optimization and cost-saving opportunities for applications. Similar to the Cloud safety management product and many internal tools made for tech leaderships and engineers, the initiative of the product is to surface data to UI as information. The good news is that the logic and correlations of the data behind this product are less complex than the Cloud safety management product that I support.


After creating prototypes to address immediate tech requirements, conducting usability testing sessions, and talking to more users to understand the current process and pain points of Cloud optimization and cost, I obtained valuable insights from the targeted users.


People's time is valuable. Data analysis taking time. Unfortunately, not everyone has the time or is eager to spend a lot of time looking into the data. I constantly hear one voice from different people when gathering feedback on internal-facing tools, "Just tell me what I need to do."


The objective of this product is to provide recommendations. However, users have to dig into a table to find out what instances are overutilized or underutilized, then finding more data to decide whether to take action and what to do.


Thinking of user's feedback and the current process, I revisited DIKW (data, information, knowledge, wisdom) Pyramid.


Fig. 1 DIKW Pyramid

Do we have the tech capability to tell people what they should do first? The team told me yes.


In the redesign, I shrank the factual information of the current state to the right side of the page title, so I can promote the low-hanging fruit as three cards above the fold and rank them based on efforts level.


When hover over each card, users will have a quick link to filter the the table below. Users can choose to look into the instances on the table, tailor the data on table, or simply export the table.


For example, if the user is interested in deleting unused instances, she could choose to view what instances are unused by clicking the green button. This action will filter the data table below and show her the unused instances along with relevant information. Or, she can directly export the data and go to AWS to remove the instances.

I worked with the PM to test the prototype with three users from different user groups. All users liked the feature. One of the users also brought another purpose of these cards. He mentioned the information on the cards is helpful for leaderships to quickly discover what can be done and the impacts associated with each actionable area.


I'm planning on the metrics to gather qualitative data of the product.

20 views0 comments

Recent Posts

See All