Segment is an Analytics Company that routes data to various tools where it can be used most effectively. We're talking about 10,000 Sources that are spreading information over Destinations.
Because I'm interested in dashboards and juggling with data metrics, I've decided to talk a bit with one of the engineer who works with this data.
My goal was to see if I could improve productivity by improving a KPI.
What is wrong with you?
The result of the interview was fruitful.
To define where's come from an alert and if it's from the Sources or Destination side, you need to check two different dashboards.
Where do I start?
Regarding the complexity of tasks, I worked on a Task Flow to define:
1. How many tasks do I have to do to find the source of the problem?
2. How to simplify the process to warn concerned team?
3. How to make the all process simple and more efficient.
Then I was able to define how to help and I've prepared a plan:
For the "Ideal Dashboard", I've started my journey with a Situation Awareness technique to get a comprehensive overview of different approaches to the measurement of situation.
I chose that tool because engineers are like pilots, or surgeons: a lot of pressure, a lot of metrics and complex tasks with a limited time.
It helps me defining how to improve the operations' Dashboard. It wasn't only about making "better". It was also about giving some relief to the team in charge of incidents in the network (Less stress, more satisfaction) and communicating with Sales and then clients with more information about ongoing incidents.
Because I'm not an engineer, I needed some information about how everything works today to define how it will need to work tomorrow.
I did a Task flow to follow every steps that my engineer has to go through to complete his tasks:
Now let's make a better world
Then I did a "ideal" Task flow. That allowed me to visualise what could be an ideal situation to fix problem:
I've started to iterate using the "Crazy 8" method". It started messy but eventually, I reached my goal (with a neater version) and I started to do some mockups :
- INNOVATION : Alert system to inform real time which Client or Destination is impacted
- IMPROVEMENT : Find a system to show Which clients are VIP
Graphs appear from the requests
INNOVATION : Provide a solution to send a report about the incident to other teams
I checked with the engineer if we were on the same page and with somes changes, I was able to propose to him to test the prototype.
LEARNINGS AND TAKEWAYS
When you want to improve a product, don't be afraid to iterate.
Every feature that doesn't work for my user wasn't a personal failure but an occasion to do better, to really please him.
I have to rethink some features, the flow of the tasks, to make it visually more impactive.
The result of this case study was positive. Metrics are challenging and I'm not afraid of challenges.
Creating a new feature from scratch could necessitate more interviews, to see if all users share the same pain points. Since it was a self-assignment, I know I could dig in depth.