Crafting a more opinionated Datadog
Turning Datadog from an informative tool into an opinionated companion through changes in
production and ways of working
Context
For years, Datadog was primarily an observability tool built around tables and a limited set of dashboards. Its guiding philosophy was to surface as much information as possible so teams could triage issues and decide which actions to prioritize.
That approach shifted with the introduction of Security Monitoring, where certain events demanded time-sensitive responses. Instead of merely presenting data for manual review, the product needed to enable faster detection and more immediate action to address security incidents.
My Role
As the manager of the Security team, I was responsible for:
Implement the new vision
The key problem we identified from our Research was that users had a hard time answering the question “where do I click next?”.
Through testing and experimenting, we decided to use that as our main philosophy. Users needed to have a clear next step for every screen in the platform.
Coordinating with the partner teams
Security Engineering is Datadog’s internal Security team. Besides making sure Datadog was compliant with all Security regulations, they also happened to be the perfect persona for our Security Monitoring solution.
I asked for their help to collaborate with us as reviewers and first responders helping is find the big rocks in our way before we engaged with finer UX Research
Provide design guidance
“Being opinionated” was a new concept in terms of design values for Datadog. My job here was to create the space for this new design language to emerge as a global pattern that could be used in the rest of the UI.
Ramp up on Research
The new components meant a considerable investment for all teams involved. To justify these changes we decided to keep a close eye on the actual UX improvements these changes had.
We paired with the UXR team to have a dedicated Researcher for Security, making sure every concept was tested before it went to production. The trade-off was speed of delivery, but in the case of Security we wanted to get it right, since part of the success of the Datadog strategy relied on getting access to that new market.
Approach
Introduce new Design System rituals
Exploring new elements for the design language meant also coordination across the different teams.
Our first stop was the DRUIDS team (Datadog’s Design System). We wanted them to feel included and be an active part of this discussion, with the hopes of expanding this new language to other products.
To pair with them we introduced a monthly check so they can review our work and raise any concerns from a DS perspective. We also assigned one designer of the Security team as a “DRUIDS ambassador”, and her role was to be the voice of the design system in our weekly syncs.
Speaking the language of Users
Security Engineering has specific frameworks and tools which are industry-standard. We relied on these concepts, such as the MITRE ATT&CK framework to allow users to find a familiar space they new and could rely on.
This not only helped our UX but also made our work easier than coming up with a new standard for classifying and triaging signals.
One of our main gains was the ability to rely and understand what signals and threats are more important or urgent, so we can guide the user to “their next click”.
Outcomes & Takeaways