Problem
What made this hard
The previous flow exposed technical dependencies too early and made recovery difficult. IT administrators had to interpret raw infrastructure commands before they could make safe cluster setup decisions.
All case studies03 / 11
Enterprise storage · Systems UX · 2023
Swarm Installation
Turn infrastructure setup into a guided sequence of decisions.
RoleSenior UX Engineer IIStatusShippedContextDataCore · Swarm object storageEvidenceNode discovery · System checks · Governance choices · Recovery guidance · Usability timing benchmark
Outcome
What changed
A guided installation model that automated node discovery, capacity checks, and network validation, while making governance choices (retention, replication, immutability, quota) explicit. Measured a ~60% reduction in setup time and a 25% improvement in first-run task completion.
The thesis
A redesigned installation experience for Swarm object storage that discovers system state, guides consequential choices, and keeps progress understandable for enterprise IT operations.
Case-study proof
The decision record
Enough context to judge the choice, inspect the work, and separate evidence from ambition.
Why this, not that
I moved automatic discovery and system checks ahead of governance choices instead of exposing every infrastructure dependency in one technical wizard. Automation handled what the system could know; consequential decisions stayed explicit.
Why this case is specific
The flow is specific to object-storage installation: node discovery, capacity and network checks, then retention, replication, immutability, quota, and recovery guidance.
What you can inspect
Shipped workflow · illustrative reconstructionThe page reconstructs the shipped sequence and its decision points in a static interface model. It communicates the workflow, backed by measured usability benchmarking.
Decision to outcome
The shipped sequence gave teams a visible next step after each check and kept high-consequence configuration readable, delivering a measured 25% improvement in setup task completion.
Interface evidence
Quantified Evidence · Methodological Transparency
Defensible Metrics & Measurement Methodology
Clearly distinguishing measured product outcomes from prototype benchmarks and design targets. Every metric is anchored in verifiable methodology.
| Metric | Evidence Type | Before State | After / Target | Measurement Method & Sample |
|---|---|---|---|---|
| Workflow Page Hops | Measured Shipped Outcome | 4 separate pages & modals | 1 persistent side panel | Workflow click-audit & production telemetry · (TapClicks AI Agents (1,200+ agency users)) |
| Setup Time Reduction | Measured Shipped Outcome | 45 mins manual node provisioning | 18 mins guided installation (~60% faster) | Benchmark usability timing tests · (DataCore Swarm Installation (Enterprise IT Admins)) |
| Task Completion Rate | Measured Shipped Outcome | 68% completion on first run | 93% completion (+25% gain) | Pre/post release task testing · (DataCore enterprise storage administrators) |
| Data Inconsistencies | Measured Shipped Outcome | Frequent missing lineage links | 40% measured reduction in link breaks | Error log telemetry & client audit tickets · (Enquero PowerMe lineage catalog) |
| Configuration Save Gates | Validated Design Target | 3 discrete save checkpoints | 1 continuous save with memory protection | Stakeholder & prototype validation · (TapClicks Unified Data Configuration Concept) |
| Time-to-First-Slide | Prototype Benchmark | Blank screen (96% drop-off) | < 2 minutes with pre-populated demo data | Interactive prototype session recording · (SmartStory Trial Onboarding reference prototype) |
01 / Understand
Translate system state into a next step
I organised the flow around four questions: what was found, what needs a decision, what could block progress, and what is safe to do next.
02 / Sequence
Automate checks, expose decisions
Discovery, capacity, and network checks run before users reach governance choices.
- Automatic node discovery
- Visible check status
- Plain guidance beside failures
- Governance choices kept explicit
03 / Safety
Make consequences readable
Retention, replication, immutability, and quota decisions include direct explanations and confirmation. People never have to infer whether a consequential change succeeded.
04 / Evidence
Defensible setup reduction and higher completion
Benchmarking tests confirmed an estimated 60% reduction in setup duration (from 45 mins to 18 mins) and a 25% improvement in task completion without admin error.
Supporting evidence
Translate system state into a next step
I organised the flow around four questions: what was found, what needs a decision, what could block progress, and what is safe to do next.
Automate checks, expose decisions
Discovery, capacity, and network checks run before users reach governance choices.
Bring me the difficult part.
Complex workflows, AI trust, and enterprise systems.
Discuss a product challenge