Full-time · First & solo product designer · 2016 H2 – 2019 H1
ThousandEyes’ Multi-Product UX Journey
As the first and solo product designer, I built the design foundations and experience strategy that carried a single-product startup into a four-product platform, through to Cisco’s ~$1B acquisition.
ROLE
First & solo product designer, later design team lead
SCOPE
Platform UX architecture across 4 product lines
PLATFORM
Enterprise SaaS, desktop web
TYPE
Full-time, in-house · 3 years
01 · CONTEXT
One flagship product became four, each with its own user.
ThousandEyes is a network intelligence platform, the “Google Maps for the Internet.” It shows where data is delayed, dropped or throttled across global networks that users can’t see.
When I joined as the sole designer, the company relied on one offering: Cloud & Enterprise Agents. During my tenure it scaled into a multi-product suite. Each new product brought its own audience, deployment constraints, data visualization needs and commercial model.
FLAGSHIP
Cloud & Enterprise Agents
Active probes simulating traffic from data centers to critical servers.
NEW
Endpoint Agents
Installed on employee laptops to monitor network quality for remote and branch offices.
NEW
Device Layer
Automated discovery and visualization of local-network infrastructure: switches, routers.
NEW
Internet Insights
Aggregated telemetry from global agents to visualize macro internet outages in real time.
Before: the original beta didn’t hold up at large data scale.
THE STARTING POINT
Each new product broke something the platform took for granted.
Data at a new scale. Endpoint Agents monitored tens of thousands of laptops. Plotting millions of fragmented data points created an unreadable “spaghetti monster.”
Deployment varied wildly. Zero install, MDM mass installs, LAN scanning. Cold configuration pages on Day 0 led to heavy trial drop-off.
Mismatched pricing. Usage-based billing next to license subscriptions. Enterprise buyers couldn’t see total cost of ownership.
02 · CORE TRADEOFF
Products can go deep in different directions and still share one architecture.
The question was double-edged: how do we honor the deep technical needs of each specialized persona while keeping the platform cohesive and intuitive? My answer was a unified Platform UX Architecture that bridged the products.
Each persona
Cohesive diversity
Every product keeps the visualization, setup and pricing model its users need. A help desk tech and a network engineer don’t get the same screen.
All product lines
Seamless consistency
Shared patterns for how data is revealed, how setup is guided and how cost is shown, so moving between products feels like one tool.
03 · DESIGN GOALS
Three goals, one for each problem the platform hit.
GOAL A · DATA VISUALIZATION
Show the alert first. Reveal the network only when asked.
Cloud Agents visualized linear path routing. Endpoint Agents generated millions of fragmented points, which plotted directly became unreadable. I set up a 3-step telemetry framework based on progressive disclosure to manage the cognitive load.
STEP 1
Summary overview
The first fold shows high-level system health and chronological trends.
STEP 2
Concentrated alerts
Anomalies are highlighted and clustered. Deep technical detail stays hidden to shield users from noise.
STEP 3
Diagnostic node grouping
On deep-dive, data streams cluster into logical node groups. Zoom and drag to pinpoint the failing router or terminal.
GOAL B · DAY 0 ONBOARDING
Show the value before asking for configuration.
Cloud Agents needed zero install. Endpoint Agents needed IT admins to push silent MDM installs to thousands of laptops. Device Layer needed complex LAN scanning. Cold configuration pages on Day 0 drove trial drop-off, so I designed a “value-first, guided setup” blueprint.
Endpoint Agent Day 0 Experience
Devices Day 0 Experience
FIRST
Illustrations & sample data
Each new product opens with scenario-specific illustrations and interactive sample data tailored to the role (network engineer vs. help desk), so users see their finished dashboard before writing any config.
THEN
Deployment wizard
One step-by-step wizard for every product. Command-line inputs or bulk software distribution, the UI wraps it in the same progress indicators, real-time feedback and error validation.
GOAL C · BILLING & QUOTING
Four pricing models, one pane of glass.
Going from one product to four brought usage-based billing alongside license subscriptions. I redesigned the Organizational Overview and created a dynamic Quoting Tool that turns tier calculations into an interface where buyers assign quotas, predict costs and manage budgets across all product lines.
Organizational Overview
Usage and quota for every product in one view.
Quoting Tool
Buyers model cost before they commit.
04 · REUSABLE FRAMEWORKS
Three interaction frameworks, reused across every product line.
01
Alert → Diagnosis progressive disclosure
Kept the first screen legible at any data scale, and gave every monitoring product the same path from overview to root cause, so users learned it once.
02
Value-first onboarding + unified deployment wizard
Replaced cold configuration pages with sample data and one guided setup pattern, addressing trial drop-off however complex the underlying install.
03
Unified billing & quoting
Brought usage-based and license-based pricing into one pane of glass, so enterprise buyers could assign quotas and predict total cost across the platform.
05 · OUTCOMES
The platform expanded and the experience held up.
GARTNER CUSTOMER EXPERIENCE
4.73/5
Across 48 company reviews, sustained through a complete product evolution.
COMPANY VALUATION
20×
Valuation grew rapidly and ended in acquisition by Cisco (~$1B).
DESIGN FUNCTION
UI → Product Design
From a solo designer to leading the product design team.
06 · REFLECTION
Differentiate first, then make it familiar.
Start from each product’s scenario and persona
A network engineer tracing a path, an IT admin rolling out agents to thousands of laptops and a help desk tech checking one user’s Wi-Fi need different things. I designed each product around its own data scale, deployment reality and job to be done before looking for overlap.
Then find the interactions they share
Once the differences were clear, the common ground showed up in behavior: every product moves from overview to alert to diagnosis, every setup is a guided sequence, and every plan resolves to quota and cost.
Consistent interactions lower the cost of the next product
Keeping those interaction models the same meant a customer who learned one product already knew how to read, set up and budget for the next one.