Virtual SaaS Transformation at Array

The Problem

Before COVID, Array had strong product-market fit with in-person pharma education events—we provided everything: the software, devices, network, on-site tech support. Our customers loved the white-glove experience for their high-stakes medical education meetings where hundreds of healthcare professionals gathered.

But they kept asking: "Can we use this for virtual events too?"

The challenge wasn't just adding video. Our entire business model was built around physical presence. Our revenue came from full-service event delivery. Our differentiator was the on-site experience. Replicating that virtually—with the same engagement, ease of use, and reliability—felt like rebuilding from scratch.

We'd started exploring virtual capabilities through a Customer Advisory Board, testing a rudimentary prototype: our existing engagement platform with a Zoom-like video pane bolted on. We were learning, but slowly.

Then COVID hit. Within days, every in-person pharma meeting was canceled.

Our customers called desperately. Their medical education programs couldn't stop—physicians needed to learn about new therapies, drug data, and treatment protocols. But live events were impossible. They needed virtual solutions immediately, not in six months.

If we couldn't deliver, they'd find alternatives. And we'd lose relationships we'd spent years building.

Discovery Under Pressure

We had weeks, not months, to figure this out. I immediately started talking to customers to understand: What do you actually need from a virtual event platform?

Early Learning from Prototype Testing:

Before COVID, we'd used our Zoom-like prototype internally and with friendly CAB customers. That early testing surfaced something we hadn't anticipated: virtual events had more stakeholder complexity than in-person.

In-person, we had one facilitator running everything on-site. Virtual was different. There were multiple roles, each with different needs:

• •Live presenters giving talks
• •Moderators managing Q&A and flow
• •Next-up presenters waiting backstage
• •Rehearsal and tech check needs before going live
• •Slide review workflows for compliance-sensitive content

Our in-person model assumed one person orchestrating everything. Virtual needed true multi-role orchestration.

The rapid COVID discovery validated and expanded this insight:

Customers confirmed multi-role support was critical, but added urgent new requirements:

• •Reliability at scale: No tolerance for video crashes or tech failures
• •Engagement tools: Polls, Q&A, breakout discussions—not just passive watching
• •Compliance-ready: Recording, archiving, attendance tracking for medical education credits
• •White-glove support: Even virtual, they wanted our support team available, not DIY

The key insight: We weren't just adding video to our platform. We were reimagining event delivery for a virtual-first world while maintaining the reliability and support that made customers trust us.

Building the Platform: Rapid Launch in <90 Days

I partnered closely with sales to understand: Which customer needs are critical vs. nice-to-have?

We couldn't build everything. We had to sequence ruthlessly.

Phase 1: Core Capabilities (Weeks 1-4)

Priority features for MVP:

• •Multi-role support: Presenter, moderator, backstage workflows clearly separated
• •Rehearsal and tech check modes: Let presenters practice before going live
• •Basic engagement tools: Polls, Q&A, chat
• •Reliable video hosting: Partnered with proven video infrastructure (didn't build from scratch)

Phase 2: Launch with Friendly Customers

• •Selected 3-4 customers willing to be early adopters
• •Ran their first virtual events with our team standing by
• •Collected feedback immediately after each event
• •Iterated on pain points within days, not weeks

Phase 3: Broader Rollout

• •Expanded to more customers based on early learnings
• •Added features informed by live event usage:
  • •Breakout rooms for small group discussions
  • •Improved backstage prep flows
  • •Better moderator controls for managing Q&A
• •Continued weekly iteration based on event data

Our customers were incredibly gracious given the circumstances. They knew we were building in real-time. The trust we'd built through years of reliable in-person events gave us room to learn quickly.

Platform Thinking While Moving Fast

Even under extreme time pressure, we made architectural choices that would matter long-term.

Instead of hardcoding features for specific event types, we built a modular event platform with reusable components:

• •Role and permission system: Define who can do what in an event (presenter, moderator, attendee, admin)
• •Video stream management: Plug in different video sources without rebuilding workflows
• •Engagement widget library: Polls, Q&A, surveys could be configured per event

Why modular architecture mattered:

Short-term: We could launch with core features, then add capabilities incrementally without breaking existing events.

Long-term: New event formats (webinars, hybrid events, multi-day conferences) could be configured, not custom-built. We could scale from 10 to 100+ simultaneous events without rewrites.

This wasn't over-engineering—it was just-enough architecture to move fast sustainably. We weren't guessing what we'd need in the future. We were building foundations that let us keep learning and adapting.

Launch & Iteration: Weekly Learning Cycles

The real insights came from running live events and watching what broke.

Feedback (from first 3 events):

• •Backstage prep was confusing—presenters didn't know when they were "live" vs. practicing
• •Moderators needed better Q&A controls—questions were coming in faster than they could manage
• •Breakout rooms were requested for small group discussions

Adjustments:

• •Added clear visual indicators for "rehearsal mode" vs. "live"
• •Improved Q&A moderation tools (mark as answered, pin important questions)
• •Started building breakout room capability

Feedback:

• •Customers wanted to see engagement data during events, not just after
• •Some presenters struggled with slide transitions (tech check wasn't catching this)
• •Recording storage and retrieval was too manual

Adjustments:

• •Real-time engagement dashboard for event organizers
• •Enhanced tech check to specifically test slide transitions
• •Automated recording archival with easier retrieval

This became our rhythm: Run events → collect feedback → iterate → run more events. Weekly releases meant customers saw improvements constantly.

Results

Business Impact

• •50% revenue growth in 6 months post-launch (from virtual event services)
• •Retained Fortune 500 pharma customers through the transition (zero churn during crisis)
• •New customer acquisition from pharma companies who hadn't used Array before but needed virtual solutions

Platform Success

• •Ran 100+ virtual events in first 6 months with 99%+ uptime
• •Scaled from 10 to 50+ simultaneous events without infrastructure rewrites
• •Modular platform enabled future products: Hybrid events, on-demand content, mobile experiences all built on same foundation

Customer Satisfaction

• •High NPS scores despite launching during crisis (customers appreciated rapid iteration and responsiveness)
• •Customer advisory board grew as more pharma companies wanted to influence virtual roadmap
• •Reference customers emerged who promoted Array's virtual capabilities to peers

Team & Organizational Impact

• •Built resilient product culture: Team learned to ship under pressure while maintaining quality
• •Established iteration discipline: Weekly feedback cycles became standard practice
• •Strengthened customer relationships: Crisis response deepened trust with enterprise accounts

Key Learnings

1. Discovery doesn't stop at launch.
The real insights came from running live events and watching what broke. Our CAB prototype testing gave us direction, but the rapid post-launch iteration taught us what actually mattered. Weekly feedback cycles became our most valuable discovery tool.

2. Constraints force clarity.
We couldn't build everything, so we had to identify what really mattered. Multi-role orchestration was the insight that unlocked the platform. If we'd had unlimited time, we might have gotten lost building features that didn't matter. Crisis forced ruthless prioritization.

3. Platform thinking enables speed.
The modular architecture we built in those first 90 days let us keep iterating for 2+ years without technical rewrites. Moving fast and building for scale aren't opposites—they're symbiotic. Just-enough architecture at the start pays dividends for years.

4. Grace and partnership matter.
Our customers gave us room to learn because we were solving their crisis too. That trust—built through years of reliable in-person events—let us experiment in ways a polished, slow launch never would have. Relationships create space for innovation.

5. "What can we do in two weeks?" is a powerful forcing function.
When I joined Dice later, I brought this question with me. It came from this experience—realizing that tight timeframes force creative solutions and prevent overthinking. The best way to learn is to ship something and get real feedback.

6. Side-by-side customer partnership accelerates learning.
Working directly with 3-4 friendly customers in the early weeks gave us feedback loops measured in days, not months. They became co-designers of the platform. When you're building something new, find customers willing to build with you.