Building the Test Experience @ Bandwidth.
Role:
Product Manager
Duration:
11 weeks
Tools:
Claude, Jira, Confluence, etc.
CONTEXT
Bandwidth's enterprise test experience process was entirely manual, with no existing framework and no fixed scope. I owned it end to end: 60% of test prospects convert, representing 80 to 90% of deals over $1M ARR, and every account required hands-on Implementation work with no self-service path.
THE PROBLEM
Setup took hours to over a week, pushing self-directed customers toward competitors with instant setup. Accounts ran with no usage caps or expiration, generating P1 escalations and government audit fees on untracked traffic. Each build took 55+ manual tasks, up to 945 clicks, and one specialist spent 15 to 20% of their time on just 3 to 4 accounts. The model couldn't scale without adding headcount, introducing barriers to new partnerships and products.
WHAT I FOUND
The test experience was critical to customer buy-in, especially high-value enterprise customers. Standardization was critical to supporting a growing number of customers, but a single experience did not match customer needs. This led to the development of a three-tier system.
THE SOLUTION
I designed a three-tier model (Self-Serve, Bundles, White Glove) segmented by ARR, with automated provisioning under 30 minutes, enforced 30-day limits with real-time usage tracking, and support scaled to tier. Implementation involvement drops from 15 to 20% of specialist capacity to roughly 2 to 5%.
Two decisions mattered most. First, tiers over one flow: tier assignment required objective revenue thresholds and required override approval. Second, making limits visible throughout the trial instead of hidden until hit, since enterprise buyers read well-defined limits as compliance readiness.
I scoped Phase 0, test account visibility, as an independent launch since Phase 1 depends on infrastructure still in progress. Phases 1 to 3 (automation, self-serve tiers, AI-driven features) are roadmapped with advance triggers for the PM team continuing the work.
OUTCOMES
25+ interviews and analysis across 7+ teams, self-scoped.
Phase 0 shipped, in preparation for future development.
Strategy (PRD), full roadmap, and interactive prototype approved by product and senior stakeholders.
Implementation load projected to drop from 15 to 20% to roughly 2 to 5% of specialist capacity once fully rolled out - PM team continuing Phases 1 to 3 using my handoff plan.
REFLECTION
I led my own product including all associated responsibilities, including making decisions and being the subject matter expert in the room. I also learned to adapt to significant changes. I worked around this by prioritizing and scoping Phase 0 as an independent launch, instead of waiting for the full build, meant shipping something real in 11 weeks and leaving a credible roadmap behind instead of an unfinished one.