The Support Plan That Does Not Match the Workload
A healthcare-adjacent startup runs a patient-scheduling platform with a 99.9% availability commitment to its customers. It currently has the lowest-tier support plan, which provides no 24/7 access to engineers and no guaranteed response times for production-down cases. Last quarter, a production outage began at 2 a.m. on a Saturday and was resolved only after a customer escalation on Monday. The CTO wants to upgrade support, but the CFO notes that the next tier up costs roughly 10x and asks for a business case. The company also has a compliance obligation to notify affected parties within 72 hours of a confirmed data incident.
1. Build the argument for or against upgrading support using the outage as evidence.
Read the worked answer
The outage is the evidence: a Saturday 2 a.m. production-down event with no 24/7 engineer access turned a technical incident into a multi-day customer-visible failure, which is a direct breach risk against the 99.9% commitment. Quantify it: hours of downtime times the contractual credit or churn risk, plus the engineering time spent working around the lack of support. Compare that to the 10x support cost. The judgement point is that support tier is an insurance and response-time decision, not a feature purchase - the question is whether the expected cost of a delayed incident exceeds the premium. For a platform with a published SLA and a 72-hour notification clock, the answer is usually yes for at least the tier that provides 24/7 production-down response, even if higher tiers with proactive review are not justified yet.
2. Which support tier features actually matter for this company, and which are noise?
Read the worked answer
Matter: 24/7 access for production-down severity, defined response-time targets by severity, and a named technical account contact who understands the architecture. Also matter: guidance during a security incident, because the 72-hour notification obligation means the company needs fast, authoritative answers about whether data was exposed. Noise for now: proactive architectural reviews and concierge onboarding are valuable but not tied to the outage risk, and the highest tier's business-level account management is unlikely to pay back at this size. The reasoning is to map each feature to a specific risk the company has actually experienced or is contractually exposed to, rather than buying the top tier because it feels safer.
3. Beyond buying support, what internal change would have shortened the outage most?
Read the worked answer
A documented severity model and on-call rotation with an escalation path. The outage was prolonged because no one was paged and the only path to help was a customer complaint. Even with the lowest support tier, an internal on-call engineer could have diagnosed and mitigated within the first hour. The support upgrade and the on-call process are complementary: on-call handles what the team can fix, support handles what requires provider-side investigation. The judgement point is that buying support without an internal escalation path still leaves the Saturday 2 a.m. gap, and building on-call without support leaves the team stuck when the fault is in the provider's service.