Australian cloud for business software — built and run in Sydney since 2002

NetLeverage
NetLeverage

Handling change requests after an ISV customer is live

Technical article
Hosting

About this article: This article is reviewed against the cited sources. This is source research, not lab validation. Check the stated versions and your own requirements before making changes.

Direct answer: After go-live, route every customer ask through a classified intake: incident (unplanned service disruption or degradation—including security, privacy or data-integrity events), standard request (known config change with playbook), defect (product not meeting documented behaviour), or enhancement (new capability)—with explicit queues, owners and SLAs per class so config tweaks are not debated like outages and product gaps are not disguised as “urgent toggles.”

Why it matters

Live customers treat every frustration as “fix it now.” Without taxonomy, support burns out reclassifying tickets, engineering receives vague “make report faster” tasks, and account managers promise enhancements on phone calls that never enter a backlog.

This article is post-live change queue design for ISVs on hosted delivery. It complements—but does not repeat—a full support boundary RACI between application, OS and network. Here the focus is customer-facing request fulfilment once the tenant is in production.

How it works

Intake form fields (minimum)

  • Tenant id and environment (prod vs UAT—reject prod experiments without change window).
  • Request type selector with plain definitions and examples.
  • Business impact: blocked users vs cosmetic vs regulatory deadline.
  • Desired date and attachments (screenshots, sample export).

Decision tree

  1. Is production usability or data integrity compromised now? If yes → incident path (war room if severity 1), even when a defect is the probable cause—start response and open a defect record in parallel rather than waiting to classify.
  2. Is there a suspected security, privacy or confidentiality incident—for example credential leak, unauthorised disclosure or confidentiality exposure—even if the service is still up and data is not modified? If yes → incident path under your security/privacy procedure; do not wait for downtime or integrity failure before classifying.
  3. Does a standard playbook exist and is the service otherwise healthy? If yes → request fulfilment with approval if financial or security sensitive.
  4. Does behaviour differ from docs/release notes? If yes → defect with reproduction steps; link to any open incident until stable.
  5. Else → enhancement; product prioritisation, not overnight patch.

Queues and owners

Class Owner team Typical artefact
Incident Support + on-call engineering Incident record, RCA for majors
Standard request CS or tier-2 config team Checklist, rollback plan
Defect Engineering Version target, workaround
Enhancement Product Roadmap item, customer communication

Change windows

Configuration that touches integrations, tax tables or bulk data may need scheduled change even when not a code release—especially on shared stacks where a bad script affects neighbours. Document customer-visible notice lead time.

Commercial boundaries

Professional services hours, included config allowance and paid change packs should appear in the intake (“this may be billable”) to reduce invoice disputes.

What to choose and why

Automate status visibility: customers see “defect—target version 4.3” not “with engineering.” Hide internal churn.

For franchise groups, route requests via entity so store managers do not open network-wide changes.

Publish a one-page cheat sheet for customer admins: what is self-service in the app vs what needs a ticket.

Risks and limits

  • Misclassified defects closed as “training” erode trust—sample QA review of closures weekly.
  • Enhancements smuggled as config can fork tenant-specific logic you cannot upgrade later.
  • Shared multi-tenant config flags need guardrails so CS cannot set global killswitches by mistake.
  • ITIL terminology used educationally; not a certified process description.

Sources

  • ISO/IEC 20000-1:2018 — ISO; accessed 2026-11-16 — service management system requirements for providers (high level).

Related reading

Learning notes

  1. Four request classes reduce ping-pong more than faster agents.
  2. Intake quality determines routing—invest in the form.
  3. Standard requests deserve playbooks, not heroics.
  4. Enhancements need product backlog honesty.
  5. Post-live config can need change windows like code deploys.