Services

What Happens After Launch? How to Structure a Maintenance Retainer/SLA So You're Not 'Paying Only When It Breaks and Running Naked the Rest of the Time'

2026.07.16 · 47 views
What Happens After Launch? How to Structure a Maintenance Retainer/SLA So You're Not 'Paying Only When It Breaks and Running Naked the Rest of the Time'

Most SMBs treat maintenance as an optional monthly fee — until the site goes down, payments break, or they get hacked. This piece gives a framework to turn maintenance from vague goodwill into a verifiable service level

Share:

The client's real question: "Didn't we finish the site? Why keep paying monthly?"

An e-commerce owner ran a year without incident, so he cancelled the NT$12,000/month maintenance. In month eight, an un-updated package was exploited, the site was injected with malicious redirect code, Google flagged it as dangerous, revenue went to zero for three days — and between remediation and reputation, losses far exceeded the four months of maintenance he saved. His question is one every SMB owner asks: after launch, what does that monthly fee actually buy? This piece turns "maintenance" from vague goodwill into a contract you can verify.

Breaking industry myths

  • Myth 1: "Once the site is built, leave it alone" — reality: frameworks, packages and servers ship security updates monthly; not updating is leaving the door unlocked. Laravel's support policy lists each version's security-update cutoff.
  • Myth 2: "Just call someone when it breaks" — reality: with no prevention or backup verification, incidents are worse and pricier, and you burn time finding someone unfamiliar with your system.
  • Myth 3: "Maintenance means someone on standby 24/7" — reality: maintenance is "lower the odds of failure + a committed response when it fails," not a full-timer. Defined via SLA, it's cheaper than you think.
  • Myth 4: "We have backups, so we're safe" — reality: a backup never restore-tested isn't a backup. The metric is "successful restore drill," not "backups running."

Core framework: the maintenance pyramid

Split needs into three layers, bottom-up, to see how high you should buy:

  • Layer 1 — Survival (essential): security updates, daily backups + quarterly restore drills, uptime monitoring (UptimeRobot), SSL renewal, outage recovery. This layer is insurance; don't skip it.
  • Layer 2 — Health (recommended): performance monitoring (Core Web Vitals), error tracking, a monthly tweak allowance, dependency inventory. Keeps the system from slowly rotting.
  • Layer 3 — Evolution (value-add): quarterly optimization, small feature projects, data reports, A/B tests. Lets the system grow with the business.

Three typical scenarios

  • Brochure site (no payments, low traffic): Layer 1 only. ~NT$3,000-6,000/month, focused on security and backups.
  • E-commerce/membership (payments, revenue-dependent): Layers 1+2, with 24/7 monitoring and payment-anomaly alerts. ~NT$12,000-25,000/month.
  • Operational platform (multi-system, growing): all three layers + monthly evolution hours, near tech-custody. NT$30,000+/month, still far below an in-house engineer.

Hidden costs (what you actually pay without maintenance)

  • Downtime loss: an e-commerce site's revenue loss per hour down often exceeds a full year of maintenance.
  • Security incidents: malware injection or a data breach — remediation and reputation costs routinely hit six figures.
  • Emergency premium: "call someone when it breaks" rush rates are typically 2-3x a retainer.
  • Knowledge loss: without continuous maintenance, a new team relearns the system from scratch — you pay the learning cost twice.
  • Tech-debt interest: skip updates long enough and a future big upgrade compounds to several times the cost.

KPI scorecard for a maintenance partner

  • ☐ A written SLA (three numbers: response, resolution, service hours)
  • ☐ Clear backup frequency and a "restore drill" mechanism
  • ☐ Proactive security updates with a monthly report
  • ☐ What monitoring/alerting covers (availability/performance/payments)
  • ☐ Defined emergency events and an escalation path
  • ☐ How monthly hours are counted and overage handled
  • ☐ Handover docs and account access held by you
  • ☐ On termination, how code and data are returned

ScriptWalker's options + when we're not a fit

We offer three modes: Survival pack (Layer 1), Standard health pack (Layers 1+2 with SLA), and Tech custody (all three + evolution hours, an external tech team). Honestly, we'll tell you not to buy our maintenance when: it's a purely static brochure page (free monitoring is enough), revenue doesn't depend on the system and long downtime is acceptable, or you have in-house engineering and just need hands (advisory fits better than full custody). Maintenance value comes from how much the system matters to you — buying the priciest plan for an unimportant system is waste.

Onboarding playbook (get the first month of takeover right)

  • Week 1: system inventory — dependency versions, server config, payment integrations, exposure map; set up backups and monitoring.
  • Weeks 2-4: apply pending security updates, run the first restore drill, set alert thresholds, deliver the first health report.
  • Day 90: review incidents and hours, adjust SLA tier, decide whether to enter Layer 3 evolution.

FAQ

After launch, what does the maintenance fee buy?

Three things: security (framework/package updates, patching), availability (monitoring, backups, recovery), and minor evolution (tweaks, reports, optimization). Not a full-timer on standby, but lower odds of failure and a committed responder when it happens. No maintenance isn't saving money; it's deferring risk into one big blow-up.

How much does a maintenance contract cost?

A common rule is 15-20% of the build fee per year, amortized monthly. A NT$500k site is often NT$8,000-15,000/month, covering monitoring, backups, security updates and agreed hours. If the system has payments, members or revenue depending on it, maintenance is insurance, not an option.

What must an SLA spell out?

Three numbers: response time, resolution time, service hours (business or 24/7); plus what's covered/not covered and the emergency definition and escalation path. As numbers, maintenance becomes a verifiable promise.

Can I self-maintain and only call an outsourcer when it breaks?

Yes, but know the cost: no one doing preventive updates and backup verification means accumulating risk, paying premium rates during incidents, and no one familiar with your system. Fine for minimal, non-revenue systems; once the system carries revenue, preventive maintenance is almost always cheaper.

Call to action

Want to know which maintenance layer your system needs? We offer a free 30-minute assessment to inventory your exposure surface, backups and SLA needs, translated into a clear plan and monthly fee. ScriptWalker provides maintenance from basic survival to full tech custody.

Share: