Log in

Blog Problem Solver

The real cost of network downtime (and why brochure numbers mislead)

7 min read

A practical way to estimate downtime cost beyond vendor SLA tables—lost revenue, labor, reputation, and the quiet tax of workarounds.

Four cost buckets

  1. Direct revenue — transactions that cannot complete
  2. Labor — overtime, truck rolls, war rooms
  3. Reputation — churn risk, SLA credits, sales friction on renewals
  4. Workaround tax — LTE hotspots, manual processes, “temporary” exceptions that become permanent

Most estimates stop at (1). The mature estimate includes (2)–(4), even if rough.

A back-of-envelope model

For a site or customer segment:

Cost ≈
  (revenue_at_risk_per_hour × outage_hours × impact_factor)
  + (loaded_labor_cost_per_hour × people × hours)
  + expected_credits_and_goodwill
  + workaround_spend

impact_factor is not always 1.0. A warehouse may lose 100% of scanning; a backup-heavy office may lose 30% of productivity but 0% of invoicing.

Why brochure downtime understates pain

  • Outages cluster at bad times
  • Partial outages (brownouts) never enter “down” counters
  • Human recovery takes longer than link recovery
  • Multi-site customers feel correlated failures harder than single-site math suggests

Using the number

Once you have a credible hourly range:

  • Prioritize monitoring on journeys that burn money fastest
  • Justify backup paths with the same units finance already uses
  • Decide which alerts deserve pages (high $ / hour) vs. tickets

Related