Benchmark Methodology
Managed Connectivity Benchmark 2026 - Methodology
Methodology for an anonymized benchmark covering availability, latency, packet loss, mobile signal, failover tests, and carrier operations.
Core answer
Anexum is preparing an anonymized Managed Connectivity Benchmark based on its managed estate of more than 200 connections. This publication defines metrics, exclusions, quality controls, and privacy rules. It intentionally contains no result values until the sample, measurement intervals, and customer approvals have been validated.
Which population is considered?
Only actively managed fixed-line and mobile connections with sufficiently complete measurement and inventory data are included. The known estate of more than 200 connections is not automatically a valid sample size for every metric.
Which metrics are planned?
Planned metrics include reachability, latency, jitter, packet loss, bandwidth, mobile signal, router state, failover state, alert duration, time to ticket creation, carrier response, and time to stable recovery.
How are different circuits compared?
Technology, bandwidth, region, site type, carrier, SLA, and measurement interval are recorded as segments. Fiber, DSL, coax, leased lines, and LTE/5G are not blended into one average without segmentation.
Which data is excluded?
Exclusions include test circuits, planned maintenance without customer impact, incomplete time series, known monitoring errors, non-representative activation periods, and records that cannot be anonymized adequately.
How is availability calculated?
The methodology must define the denominator, interval, planned maintenance, partial outages, and degradation. A successful ping alone does not automatically mean a usable service. Calculations and exclusions accompany every result.
How is privacy protected?
Customer names, exact locations, IP addresses, contract numbers, and individual carrier prices are not published. Small segments are aggregated or suppressed where re-identification could be possible.
How is quality checked?
Before publication, the process checks periods, missing values, time zones, duplicates, target changes, router replacements, and outliers. Every published metric includes sample size, period, and calculation definition.
When will results be published?
Results are published only after a complete reporting period, documented quality review, and required customer approvals. Until then, this methodology is the explicit boundary against fabricated or premature benchmarks.