Peer group and statistical methods

Benchmarks do not show individual organizations. They show what the distribution of peers looks like.

Peer group construction

A peer group is typically defined by:
  • Industry
  • Size band (usually revenue)
  • Region or geography
Rules applied in the background:
  • Minimum sample sizes before a metric is reported
  • Exclusion or trimming of extreme outliers in some datasets
  • Anonymization that prevents identifying individual contributors

Percentiles and quartiles

Common statistics:
  • Median – the 50th percentile
  • Lower quartile – the 25th percentile
  • Upper quartile – the 75th percentile
  • Minimum and maximum, sometimes with outliers trimmed
Charts usually represent:
  • Median as a line or dot
  • Interquartile range as a band or box
  • Your organization as a separate marker
Interpreting:
  • If you are near median, you are typical
  • Inside the band, you are within a normal range
  • Outside the band, you are structurally different and should know why
    Interactive Benchmarking box plot

Archetypes

Archetypes are patterns derived by clustering organizations based on their spend structure.

Types supported with Interactive Benchmarking data:
  • Hardware Technology-centric: A Technology operating model where delivery relies heavily on owned or directly managed infrastructure. Typical characteristics include:
    • Larger on-prem or colocation footprint, more physical infrastructure to run.
    • More spend on infrastructure platforms and underlying tech stacks.
    • Often more internally managed operations (or at least hardware ownership), even if some labor is outsourced.
  • Software Technology-centric: A Technology operating model where delivery relies heavily on software platforms, tooling, and licenses rather than owned hardware. Typical characteristics include:
    • Higher use of SaaS, platform subscriptions, automation tooling, and enterprise software stacks.
    • Infrastructure may be abstracted (cloud or managed), shifting emphasis away from hardware ownership.
    • More spend on security, observability, ITSM, collaboration, developer platforms, and enterprise apps.
  • Vendor-centric: A Technology operating model where a large portion of work is performed by third parties. Typical characteristics include:
    • Heavy managed services, systems integrators, outsourcing partners, and contractors.
    • Internal teams focus more on governance, architecture, vendor management, and product ownership.
    • Delivery capacity is scaled by vendors more than hiring.
  • People-centric: A Technology operating model where capability is delivered primarily by internal employees. Typical characteristics include:
    • Larger in-house engineering and operations teams.
    • Strong internal ownership of delivery and run activities.
    • Vendors may exist, but internal teams do most of the work.
Use cases:
  • Understanding which broad pattern your organization resembles.
  • Explaining that you look different from the median because you intentionally follow a different archetype.
Archetype radar charts appearing in the “Industry & OpEx Benchmarks” report in Costing.