![\"\"](\"https:\/\/www.hammerdb.com\/blog\/wp-content\/uploads\/2026\/01\/C9benchmark.jpg\")
<\/a><\/p>\nThat original idea still matters.<\/p>\n
\n
The Core of a Benchmark: Repeatability<\/h2>\n
At its heart, a benchmark is about repeatability<\/strong>.<\/p>\n If you can\u2019t repeat a measurement and get a comparable result, you don\u2019t really have a benchmark \u2014 you just have a number.<\/p>\n That\u2019s as true for database benchmarks as it was for surveying:<\/p>\n the same workload<\/p>\n<\/li>\n the same schema<\/p>\n<\/li>\n the same scalability<\/p>\n<\/li>\n the same rules<\/p>\n<\/li>\n the same measurement method<\/p>\n<\/li>\n<\/ul>\n Change those, and you\u2019ve changed what\u2019s being measured.<\/p>\n Experimentation isn\u2019t the problem \u2014 it\u2019s the point. A good benchmark makes experimentation meaningful.<\/p>\n When the benchmark is repeatable, changing a single variable \u2014 hardware, configuration, software version, or scale \u2014 means any difference you observe can reasonably be attributed to that change. Cause and effect become visible.<\/p>\n But without a stable reference point, you\u2019re guessing.<\/p>\n Many historic survey benchmarks, particularly metal markers, include a warning stamped directly into them:<\/p>\n \u201cFine or imprisonment for disturbing this mark.\u201d<\/strong><\/p>\n<\/blockquote>\n But the mark itself wasn\u2019t valuable. However, if it could be moved, every future measurement taken from it would be suspect. Accumulated knowledge depended on that reference point remaining exactly where it was. Instead, what was valuable was trust in the measurement.<\/p>\n The problem is that production is rarely a single system.<\/p>\n Over time, organisations accumulate many databases, platforms, and teams. People move on, systems evolve, and each measurement captures a moment that can\u2019t easily be compared with the next.<\/p>\n Each result may be valid on its own, but without a common benchmark none of them relate to each other. You don\u2019t end up with a strategy \u2014 you end up with isolated measurements tied to individual systems and people with different skill-sets, opinions and bias.<\/p>\n And once that happens, you don\u2019t really have a database strategy at all.<\/p>\n You end up with a sprawl: different databases, different platforms, different operating systems, some on-prem, some in the cloud. Each measured differently, at a different time. The numbers don\u2019t line up, and there\u2019s no consistent view of performance \u2014 or of what any of it actually costs.<\/p>\n Once a benchmark is accurate and repeatable, something important changes: you can start to understand why<\/em> things behave the way they do.<\/p>\n You can:<\/p>\n compare hardware generations<\/p>\n<\/li>\n evaluate configuration changes<\/p>\n<\/li>\n understand scaling behaviour<\/p>\n<\/li>\n measure cost versus performance<\/p>\n<\/li>\n make decisions based on evidence rather than instinct<\/p>\n<\/li>\n<\/ul>\n The goal isn\u2019t to recreate every detail of production. It\u2019s to create a stable reference point you can return to \u2014 today, next month, or next year \u2014 and trust that the comparison still holds.<\/p>\n The concept of benchmarks have lasted for centuries because they\u2019re repeatable<\/strong>.<\/p>\n Long before databases or computers, someone looked at the problem of measurement and said: \u201cIf we want to measure something time and time again, we need a fixed reference point we can trust.\u201d<\/em><\/p>\n That simple idea is still the foundation of benchmarking today.<\/p>\n","protected":false},"excerpt":{"rendered":" Most people don\u2019t realise that benchmarks pre-date computers by decades. The idea comes from land surveying. Long before software or databases existed, surveyors faced a simple problem: how do you measure something today and be confident you\u2019re measuring the same thing months or years later? The answer was the BENCH MARK \u2014 a fixed reference … <\/p>\n\n
\n\u201cFine or Imprisonment for Disturbing This Mark\u201d<\/h2>\n
\n
\n\u201cWe Only Benchmark Our Production Systems\u201d<\/h2>\n
\nWhy Repeatability Unlocks Understanding<\/h2>\n
\n
\nIn summary<\/h2>\n