Celiq
v0.12Open app โ†—
Docs/Building with Celiq/Benchmarks

Benchmarks

See your live data and an industry benchmark in the same answer. Your CAC isn't just a number โ€” it's a percentile against comparable companies.

TracerWeaverLuminaryUpdated Jun 2026 ยท 7 min read
โœฆ
In this section
Part of Celiq's semantic layer platform. Connect your warehouse, model your data once, query it everywhere.
๐Ÿ’ก
Tip
What it is: A curated benchmark dataset that Orion injects into answers โ€” so when you ask about a metric, you get your value and how it compares to your industry.
Why use it: Context. Knowing your CAC is $38 is useful; knowing it beats ~65% of comparable ecommerce companies is actionable.
Where to find it: Automatically in Orion answers, and on the Benchmarks page.
Who can use it: All roles.

Benchmark mode answers the question every metric quietly raises: is this good? When Orion produces a metric value, it checks whether Celiq hosts an industry benchmark for that metric โ€” and if so, it tells you where you stand, in the same response. No other step, no separate report.

Ask "how does my CAC compare to ecommerce benchmarks?" and Orion answers with your live data and the industry benchmark together. You can also browse the full benchmark set on the Benchmarks page.

How benchmarks appear in answers

Whenever an Orion answer includes a specific metric value, Orion looks up a matching benchmark for your workspace's industry. If one exists, it appends a benchmark line to the response:

๐Ÿ“Š Benchmark: Your CAC of $38 is better than ~65% of ecommerce companies
(industry median: $35, top quartile: $20)

This is automatic โ€” there's nothing to enable. Orion detects your industry from your workspace memory or domain, finds the closest matching benchmark, and works out your percentile. If no benchmark matches the metric, the answer simply omits the benchmark line.

The Benchmarks page

The Benchmarks page is a browsable explorer for the full benchmark set. Filter by industry and read each metric's percentile distribution:

Industry Benchmarks
[Filter: Ecommerce | SaaS | Marketing]

Metric            p25      Median   p75      p90
CAC               $20      $35      $65      $120
Return Rate       3%       6%       12%      18%
Conversion Rate   1.2%     2.1%     3.8%     5.5%
AOV               $45      $72      $120     $210
LTV               $85      $145     $280     $520

Each row expands to show the source citation and a methodology note, so you can judge how comparable a benchmark is to your own business before you act on it.

Which metrics have benchmarks

Celiq ships curated benchmarks for the most-watched metrics in each industry. Coverage grows over time, but at launch the set includes:

Ecommerce

Metricp25Median (p50)p75p90
CAC$20$35$65$120
Return rate3%6%12%18%
Conversion rate1.2%2.1%3.8%5.5%
AOV$45$72$120$210
LTV$85$145$280$520

SaaS

Metricp25Median (p50)p75p90
Churn rate2%5%8%12%
NPS20355065
CAC payback period8 mo14 mo22 mo36 mo

If you ask about a metric that isn't in this set, Orion answers normally โ€” it just won't add a benchmark line.

How to read your percentile

Each benchmark reports four percentiles โ€” p25, p50 (the median), p75, and p90. Your percentile tells you what fraction of comparable companies you sit above or below.

For "lower is better" metrics like CAC, return rate, or churn, a lower value is a higher percentile of performance. A CAC of $38 sits just above the $35 median, which Orion expresses as "better than ~65% of ecommerce companies" โ€” because most companies pay more to acquire a customer.

For "higher is better" metrics like conversion rate, AOV, LTV, or NPS, a higher value is a higher percentile.

๐Ÿ“
Note

Orion always phrases the comparison in plain terms โ€” "better than ~65%" or "top quartile" โ€” rather than leaving you to interpret the raw number. The percentile is directional: it already accounts for whether higher or lower is the good direction for that metric.

How benchmarks are sourced

Benchmarks are a curated, hosted dataset maintained by Celiq, assembled from public industry surveys and reports. Each benchmark records its source and the year it reflects, both visible when you expand a row on the Benchmarks page. They are updated periodically as new survey data becomes available.

Benchmark data is shared across all workspaces โ€” it is not derived from, mixed with, or influenced by your own data, and your data is never contributed back into the benchmark set. The comparison runs one way: your live value is measured against the public benchmark.

Limitations

  • Industry-wide, not segment-specific. Benchmarks reflect a whole industry. A luxury retailer and a discount marketplace are both "ecommerce" here, so treat the comparison as a broad reference point, not a precise peer match.
  • Not every metric is covered. Only the curated set above has benchmarks today. Other metrics answer normally, without a benchmark line.
  • Point-in-time sources. Each benchmark reflects a specific source and year. Check the methodology note on the Benchmarks page to confirm it's still relevant to your situation.
  • Your metric definition must align. Orion matches benchmarks by metric name. If your workspace defines a metric differently from the industry convention (for example, a CAC that excludes certain spend), the comparison may not be apples-to-apples. Review the metric's definition before acting on the percentile.