A benchmark is a comparison point. It tells you whether a number is good or bad by showing what it's being measured against.
For example, SecondBank's analytics show that 12% of weekly active cardholders visit the rewards tab. That's the baseline. When a redesign lifts visits to 31%, the benchmark is what makes the improvement visible.
Why it matters
A number on its own says very little. Task success of 78% sounds good, but nobody knows whether it's better than last quarter, better than competitors or simply normal. Without a comparison point, teams celebrate ordinary results and miss real problems.
Benchmarks give every metric and signal context. They show how far a change actually moved things, they let leaders compare options on the same footing, and they make it harder to cherry-pick the comparison that tells the nicest story.
Set a benchmark before the work starts, record where it came from and when, and use the same one when judging the result.
How we write one
Every benchmark is written as one sentence that names the metric, its baseline value and where it came from:
A metric baseline is a value from a source.
Rewards tab visits baseline is 12% of weekly actives from analytics.
Writing it this way ties every comparison point to a source, so anyone can check where it came from and whether it still applies.
How it differs from similar ideas
-
A UX metric is the current measurement. A benchmark is the point it's compared against.
-
A target in a business intent or result is where the business wants to get. A benchmark is where things stand, or where others stand, today.
-
An outcome is a change from one value to another. A benchmark supplies the starting point or reference for that change.
-
A test's expected result is a prediction for one test, such as 70% task success. A benchmark is an established reference that applies beyond one test.
-
A signal interprets evidence. A benchmark is what a signal uses to judge whether the evidence is strong.
A good benchmark
-
Names one specific metric
-
Gives a baseline value
-
Says where it came from
-
Has a date
-
Says whether it's internal history, a competitor or an industry figure
-
Matches the audience and context it's compared against
What happens to it
-
Found. Ray or the team pulls it from analytics, past research, a competitor review or an industry report.
-
Set. It's linked to the metric it anchors, ideally before work begins.
-
Used. It anchors Comparing in Focus and Mapping in Lead, and helps judge signal strength.
-
Refreshed. When newer data arrives, a new benchmark is stored and the old one is kept for history.
-
Retired. When the product, audience or measurement changes enough, it stops applying.
For Ray
This part tells Ray how to find benchmarks in an input, such as an analytics export, a past research report, a competitor review, an industry study or a planning doc, and how to store them.
How to find one
-
Mark reference values. Read the input and mark every number presented as a baseline, average, norm, history or comparison.
-
Find the metric. Identify exactly what is being measured, and match it to an existing UX metric if one exists.
-
Find the value and date. Pull out the baseline value and when it was measured. If the date is missing, leave it blank.
-
Find the source. Record where it came from: an analytics tool, a past study, a competitor or a published report.
-
Tag the type. Internal history for the product's own past data, Competitor for a rival's figure, Industry for a published norm or average.
-
Check the fit. Note the audience and context it applies to, and whether it matches what it will be compared against.
-
Write the sentence. Use the format: [Metric] baseline is [value] from [source].
-
Fill the record, score it, link it and store it using the rules below.
Cues to look for
-
Comparison language: "baseline", "benchmark", "average", "industry norm", "compared to", "last quarter"
-
Historical values in analytics exports and dashboards
-
Competitor figures in market or competitive reviews
-
Published norms, such as a SUS average of 68
What doesn't count
<table xmlns="http://www.w3.org/1999/xhtml" style="min-width: 50px;"><colgroup><col style="min-width: 25px;"><col style="min-width: 25px;"></colgroup><tbody><tr><th colspan="1" rowspan="1"><p>If the statement...</p></th><th colspan="1" rowspan="1"><p>Store it as</p></th></tr><tr><td colspan="1" rowspan="1"><p>Reports a new measurement from current testing</p></td><td colspan="1" rowspan="1"><p>UX Metric</p></td></tr><tr><td colspan="1" rowspan="1"><p>States where the business wants to get</p></td><td colspan="1" rowspan="1"><p>Business Intent or a Result target</p></td></tr><tr><td colspan="1" rowspan="1"><p>Predicts the result of a single test</p></td><td colspan="1" rowspan="1"><p>The expected result on the Test</p></td></tr><tr><td colspan="1" rowspan="1"><p>Describes a change from one value to another</p></td><td colspan="1" rowspan="1"><p>User, Product or Business Outcome</p></td></tr><tr><td colspan="1" rowspan="1"><p>Describes what's happening without a number</p></td><td colspan="1" rowspan="1"><p>Situation</p></td></tr></tbody></table>
The record
<table xmlns="http://www.w3.org/1999/xhtml" style="min-width: 75px;"><colgroup><col style="min-width: 25px;"><col style="min-width: 25px;"><col style="min-width: 25px;"></colgroup><tbody><tr><th colspan="1" rowspan="1"><p>Field</p></th><th colspan="1" rowspan="1"><p>What to fill in</p></th><th colspan="1" rowspan="1"><p>Values</p></th></tr><tr><td colspan="1" rowspan="1"><p>Statement</p></td><td colspan="1" rowspan="1"><p>The one-sentence benchmark</p></td><td colspan="1" rowspan="1"><p>[Metric] baseline is [value] from [source].</p></td></tr><tr><td colspan="1" rowspan="1"><p>Metric</p></td><td colspan="1" rowspan="1"><p>What is being measured</p></td><td colspan="1" rowspan="1"><p>Short name; link to UX Metric if one exists</p></td></tr><tr><td colspan="1" rowspan="1"><p>Baseline value</p></td><td colspan="1" rowspan="1"><p>The reference value</p></td><td colspan="1" rowspan="1"><p>Number with unit</p></td></tr><tr><td colspan="1" rowspan="1"><p>Type</p></td><td colspan="1" rowspan="1"><p>Where the comparison comes from</p></td><td colspan="1" rowspan="1"><p>Internal history, Competitor, Industry</p></td></tr><tr><td colspan="1" rowspan="1"><p>Source</p></td><td colspan="1" rowspan="1"><p>The tool, study, competitor or report</p></td><td colspan="1" rowspan="1"><p>Short text</p></td></tr><tr><td colspan="1" rowspan="1"><p>Date</p></td><td colspan="1" rowspan="1"><p>When the baseline was measured</p></td><td colspan="1" rowspan="1"><p>Date or period</p></td></tr><tr><td colspan="1" rowspan="1"><p>Audience and context</p></td><td colspan="1" rowspan="1"><p>Who and what it applies to</p></td><td colspan="1" rowspan="1"><p>Link to Audience; short text</p></td></tr><tr><td colspan="1" rowspan="1"><p>Anchors</p></td><td colspan="1" rowspan="1"><p>What it's used to compare</p></td><td colspan="1" rowspan="1"><p>Links to UX Metric, Signal, Method or Initiative</p></td></tr><tr><td colspan="1" rowspan="1"><p>Input</p></td><td colspan="1" rowspan="1"><p>The input and exact table, chart or passage</p></td><td colspan="1" rowspan="1"><p>Link to Input</p></td></tr><tr><td colspan="1" rowspan="1"><p>Status</p></td><td colspan="1" rowspan="1"><p>Where it stands</p></td><td colspan="1" rowspan="1"><p>Found, Set, Used, Refreshed, Retired</p></td></tr><tr><td colspan="1" rowspan="1"><p>Confidence</p></td><td colspan="1" rowspan="1"><p>How reliable the comparison point is</p></td><td colspan="1" rowspan="1"><p>High, medium, low</p></td></tr></tbody></table>
New benchmarks start as Found. They move to Set once linked to the metric they anchor.
Scoring confidence
Confidence measures how reliable and well matched the comparison point is, not whether the number is flattering.
<table xmlns="http://www.w3.org/1999/xhtml" style="min-width: 50px;"><colgroup><col style="min-width: 25px;"><col style="min-width: 25px;"></colgroup><tbody><tr><th colspan="1" rowspan="1"><p>Level</p></th><th colspan="1" rowspan="1"><p>When to use it</p></th></tr><tr><td colspan="1" rowspan="1"><p>High</p></td><td colspan="1" rowspan="1"><p>From a named source with a date, measured the same way and for the same audience as what it's compared against</p></td></tr><tr><td colspan="1" rowspan="1"><p>Medium</p></td><td colspan="1" rowspan="1"><p>Source and value are clear, but the date is missing or the audience or method differs somewhat</p></td></tr><tr><td colspan="1" rowspan="1"><p>Low</p></td><td colspan="1" rowspan="1"><p>Quoted without a source, out of date, or measured in a way that doesn't match</p></td></tr></tbody></table>
If any part of the record can't be filled from the source, leave it blank and lower confidence by one level.
Linking
-
Check for duplicates first. If an existing benchmark has the same metric, source and date, add the new input to it instead of creating another.
-
Link to the metric it anchors and to its source input.
-
Link to the Comparing and Mapping work it supports: the comparisons and initiatives in Focus, and the outcomes in Lead.
-
Link to signals whose strength it helps judge.
-
Note replacements. When a newer baseline is stored, link it to the old one and mark the old one Refreshed.
When to store and when to ask
-
Store automatically when confidence is high.
-
Add to the review digest when confidence is medium or low, when two benchmarks for the same metric disagree, or when it anchors a bet.
When to flag
-
A UX metric or outcome has no benchmark to compare against
-
A benchmark is more than a year old and still in use
-
Two benchmarks for the same metric disagree
-
A benchmark's audience or method doesn't match what it's compared against
-
A benchmark was set after the result came in
Worked example
Input: a weekly analytics export for the SecondBank card app, September 2026.
"Rewards tab: unique visitors / weekly active users = 12.1% (4-week average, Sept 1 to 28). New cardholders, first 30 days: 12.0%."
What Ray finds:
-
"Rewards tab visitors / weekly active users" is the metric, matching the existing rewards tab visits metric.
-
"12%" is the baseline value, from a four-week average in September 2026.
-
The source is the product's own analytics, so the type is Internal history.
-
The new cardholder figure is a separate benchmark for that segment, linked to this one.
Stored record:
<table xmlns="http://www.w3.org/1999/xhtml" style="min-width: 50px;"><colgroup><col style="min-width: 25px;"><col style="min-width: 25px;"></colgroup><tbody><tr><th colspan="1" rowspan="1"><p>Field</p></th><th colspan="1" rowspan="1"><p>Value</p></th></tr><tr><td colspan="1" rowspan="1"><p>Statement</p></td><td colspan="1" rowspan="1"><p>Rewards tab visits baseline is 12% of weekly actives from analytics.</p></td></tr><tr><td colspan="1" rowspan="1"><p>Metric</p></td><td colspan="1" rowspan="1"><p>Rewards tab visits as a share of weekly active users</p></td></tr><tr><td colspan="1" rowspan="1"><p>Baseline value</p></td><td colspan="1" rowspan="1"><p>12%</p></td></tr><tr><td colspan="1" rowspan="1"><p>Type</p></td><td colspan="1" rowspan="1"><p>Internal history</p></td></tr><tr><td colspan="1" rowspan="1"><p>Source</p></td><td colspan="1" rowspan="1"><p>Card app product analytics</p></td></tr><tr><td colspan="1" rowspan="1"><p>Date</p></td><td colspan="1" rowspan="1"><p>September 1 to 28, 2026</p></td></tr><tr><td colspan="1" rowspan="1"><p>Audience and context</p></td><td colspan="1" rowspan="1"><p>Customers: all weekly active cardholders</p></td></tr><tr><td colspan="1" rowspan="1"><p>Anchors</p></td><td colspan="1" rowspan="1"><p>Rewards tab visits Product KPI; A/B test of progress bar variants</p></td></tr><tr><td colspan="1" rowspan="1"><p>Input</p></td><td colspan="1" rowspan="1"><p>Weekly analytics export, September 2026</p></td></tr><tr><td colspan="1" rowspan="1"><p>Status</p></td><td colspan="1" rowspan="1"><p>Set</p></td></tr><tr><td colspan="1" rowspan="1"><p>Confidence</p></td><td colspan="1" rowspan="1"><p>High: named source, dated, same measure as the comparison</p></td></tr></tbody></table>
Next step: stored automatically and linked to the A/B test of progress bar variants, so its results can be compared against 12%.
Edge cases
-
Mismatched comparisons. A competitor's figure for all users isn't a fair benchmark for new cardholders. Store it, note the mismatch, and keep confidence at medium or lower.
-
Moving baselines. If tab visits drift from 12% to 14% before the change ships, store the newer baseline and link it as a refresh; don't overwrite the old one.
-
No source given. "Industry standard is around 30%" with no source is low confidence. Store it and send it to review.
-
Benchmark set after the fact. A baseline picked once the results are in can make any change look good. Flag it.