Vanity Metrics
Date: 2026-08-17
Numbers that reliably go up and don’t inform any decision. The test isn’t whether a metric is impressive — it’s whether a different value would have changed what you did, and most cumulative totals fail it by construction.
A vanity metric is one that looks good when reported but doesn’t connect to a decision — typically cumulative, or trivially easy to inflate.
The test
“If this number were half what it is, what would I do differently?”
If the answer is “nothing”, or “I’d have to look at something else to find out”, it’s a vanity metric. Applied honestly this eliminates most of what appears on a monthly reporting deck.
metric if it halved, what changes?
total registered users nothing — most never returned anyway
cumulative orders all-time nothing — it can only go up
page views nothing without knowing which pages, and why
social followers nothing
email list size nothing — deliverability might even improve
↑ all vanity
7-day repeat purchase rate investigate immediately — something broke
or the cohort changed
contribution margin per order pricing and shipping decisions change
checkout completion rate a specific funnel investigation starts
↑ all actionable
The three shapes they take
Cumulative totals. Anything all-time or since-launch can only rise. It measures the passage of time, not performance.
total orders, all time
Jan 412,000
Feb 431,000 +19,000
Mar 447,000 +16,000 ← the business is slowing
Apr 461,000 +14,000 ← and slowing again
May 473,000 +12,000
the headline number went up every month.
the only informative figure is the column nobody puts on the slide.
Rates over rising denominators. A ratio that improves because the denominator fell is not an improvement, and it’s the most common way a vanity metric passes as an actionable one — Metric Design.
Counts with no denominator. “4,200 add-to-carts this week” is uninterpretable without knowing how many people were on the site. Rates over counts, nearly always.
Why they persist
Worth being unsentimental about this, because “just use better metrics” has been said for fifteen years and hasn’t worked.
- They’re the only numbers guaranteed to look good. A metric that can go down is a metric someone might be blamed for
- They’re easy to collect, so they’re available first and become habitual
- They make a compelling narrative for anyone not close enough to check
- Nobody is accountable for them, which is precisely why they’re comfortable to report
The reporting culture creates them. If a monthly deck must show progress, the metrics that survive are the ones that always show progress. Fixing the metrics without fixing the expectation just produces new vanity metrics.
Rehabilitating one
Most vanity metrics have an actionable version hiding behind them, and the fix is usually a denominator, a time window, or both.
vanity → actionable
total registered users → users who purchased in the last 90 days
page views → product page → add-to-cart rate, by template
email list size → engaged subscribers (opened in 30d) and the
trend in that as a proportion of the list
app downloads → day-7 retention of installs
total revenue → contribution margin per active customer
Contribution Margin
sessions → sessions per user, and purchasing users
Sessionisation
Each right-hand metric can go down, which is the property that makes it worth having. Two of them lean on definitions worth having settled first: what counts as a session — Sessionisation — and what’s left after variable costs, which is Contribution Margin rather than revenue.
The ones that are legitimately not decisions
Not every number needs to drive an action, and over-applying the test is its own error:
- Scale for context. Knowing you have 400,000 customers rather than 40,000 changes what problems are worth solving, even if the number itself is never acted on
- External reporting. Investors and boards ask for totals, reasonably. Report them, and don’t run the business on them
- Denominators. Traffic isn’t a vanity metric when it’s the base of a rate you do act on
- Diagnostics. A number that explains a movement in an actionable metric is doing useful work even though nobody acts on it alone — Secondary and Diagnostic Metrics
The distinction is the claim being made. “We have 400,000 customers” as context is fine; “customers grew to 400,000, so the strategy is working” is a vanity metric doing argumentative work it can’t support.
Where it interacts
- North Star Metric — a north star that only ever rises has become a vanity metric with better branding
- Metric Design — the denominator choice is where most of this is decided
- Engagement Metrics — the category most prone to this, since “engagement” is definable to suit whatever went up
- Dashboard Design — vanity metrics survive because there’s room for them; a dashboard built around a decision has no space