Which dashboard metrics survive an audit — and which don't
Every quarter, someone on your exec team asks for the same dashboard: impressions, reach, total clicks, follower growth, "engagement rate." It takes a day to build, gets glanced at twice, and tells nobody whether the business made money. Meanwhile the dashboard that actually drives decisions — true ROAS by channel, contribution margin after returns, cost per retained customer — sits three tabs deep because nobody asked for it directly.
This is the vanity metrics problem, and it's not a reporting issue. It's a trust issue. Execs ask for metrics they recognize, not metrics that matter, because nobody's audited the dashboard stack in two years and explained the difference. Here's how to actually do that audit.
Why vanity metrics survive this long
Vanity metrics aren't stupid — they're legible. Impressions go up and to the right almost by default as you spend more. Follower count never goes negative. Engagement rate feels like proof of "brand health" even when nobody can define what that means in dollars. These metrics survive because they're easy to report, easy to feel good about, and hard to directly blame for bad outcomes.
Real performance metrics are the opposite. True ROAS can go down even when you're doing everything right, because it accounts for returns, discounts, and blended CAC across channels. Contribution margin can look ugly in a growth phase. Nobody wants to be the one presenting a chart that says "we spent more and made less efficiently," even if that chart is correct and actionable.
So the dashboard calcifies around what's comfortable to report, not what's useful to act on. An audit has to name this dynamic before it can fix it.
Start by mapping metrics to decisions
The fastest way to expose a vanity metric is to ask one question of every number on the dashboard: what decision does this change? Not "what does this tell us" — what decision.
- Impressions: does a change in impressions ever change a budget allocation, a creative decision, or a channel investment? Usually no.
- Engagement rate: has anyone ever paused a campaign because engagement rate dipped 0.3%? Rarely, and if they did, was it the right call?
- True ROAS by channel, including returns and discount cost: this directly drives budget shifts every month.
- Cost per retained customer (not just cost per acquired customer): this changes retention spend and loyalty program investment.
- Contribution margin by SKU or category: this changes what gets promoted and what gets discontinued.
Run every metric on the current dashboard through this filter. If nobody can name the specific decision a metric has influenced in the last two quarters, it's a vanity metric wearing a business-metrics costume. Flag it, don't delete it yet — you'll need it for the next step.
Trace the data lineage, not just the number
Before you can argue to remove a metric, you need to understand where it actually comes from, because half the time the real problem isn't that the metric is vanity — it's that it's broken and nobody noticed because nobody uses it to decide anything.
Pull up the pipeline for each core metric. Where does "ROAS" in this dashboard actually originate? Is it platform-reported ROAS from Meta or Google, which counts view-through conversions and ignores returns entirely? Or is it blended ROAS calculated from your own order data, net of refunds, matched back to ad spend by UTM and order attribution window? These are wildly different numbers that often share the same label on a slide.
This is where most dashboard audits fall apart — not on philosophy, but on plumbing. You find:
- Platform-reported conversions that double-count because of overlapping attribution windows across channels
- Revenue figures that don't net out returns, so a product with a 40% return rate looks like a star performer
- "New customer" counts that count account creation, not first purchase, inflating acquisition numbers
- Spend data that excludes agency fees or platform fees, understating true cost per acquisition
Document the actual source, the join logic, and the attribution window for every metric that survives the decision filter. If you can't trace it cleanly, that's a data reliability problem to fix before it's a dashboard design problem. A clean definition of true ROAS — spend fully loaded, revenue net of returns and discounts, attribution window consistent across channels — is worth more than any amount of dashboard redesign.
Reframe vanity metrics instead of killing them
You will not win by telling an exec "impressions don't matter, stop asking for them." That's a fight about ego, not data, and you'll lose it. The better move is to keep the metric but change its context so it can't be misread as success on its own.
Pair every surviving vanity metric with the business metric it should never be viewed without:
- Impressions, shown next to cost per impression and conversion rate from that channel — so a spike in reach with a falling conversion rate reads as a warning, not a win.
- Engagement rate, shown next to revenue per engaged session — so likes and comments stop being treated as a proxy for purchase intent.
- Follower growth, shown next to the percentage of revenue coming from that channel's audience — so growth without monetization gets called out immediately.
- Total clicks, shown next to cost per acquisition and return rate for that traffic source — so cheap, low-quality clicks can't hide behind a big volume number.
This does two things. It satisfies the exec who wants to see the familiar number, and it makes it structurally impossible to present that number as a win without also showing whether it drove real outcomes. Over a few reporting cycles, the paired business metric becomes the thing people actually discuss in the room, and the vanity metric quietly fades into a footnote.
Build the dashboard around decisions, not platforms
The underlying structural mistake in most marketing dashboards is that they're organized by data source — a Meta tab, a Google tab, an email tab, a Shopify tab — instead of by the decision they inform. That structure practically guarantees vanity metrics survive, because each platform reports what makes that platform look good, and nobody forces the numbers to compete on a level playing field.
Rebuild around three or four core decision areas instead:
- Budget allocation: true ROAS by channel, computed the same way regardless of platform, net of returns and fully loaded for spend
- Customer economics: cost per acquisition and cost per retained customer, blended and by channel
- Product and margin: contribution margin by SKU or category, including return rate and discount depth
- Growth health: net new customer revenue versus repeat customer revenue, tracked over time
Every platform feeds into these four buckets through a consistent pipeline, with the same definitions and attribution logic no matter the source. This is also where automation pays off — a reliable data pipeline that pulls order data, ad spend, and returns into one source of truth means you're not manually reconciling five exports every Monday morning, and the dashboard stops drifting out of sync with reality.
Once the dashboard is organized this way, requests for vanity metrics tend to drop on their own. It's hard to ask for "total impressions" as a standalone slide when the dashboard in front of you is organized around "which channel do we fund next month" — the question answers itself with better data, and the old request starts to look like what it always was: noise dressed up as signal.