Google Ads metrics that lie to you.

No metric in Google Ads is dishonest. Every one of them measures exactly what it measures. The lying happens in the gap between what a metric measures and what the person reading it believes it measures — and that gap has funded more bad decisions than any platform bug ever will. These are the five gaps we correct most often in audits, and what we read instead.

CTR: a relevance signal, not a success signal

Click-through rate measures how well the ad attracts clicks — full stop. It says nothing about whether those clicks were worth attracting. "Free," urgency theatrics, and sensational claims all raise CTR while filling the funnel with people who wanted something you do not sell. The inverse is more useful: an ad with modest CTR and excellent conversion rate is pre-qualifying traffic, doing your filtering for free at the auction stage. Read CTR paired with downstream conversion rate, never alone, and treat a CTR spike after a copy change as a hypothesis to check — not a win to report.

Conversion rate: the denominator moves

Conversion rate feels like a quality score for your whole funnel, but its denominator — the traffic mix — moves constantly. Expand into broader keywords and conversion rate falls even if the expansion is profitable; tighten targeting to your warmest audience and it rises while total volume shrinks. Teams that manage to conversion rate end up optimizing for a smaller, safer funnel. The honest read is conversion rate segmented by traffic tier (brand vs non-brand, a split that changes everything it touches) alongside total conversion volume and marginal CPA. Growth almost always looks like conversion rate falling while conversions rise.

Impression share: whose impressions?

Impression share sounds like market share, but the denominator is auctions Google deemed you eligible for — a set that expands and contracts with your own keyword and match-type choices. Add broad match and your impression share can crash while your actual presence on money terms is unchanged, because the eligible universe just tripled. Read it at the campaign level on exact-match brand and core commercial terms, where the denominator is stable and the metric means what you think. And distinguish share lost to budget (a spending decision) from share lost to rank (a quality or bid problem) — the fix for one is the opposite of the fix for the other.

Quality Score: a diagnostic, not a KPI

Quality Score is a 1-10 helper metric summarizing expected CTR, ad relevance, and landing page experience. It is useful exactly once: when it is low, its three components tell you where to look. Managed as a KPI — "raise all keywords to 8+" — it produces absurdities like pausing profitable keywords because a score hurt someone's feelings. We check it during diagnostics and never report it to clients as an outcome. Money metrics are outcomes; Quality Score is plumbing.

Platform ROAS: grading its own homework

Platform-reported ROAS is the platform's own attribution of revenue to its own ads, under its own model, counting many customers who would have arrived anyway. It is genuinely useful for relative decisions inside the platform — which campaign, which product group, which audience. It is structurally unfit for the absolute question of what the channel adds. For that, pair it with MER and incrementality checks. The practical rule we run accounts by: platform metrics allocate, blended metrics evaluate, and neither is allowed to do the other's job.

None of this argues for ignoring dashboards. It argues for reading each number as the answer to its actual question — and for building the small set of paired reads (CTR with conversion rate, conversion rate with volume, ROAS with MER) that make lying to yourself harder. That pairing discipline is most of what separates reporting that drives decisions from reporting that decorates them.