There is a scene that repeats itself in almost every measurement meeting: someone projects a dashboard with twenty metrics in green, and even so, no one knows whether the campaign worked. The problem is almost never the tool. It is the layer of beliefs that has been built around digital marketing KPIs. Here are six common myths, along with the reality that usually appears when you look at the raw data.
Myth 1: “The more KPIs, the more control”
A dashboard with forty metrics doesn’t provide control; it provides noise. In practice, teams that function well work with three or four indicators per channel and one star metric per business objective. The rest are diagnostic metrics that are consulted when something deviates, not every morning.
Reality: the usefulness of a KPI depends on whether someone would make a different decision upon seeing it. If the answer is no, it’s superfluous.
Myth 2: “A high CTR is always a good sign”
A high CTR can mean two opposite things: a highly relevant ad or an ad that promises something the landing page doesn’t deliver. In lead generation campaigns with long forms, a high CTR accompanied by a very low conversion rate usually indicates a mismatch between the ad’s message and the landing page.
Expert nuance: CTR in isolation is not useful for judging creatives. It must be read alongside the conversion rate and cost per acquisition, and even then it can be misleading when volume is low: with few impressions, percentage variations are statistically unstable.
Myth 3: “ROAS says everything about a campaign”
Return on ad spend is convenient because it’s a single number, but it ignores margin. A ROAS of 4 can be excellent for a product with a 70% margin and ruinous for one with 15%. It also ignores the part of the funnel that doesn’t buy at the moment: in long-cycle sectors, attributing the entire sale to the last click inflates the performance of brand campaigns and penalizes discovery campaigns.
According to official documentation from the main analytics platforms, data-driven attribution models distribute credit among multiple touchpoints, but they still depend on the quality of tagging. If event tracking is poorly implemented, no model saves the measurement.
Myth 4: “Vanity metrics are useless”
It’s repeated so often that it has become another cliché. Followers, reach, or view time are not conversion KPIs, but they are early signals: a collapse in organic reach can anticipate a traffic drop within weeks. The key is to assign them an alert role, not a results-reporting role.
A specific case that illustrates the problem
In an accessories ecommerce project, the team celebrated the growth of organic sessions every month. When cross-referencing that data with the conversion rate by device, it emerged that more than 60% of the increase came from mobile devices with a bounce rate above 85%: category pages that loaded slowly. Sessions were rising and revenue was stagnating. The missing KPI wasn’t about traffic, it was about session quality segmented by device. That type of finding isn’t visible in an aggregated dashboard.
Myth 5: “If the KPI drops, you have to change the strategy”
Not every drop is a failure. A decrease in impressions can come from an improvement in intent-based targeting; an increase in cost per lead can reflect a seasonal auction change. Before touching anything, it’s worth reviewing three things: whether the metric’s definition changed, whether the data source changed, and whether the market context changed.
Honest limitation: KPIs measure what the measurement system is capable of capturing. Cookie consent, ad blockers, and browser tracking restrictions reduce the real visibility of conversions. Any year-over-year comparison must assume that margin of error.
Myth 6: “A KPI works for any business”
The cost per lead of a dental clinic and that of a B2B SaaS are not comparable, not even within the same sector. What can be compared is the indicator’s trend against your own historical series. Public sector benchmarks, according to industry studies, serve as rough guidance, never as a contractual target.
How to know if your KPI is well chosen
- It has a clear owner within the team.
- It can be influenced by concrete actions.
- It has a written definition and a documented data source.
- There is an agreed threshold that triggers a review.
If an indicator doesn’t meet these four conditions, it’s probably just a curious data point rather than a KPI.