How to decide with data without drowning in dashboards: the 5 metrics that really matter

Series: Decisions that Move the Needle: Article 5 of 10

The promise of the data era was simple: more information, better decisions. Ten years later, most executives live a different reality: dozens of dashboards, hundreds of indicators, meetings where no one debates (because everyone is looking at different screens) and a persistent feeling that the company has all the data except what it needs to make decisions.

This article defends a contrarian thesis backed by evidence: the problem is not a lack of data, it is a lack of hierarchy. And it proposes a simple structure: instead of accumulating KPIs, organize metrics by the type of decision they serve, and stick with five.

The data paradox: too much measurement, too little decision-making

I start with the two extremes of the paradox.

Well-used data pays off. The classic study by Brynjolfsson, Hitt, and Kim on thousands of firms found that those adopting data-driven decision-making (DDD) achieve productivity 5–6% higher than expected given the rest of their investments (Brynjolfsson, Hitt & Kim, 2011). McKinsey, for its part, measured that intensive users of customer analytics are 23 times more likely to clearly outperform their competitors in acquiring new customers than non-intensive users (McKinsey, 2014).

Misused data is costly. The same decade that produced those benefits also yielded another, less publicized finding: Gartner estimates that poor data quality costs organizations, on average, at least US$12.9 million per year (Gartner, 2020). And the annual survey by Wavestone/NewVantage Partners of large company executives found that, despite heavy investments in data and AI, only 23.9% managed to build a data-driven organization, and barely 20.6% established a data culture (Wavestone/NewVantage Partners, 2023).

Read the complete paradox: companies that use data best gain huge advantages (5–6% productivity, 23x in acquisition), while the majority accumulates costs (US$12.9 million from dirty data) and fails to achieve transformation (23.9%). The difference between the two groups is not the volume of dashboards. It is the discipline of connecting each data point with a decision.

Why dashboards fail: three concrete mechanisms

1. Metrics without an owning decision. Every KPI that no one can connect to a concrete action is noise with professional formatting. The filtering question is brutal and simple: "what decision would change if this number goes up or down?" If the answer is "none in particular," that indicator is merely decorating a screen.

2. Goodhart's Law operating in silence. When a measure becomes a target, it ceases to be a good measure, and teams know this before leadership does. The KPI for "inquiries generated" produces inflated inquiries; the one for "tickets closed" produces rushed ticket closures. It is not bad faith: it is the system optimizing exactly what you asked it to measure, at the expense of what you actually wanted.

3. Choice overload. Just like with tools and strategies, more options do not produce better decisions: they produce paralysis. A dashboard with 40 indicators is not 40 helpful aids; it is 40 candidates for an explanation, and the meeting is spent trying to agree on which one to look at.

An example of what discipline looks like in practice

To bring this hierarchy down to earth, let's see how it translates into a typical week for a service SMB with 30 employees (an illustrative example, with hypothetical figures):

Notice what does NOT happen during that week: no one debates metrics that do not connect to an action, no one proposes "adding a new dashboard," and no decision is made on pure intuition, but neither does anyone wait for the perfect report. That is the balance that evidence rewards: data serves the decision, not the ritual.

The hierarchy: one metric per decision type

The antidote is to arrange metrics by the decision they serve. In an SMB, decisions fall into five types, and each type needs a core metric: just one, the one that if it changes today, forces action this week.

Three clarifications regarding this table. First: core metrics are structural suggestions: the exact name varies by industry (in subscriptions it is churn; in retail, repurchase; in services, repeat rate), but the function is the same: one metric per decision type. Second: frequency matters as much as the metric; looking at CAC every day does not improve decisions, it improves anxiety. Third: this hierarchy does not replace the team's operational dashboards, it replaces them in the leadership meeting, which is where paralysis costs the most.

The 5 rules for operating the hierarchy without drowning

1. One metric, one predefined action. For each core metric, define the healthy range and what to do outside of it in advance. "If the CAC for channel X exceeds Y for two consecutive months, we reallocate Z% of the budget": that way data triggers decisions instead of debates.
2. Goals are chosen before results. To protect yourself from Goodhart's Law, set the target and measurement method before the team begins executing. A KPI defined after the fact always conforms to the result someone wanted to show.
3. Fewer sources, higher quality. Every new data source brings maintenance costs and the risk of contradiction. Before adding a source, ask yourself if it addresses a decision in the table: if not, it is another shiny object in database format.
4. The leadership meeting looks at 5 numbers, not 50. A one-page dashboard with the five core metrics, their trend, and their status traffic light. Detailed operational dashboards continue to exist, but at their own level, with their respective teams.
5. Quarterly data audit. Dirty data does not clean itself: set aside a quarterly consistency review (do the numbers from different sources match up?): it is the practical defense against the average cost of US$12.9 million measured by Gartner (Gartner, 2020).

The final criterion

The competitive advantage of the next decade will not be having data: everyone will have it. It will be making decisions with less data than the competition, but better chosen. The evidence from Brynjolfsson and McKinsey shows that the prize goes to whoever uses data with intensity and discipline, not to whoever accumulates screens. Five well-chosen metrics, each connected to a real decision, outperform forty ownerless indicators. Start by eliminating: your next breakthrough in decision-making probably will not come from a new dashboard, but from one less.

Sources consulted

1. Brynjolfsson, E., Hitt, L. M. & Kim, H. H. (2011). Strength in Numbers: How Does Data-Driven Decisionmaking Affect Firm Performance? SSRN. Accessed October 14, 2026: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1819486
2. McKinsey & Company (2014). Five facts: How customer analytics boosts corporate performance. Accessed October 14, 2026: https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights/five-facts-how-customer-analytics-boosts-corporate-performance
3. Gartner (2020). Data Quality (topic page, research on the cost of poor data quality). Accessed October 14, 2026: https://www.gartner.com/en/data-analytics/topics/data-quality
4. Wavestone / NewVantage Partners (2023). Data and Analytics Leadership Annual Executive Survey: Survey Findings. Accessed October 14, 2026: https://static1.squarespace.com/static/62adf3ca029a6808a6c5be30/t/63a477e23d7f9162b7771d6b/1671722982363/WAVESTONE+NVP--+Data+%26+Analytics+Executive+Leadership+2023--Survey+Findings.pdf