Contents7
A leadership dashboard is there to show what is happening in the company without waiting for the month-end close or a customer’s chaser. It still has to track the right indicators, the KPIs, and leadership has to have chosen them itself, for the decisions they are meant to inform. This guide from our section on process automation and business bots explains how to define them, how to feed them without manual data entry, and how to use them to see where the flow slows down, including when people are slowing it down despite themselves.
In short
A KPI is chosen for a decision: if it changes nothing about what leadership will do, it has no place.
Financial figures say what happened, flow indicators announce what is coming: you need both.
Each indicator has a written definition, a source, an owner, a threshold and a time when it is read.
The dashboard measures the flow, not the people: in France, the data protection authority, CNIL, strictly regulates monitoring of employees’ activity.
It feeds itself from the tools: one more manual entry is enough to doom it.
What is a leadership dashboard for?
A report tells the story of the past month; a dashboard is for deciding this week. In 1992, professors Robert Kaplan and David Norton built their balanced scorecard on an observation that has become a classic:
What you measure is what you get.
They added that financial measures alone give misleading signals for continuous improvement and innovation. The month’s revenue tells you whether the flow ran well; it does not tell you where it is starting to seize up. For that, you need to track what is moving through it.
This guide covers the whole flow, from a customer’s request to their payment. For marketing on its own, our guide to automated marketing reporting details the figures that show what pays.
Which KPIs show what is moving?
Two families of indicators complement each other. The US workplace safety administration, OSHA, distinguishes them in the context of accident prevention: leading indicators reveal potential problems, lagging indicators measure what has already happened. The distinction holds for any flow:
Leading indicators: what is building up
Files waiting at each step
Waiting time before each step
Age of the oldest file in progress
Lagging indicators: what happened
The month’s revenue and margin
Average lead time over the past quarter
Complaints received
To see the flow, six indicators cover the essentials:
- files in progress, step by step: the pile that grows before the delay shows;
- end-to-end lead time, from a customer’s request to delivery, or from an invoice received to its payment;
- waiting time before each step: this is what points to the bottleneck;
- the age of the oldest file: a reassuring average can hide a customer forgotten for three weeks;
- the rework rate, the share of files that go back a step for a missing document or because of a mistake;
- throughput, the number of files leaving the flow each week.
How do you define KPIs with leadership?
Indicators are chosen in a workshop, with leadership and with the people who do the work: the former know which decisions they have to make, the latter know what is easy to count and what a figure will never say.
- Start from the decisions.
List those leadership makes every week and every month: strengthening a team, following up a customer, revising an approval threshold. Each indicator must inform one of them.
- Write one sheet per indicator.
Its exact definition, its calculation, the tool it comes from, its owner and when it is read. Two departments that do not count a “file in progress” the same way will end up arguing over the figures.
- Set a threshold and an action.
Above what value must someone act, and who acts? An indicator without a threshold gets looked at, not steered by.
- Keep few indicators.
A handful per flow is enough; beyond that, nobody knows which ones to look at.
- Plan for revision.
An indicator that has not triggered a single decision in a quarter gets replaced.
How do you stop an indicator turning against you?
An indicator changes the behaviour it measures. British anthropologist Marilyn Strathern summed it up in 1997, writing about what is known as Goodhart’s law, after the economist Charles Goodhart:
When a measure becomes a target, it ceases to be a good measure.
A processing time turned into an individual target pushes people to close files too quickly, or to pick the easiest ones. The remedy fits in two rules: measure the flow rather than each person, and read the figures together, in a review, to decide on an improvement rather than to single someone out.
Watch out
Measure the flow, do not monitor the people
In France, monitoring employees’ activity is only lawful if it is justified and proportionate, presented to staff representatives and made known to employees, as the CNIL points out, adding that “constant surveillance is excessive”. On 23 December 2025, in the Amazon France Logistique case, the Conseil d’État, France’s highest administrative court, held the warehouses’ real-time indicators lawful, but confirmed that keeping all of them for 31 days breached the data minimisation principle, and reduced the fine from 32 to 15 million euros.
A flow dashboard needs no personal data to show that a queue is growing in front of a step. When tracking by person is justified, the CNIL cites as proportionate a piece of software that transparently counts the number of files processed per quarter and per employee.
How do you feed the dashboard without data entry?
A dashboard filled in by hand does not survive long: entry takes time, arrives late, and nobody trusts a copied figure. The figures must come from the tools where the work already happens:
Where a dashboard’s figures come from
- The tools
The CRM, invoicing and support record each event as it happens.
- Gathering
An automation collects these events at a fixed time, with no data entry.
- The database
A single table, with a written definition per indicator.
- The dashboard
Thresholds and alerts, read on a fixed schedule.
This gathering is a matter of simple rules, without AI: a model has no business calculating an indicator, as our guide Automation without AI: when a rule is enough explains.
On the software side, open source tools do the job and can be hosted on your own server: Metabase and Grafana are released under the AGPL licence, Apache Superset under the Apache 2.0 licence. The real work lies elsewhere: getting the tools to talk to each other, which our article Connecting your business tools: enter data only once details.
From the field
Measuring silence too
On Palmora Property, a property portal we run, the morning report does not just count what happened: it gives the age of the latest backup, and an alert goes out if a task has not reported in for 30 hours. A task that has stopped produces no error: only its absence gives it away.
Where to start?
Start with a single flow and three indicators: files in progress, end-to-end lead time and waiting time before each step. Read them every week for a month before adding more. The reading rhythm matters as much as the figures: our guide to management rituals suggests a rhythm that holds without multiplying meetings, and our article Running your company as a system, not in reaction explains the overall logic.
Our digital and AI maturity audit lays the groundwork: a map of your processes as they really run, and what each of your tools can pass on to the others. Then our automation offer removes the manual compiling of figures, with a meter showing what each automation costs.
Frequently asked questions
How many KPIs should you track?
Few: a handful per flow, and ten at most for the whole of leadership. Each indicator must inform a specific decision and have a threshold that triggers an action. Beyond that, attention scatters and nobody knows which figures to look at first. Start with three flow indicators on one flow, then add one only when a decision calls for it.
What is the difference between a dashboard and reporting?
Reporting tells the story of a past period, often once a month, to account for it. A dashboard is for deciding: it shows the state of the flow now, with thresholds that flag what needs action. The two complement each other. Reporting relies mostly on lagging indicators, such as revenue; the dashboard adds leading indicators, such as files waiting.
Which tool should you choose for a leadership dashboard?
The tool matters less than the data feeding it. Open source software such as Metabase, Grafana or Apache Superset is enough for most companies and can be hosted on their own server. The main work consists of connecting business tools, such as the CRM or invoicing, to a shared database, with a written definition per indicator. Without that work, the best software displays wrong figures.
Can you track employee productivity with a dashboard?
With great care. In France, the CNIL requires monitoring of activity to be justified, proportionate, presented to staff representatives and known to employees, and considers constant surveillance excessive. A management dashboard generally does not need it: it measures the flow, the queues and the lead times, without personal data. That is also what stops the figures being gamed.
How often should a dashboard be updated?
The data can update every day, or even continuously, since it comes from the tools with no data entry. Reading, however, follows a fixed rhythm: a daily glance at the alerts, a weekly review of queues and lead times, a monthly review of trends. What matters is that each reading leads to a decision, even the decision to change nothing.
Sources
- Kaplan and Norton, The Balanced Scorecard, Harvard Business Review, 1992 (new tab)hbr.org
- OSHA, Using Leading Indicators to Improve Safety and Health Outcomes (new tab)osha.gov
- CNIL, monitoring employees' activity (new tab)cnil.fr
- Conseil d’État, decision No. 492830 of 23 December 2025 (new tab)conseil-etat.fr
- Metabase licence (new tab)metabase.com
- Grafana licence (new tab)github.com
- Apache Superset repository (new tab)github.com
Kaplan and Norton, The Balanced Scorecard, Harvard Business Review, 1992
Read on hbr.org(new tab)OSHA, Using Leading Indicators to Improve Safety and Health Outcomes
Read on osha.gov(new tab)Jim · founder of Brain Plus AI
Also in automation and bots
Understand9 min
Running your company as a system, not in reaction
Most companies move forward in reaction: a customer chases, someone searches, things get fixed. Thinking of the organisation as a system of loops shifts the effort: automations relay information, a dashboard shows where it piles up, and people keep the judgement.
Understand8 min
Business bottlenecks: how to find one and free it
A quote waiting for a signature, a single person who knows how: the bottleneck sets the pace for the whole company. How to spot it, why speeding up elsewhere achieves nothing, and how automation clears it without taking the decision away.
Practical guide8 min
Automated marketing reporting: track what pays off
A useful marketing report does not copy out figures: every week it says what each channel earned, straight from the tools and unedited. The figures to track, how to connect Analytics, Search Console and your CRM, what AI can comment on, and the data pitfalls to know before you start.