KPI dashboard for a business with a team: a few metrics you actually read
A KPI dashboard is one page with the handful of metrics that tell you whether your business is running: money, customers, workflows. Many businesses with a team either don't have one, have one nobody updates anymore, or have one for which the owner pulls the numbers together personally every Monday. In that last case, even the view of the business depends on you. In this article, we show which KPIs for small business are enough for a team of 3 to 15 people, what every metric needs so that it actually triggers something, and how the numbers appear on their own when processes send a report after every run.
What a KPI dashboard is, and what it isn't
KPI stands for key performance indicator. A dashboard shows several of them at a glance. That also makes clear what it isn't: not a report with dozens of charts, and not a project that requires you to introduce new software first.
The difference between a number and a metric is simple. A number describes something, for example revenue in October. A metric also has a threshold and a consequence: if the value drops below it, something happens, and it's clear who acts. A dashboard full of numbers without thresholds gets looked at and closed again. A dashboard with a few metrics and clear consequences becomes a tool.
Which business metrics a business with a team needs
Which metrics matter for your business depends on what you do. There's no list that applies to everyone, and we deliberately don't give industry benchmarks here. As a starting point, three areas help, and you pick two or three metrics from each:
Money. For example monthly revenue, outstanding invoices and how long the money in the bank will last. These numbers usually come from bookkeeping and the bank, and your tax advisor can tell you which ones are meaningful for your business.
Customers. For example new inquiries, quotes sent and the orders that came out of them. They show you whether enough is coming in and where sales get stuck.
Workflows. For example open tasks, the turnaround time of important processes and exceptions a person had to sort out. These metrics are often missing from small business dashboards, even though they show where the work gets stuck.
Less is more here. Our house rule: the dashboard fits on one screen. Whatever doesn't fit is probably not a key metric.
What every metric needs
A metric without an owner doesn't get maintained. A metric without a threshold doesn't trigger anything. That's why we write down five things for every metric:
- Definition: What exactly gets counted, and what doesn't? "Inquiries", for example, means all new customer inquiries by email, phone and form, not follow-up questions about ongoing orders.
- Source: Where does the number come from? Which system, which report?
- Frequency: How often is it updated? Daily, weekly, monthly?
- Responsibility: Which role looks at it and acts when the threshold is crossed?
- Threshold and consequence: At which value does what happen?
As a template to copy, it looks like this:
Metric: name.
- Definition: What gets counted, what doesn't?
- Source: system or report.
- Frequency: daily, weekly or monthly.
- Responsibility: a role, not a name.
- Threshold: At which value do we act?
- Consequence: What happens then, and by when?
Example: quotes without follow-up. Definition: quotes sent whose follow-up date has passed without anyone following up. Source: the quotes folder or the report of the quoting process. Frequency: weekly, on Mondays. Responsibility: office. Threshold: you set it, for example any quote more than a week past its follow-up date. Consequence: the office follows up by Wednesday and notes the result.
Item 4 is the important one. If the owner is listed for every metric, you have a dashboard, but the business still depends on you. How to hand off responsibility so that it doesn't come back to you as a question is covered in Delegating tasks without them coming back to you as questions.
The real problem: who collects the numbers?
A dashboard often fails less because of which metrics were chosen than because someone has to pull the numbers together. Often that's the owner, on the weekend or on Monday morning. They open the bookkeeping, count the quotes, ask the team for an update. That takes time, and it makes the view of the business depend on one person.
The answer is the same as for any other process: collecting the numbers is a process in itself. It has a trigger, such as every Monday at eight, a start, the work, an end and a report. It can be written down and handed off, to a person on your team or to an AI employee. How to find and choose a process like this is covered in Creating a process map: an example and choosing process no. 1.
Even better is when the numbers don't have to be collected at all, because the processes produce them on their own.
Process metrics from your AI employees' reports
A process that runs at Autonomy Level 2, Autopilot, in our approach starts on its own on a schedule or trigger, runs through to the final report and hands anything uncertain to a person. After every run, you get a report: what's done, what's still open, where there's an exception.
From these reports, you can read process metrics without anyone counting them separately:
- Runs: How often did the process run?
- Done: How many tasks were completed?
- Open: What's still waiting, and since when?
- Exceptions: How many cases went to a person through the escalation path?
- Level: Which autonomy level is the process at right now? You set that in the monthly review.
The last metric is especially telling for you as the owner. On the map of your processes, you can see for each process how far it already runs on its own. How the four levels are built is covered in The four autonomy levels: how a process in your business runs on its own.
The exceptions deserve a second look. If they rise in a process, a rule is usually missing. That's why, once a month, we look at all processes together: what stays, what moves up a level, what's missing. How this monthly review works is covered in Keeping automated processes under control: reports, exceptions and the monthly review.
From dashboard to morning report
You have to open a dashboard. A report comes to you. For processes that run on Autopilot, the next step is therefore often not a nicer dashboard but a report that bundles the exceptions.
At Autonomy Level 3, Audited, an AI manager reviews the AI employees' work and bundles the reports, so you only see the exceptions. In the Owner Autonomy program, that's a goal after the first process: an AI manager bundles the reports into a morning report and flags exceptions. In Kevin's business, that role is still on trial.
For the dashboard, this means the process metrics no longer have to sit in a spreadsheet you go looking for. They're in a short text that arrives in the morning, and you only have to act where a threshold has been crossed. The money and customer metrics still stay on a page you look at regularly, because that's where you decide on direction.
Which tool?
A spreadsheet is enough to start. More important than the tool is that every metric has the five details above and that it's clear who delivers the numbers. If you later want to connect more sources, you can move to a dedicated reporting tool. Which one fits depends on the software you use today, and we sort that out case by case, not on a blog.
A dashboard in new software that nobody fills in is worse than a spreadsheet that's right every Monday. And it belongs in the same place as your process profiles, checklists and rules, as described in Business systems: how the knowledge in your head becomes something your team can use.
What this looks like for us
We don't offer a "KPI dashboard" product. What we build are processes that report after every run, and the process metrics come out of those reports.
At a tax advisory firm, an AI employee files the digital tax assessment notices. Before, there were up to 30 or 40 a day, and the office team needed about 2 minutes for each one. Today, filing runs on its own and reports back with a completion email. Anything the AI employee can't match with certainty goes to the office team and gets trained afterwards. The completion email after every run is the source from which such process metrics can be drawn. Both of the firm's processes are at Autonomy Level 2, Autopilot, as of September 2026.
In Kevin's own business, 14 AI employees work, each with a personnel file, a fixed role and clear limits. An AI manager reviews their work; that role is still on trial. One example of a metric with a time frame and a comparison: kevinwelter.com had 1,942 clicks from Google Search in 28 days, from Sep 2 to Sep 29, 2026; in the 28 days before, it was 88. The site is looked after by an AI SEO employee.
That's a number with a time frame, a comparison and a responsible role. To be a metric in the sense of this article, it still needs a threshold and a consequence.
Runs without me
Every two weeks, one process that runs without the owner: what came in, what the AI did, where a person decided. Two minutes to read.
Worksheet
Process Potential Check
With five questions and a first process list to fill in, you find the process that could run on its own first. Comes as a PDF by email once you confirm your sign-up.
Strategy session
In a 20-minute strategy session, you tell us where things get stuck. We'll tell you honestly whether an AI employee is a good fit for you.
Book a strategy sessionFrequently asked questions
How many metrics belong on a KPI dashboard?
As many as fit on one screen and that someone acts on. For a business with a team, that's usually two or three per area: money, customers, workflows. That's our recommendation, not a standard.
What software do I need for a KPI dashboard?
A spreadsheet to start. A dedicated tool only pays off once several sources need to be connected regularly. What matters is that the numbers get into the dashboard without you.
How often should I look at the metrics?
That depends on the metric. For processes on Autopilot, process metrics arrive with every run. Money and customer metrics are often fine weekly or monthly. Once a month, it's worth looking at all processes together.
What's the difference between business metrics and KPIs?
Business metrics are all the numbers that say something about your business. KPIs are the few of them you actually steer by. Every business has many metrics, but only a handful belong on the dashboard. The test: if this number changes, does anything change about what you or your team do? If not, it stays in the archive.
Can an AI employee fill in the dashboard?
It can deliver the process metrics, because they're in its reports, and it can gather numbers from other systems if that workflow is written down. What follows from the numbers is decided by a person, especially when money is involved.
Read more
Who checks when several AI employees report back? Kevin covers that question in Does Your Business Need an AI Lead Role? And why metrics can also help when selling a business is covered in Selling a small business that depends on you: how to make it ready to sell.
Your next step
A good KPI dashboard fills itself from processes that report, not from your Monday morning. On the call, we look at whether Owner Autonomy is a fit for your business and which process is worth starting with. That's also where we go over how it works and what it costs.