A KPI dashboard for small businesses: what to track, where the data comes from and how to make it stick

In many small businesses, knowing how things are going means someone spends half a day a month gathering data: exporting sales from the CRM, payments from the accounting system, orders from the online store, pasting it all into a spreadsheet and building a summary. By the time it reaches management it is weeks old. And when someone asks “where does this number come from?”, nobody is quite sure.

A well-built dashboard fixes that: a few metrics that update themselves, each with a clear definition and a single source. This guide covers what to measure, where to get the data, which tool to choose and, above all, what makes a dashboard get used instead of forgotten.

What a dashboard is (and is not)

A dashboard is a small set of indicators that support decisions, updated with no manual work. It is not a forty-chart report or a copy of every screen in your CRM.

The test is simple: if an indicator moved a lot from one week to the next, would anyone do anything differently? If not, it does not belong. In a small business, five to eight indicators on the main view are usually enough, with detail available for whoever needs it.

The metrics people actually look at

There is no universal list, but these indicators turn up in almost every small-business dashboard:

• Sales: revenue for the period against target, new customers, value of open deals and quote conversion rate.
• Cash: outstanding receivables, overdue invoices and how old they are, expected balance over the coming weeks.
• Operations: orders in progress, actual delivery time, incidents or returns.
• Customers: open tickets, time to first response, repeat customers.
• Marketing: new contacts by channel and how many become customers.

Every indicator needs a definition sheet, even a one-line one. It is what stops arguments about what each number means:

Where the data comes from

A small business’s data is usually spread across the CRM, the ERP or accounting system, the bank, the online store, web analytics and a spreadsheet or two. There are three ways to bring it together, from simplest to most involved:

1. Direct connection. The dashboard tool connects to each source through a connector. Fastest when there are few sources and connectors exist for them.
2. Workflows that load data into an intermediate store. An automation (with n8n, Make or similar) pulls the data from each tool every day, cleans it and puts it in a spreadsheet or a simple database, and the dashboard reads from there. The most common option when a source has no connector or data from several sources has to be combined.
3. A data warehouse. With high volume, many sources or a need for detailed history, data is centralised in a database built for analysis. Most small businesses do not need one to start with.

Start with the simplest option that works. Moving from the first to the second, or the second to the third, makes sense once the dashboard is in use and outgrows it. If you need to connect tools that do not talk to each other, see integrations and custom software.

Data Studio (Looker Studio), Power BI or Metabase

There are many dashboard tools; these three cover most small-business cases:

• Data Studio, which Google renamed back in April 2026 after a few years as Looker Studio, keeps a free version and a paid Pro version per user. It is the fastest way to start if your data already lives in Google tools.
• Power BI lets you build reports for free with Power BI Desktop, but sharing them with your team needs Pro or Premium Per User licences, or a Fabric capacity of a certain size. Very strong if you work in Excel and Microsoft 365.
• Metabase has an open-source edition (AGPL licence) you can host on your own server with no licence fee, plus paid plans in its cloud. A good fit when the data is already in a database.

The tool matters less than what sits underneath: clean data and agreed definitions. A beautiful dashboard on inconsistent data just lets people argue faster.

What makes a dashboard get used

Most abandoned dashboards fail not because of the tool, but for one of these reasons:

• Nobody trusts the numbers. If the dashboard says one thing and the accounts another, people go back to their spreadsheets. Definitions must be written down and each figure must have a single source.
• The source data is messy. Duplicate customers in the CRM, deals left open for months or orders without dates spoil any indicator. Often the first job is tidying the CRM; that is part of our CRM setup and integration service.
• There is no moment to look at it. A dashboard that is not reviewed in a regular meeting stops being opened. Fifteen minutes a week with the owners of each indicator is enough.
• There is too much on it. Forty charts do not get read. Better a few headline indicators with detail one click away.
• It gets looked at, but it does not alert. For urgent things, like a large overdue invoice or a spike in incidents, an email or chat alert beats waiting for someone to open the dashboard.

Example (hypothetical). A services company with about twenty staff builds a dashboard with six indicators: revenue this month against target (accounting system), overdue receivables by age (accounting), open deals and their value (CRM), quotes sent and accepted (CRM), billable hours against available hours (timesheet spreadsheet) and open tickets (support tool). A nightly workflow loads the data into an intermediate spreadsheet, the dashboard reads from it and, every Monday, management reviews it in a fifteen-minute meeting. If a large invoice falls overdue, an alert goes out that same day.

Where AI comes in

AI does not replace a good dashboard, but it can add value on top of one:

• Automatic summaries. Every Monday, a short note emailed to management on what changed from last week and why it might have.
• Anomaly detection. Alerts when an indicator leaves its usual range, without setting every threshold by hand.
• Questions in plain language. Several tools let you ask “which customers have spent less this quarter?” and get an answer. Useful, but check the important figures: AI can misread an ambiguous question.

Where to start

1. Choose the five indicators that drive the most decisions and write their definition sheets.
2. Find which tool each one comes from and whether it has a connector or API.
3. Build a first version with the simplest data option that works.
4. Set a weekly review and add alerts for urgent things.
5. Extend it only once the first version is genuinely in use.

This is the kind of project we do in analytics and dashboards. If you would like us to look at your case, in the free 30-minute audit we go through what data you have, where the indicators would come from and what it would cost to build. Write to us from the contact page. And if you are still deciding what to automate first, start with which processes to automate first.

Frequently asked questions

How many KPIs should a dashboard have?

On the main view, five to eight is usually enough for a small business. Each should support a decision; the detail can sit one click away.

Do I need a data warehouse for a dashboard?

Usually not to start with. With few sources, connecting them directly or loading the data daily into a spreadsheet or simple database is enough. A warehouse makes sense with high volume or many sources.

How often should it update?

It depends on the indicator: for most, a daily update and a weekly review are enough. For urgent things, like a large overdue payment, an alert beats a real-time refresh.

Why do so many dashboards get abandoned?

Mostly because nobody trusts the numbers, the source data is messy, or there is no regular moment to review them. The tool is rarely the problem.