Power BI, Looker Studio, or something simpler?

Looker Studio is free and suits businesses whose data already lives in Google — Analytics and Sheets. Power BI is stronger at modelling and enterprise connections, and costs per user. For a small business with three sources, a well-built weekly report in WhatsApp often beats both, because it arrives instead of waiting.

How much data do you need for a dashboard to be worth it?

Not volume but regularity. A dashboard becomes useful once there is enough history to show a trend, usually at least three months. A business that started measuring two weeks ago gets a chart of noise, not direction.

How often does the dashboard refresh?

It depends on licence and source. Power BI Pro allows up to 8 scheduled refreshes a day, Premium up to 48. But the real constraint is usually the source API rather than the tool. For most management decisions, once a day is enough.

We built a dashboard and nobody opens it. Why?

Because it answers a question nobody asked. A dashboard gets opened when it is tied to a decision made on a known rhythm, such as which leads went untouched this week. A dashboard showing general business health gets opened twice and then forgotten.

BI Dashboard for Business: What Goes On It and What Stays Off

Why most dashboards stay closed

The common failure of a BI dashboard is not technical. It is that it is built to answer “how is the business doing”, which is not how anyone phrases a question before making a decision.

A manager does not open a dashboard to be impressed. They open it when they have something to decide: who to reinforce this week, which leads are falling through, whether the campaign launched a fortnight ago is paying for itself. A dashboard not tied to a decision that recurs on a known rhythm never enters the routine, and no amount of design polish rescues it.

The simplest test before building: who exactly will open this, on Tuesday morning, and what will they do differently as a result.

What goes on the screen

The rule that works is one screen, no scrolling, and every number tied to an action.

What goes on:

One or two outcome metrics: sales, new leads, enquiries closed
A process metric that explains the outcome: average response time, share of leads with no reply, stage-to-stage conversion
One breakdown that enables action: by channel, by rep, by service type
A comparison: against last week or against target. A number without context says nothing

What stays off:

Metrics nobody owns
Charts that are handsome and imply no decision
Breakdowns that matter once a quarter. Those belong in a separate report, not the daily screen

A business that insists on twelve metrics on one screen effectively gets zero, because the eye has nowhere to land.

Report or dashboard

Deciding this early saves money.

A report is pushed to you on a schedule and answers a known question. It wins when the metric is stable and the audience is non-technical. It arrives in WhatsApp or email and requires nobody to remember to log in.

A dashboard waits for you and allows exploration. It wins when the questions change and someone genuinely sits and digs.

Many businesses build a dashboard when they needed a report. The tell: everyone keeps asking for “a screenshot of the dashboard” in the WhatsApp group. If that is happening, build a report.

Choosing a tool

Looker Studio — free, connects well to Google Analytics, Sheets and Ads. Suits a business whose data already lives in Google’s world
Power BI — strong at modelling and enterprise sources, costs per user, and needs someone who can maintain it
A structured report sent over WhatsApp — not “BI” and frequently the better answer. Zero learning curve, zero chance it goes unopened

On freshness: Microsoft specifies up to 8 scheduled refreshes per day on Power BI Pro and up to 48 on Premium and Fabric capacity, with manual and API refreshes counting against the same quota. A detail worth knowing: if nobody views the report for two months, scheduled refresh stops on its own. That is a useful test of whether the dashboard is genuinely in use.

The definitions layer is the real work

The reason two people get two different numbers for the same question is almost always a definition, not a bug. Is an “active customer” someone who bought this year or this quarter. Is revenue counted on the invoice or on the payment received. What happens to credits and refunds.

These definitions need to live in one place the dashboard reads from, rather than being duplicated in every chart. Gartner predicts that by 2030 universal semantic layers will be treated as critical infrastructure alongside data platforms and cybersecurity. In a small business this is not a platform, it is one tidy definitions file, but the principle is identical.

How to start without wasting the budget

Pick one decision that recurs on a known rhythm
Write down the two numbers needed to make it
Build them in the cheapest tool that connects to the source
Run it for a month and see who opens it
Add a third metric only if someone asks

This order prevents the project small businesses commonly run: three weeks of building, an impressive screen, and zero visits in month two.

Sources

Configure scheduled refresh - Power BI — Microsoft Learn
Gartner Announces Top Predictions for Data and Analytics in 2026 — Gartner, March 2026

Bottom Line

Want to know which automation is right for your business? A free 30-minute diagnostic call, no commitment, we'll map out your processes together and honestly tell you if and how automation can help.