Business intelligence dashboards: from spreadsheets to one source of truth

A business intelligence dashboard is a screen that shows a business's key numbers from several systems, joined and refreshed automatically, so everyone reads the same figures. For a business that spends on advertising, the version worth building joins sales or bookings, the CRM, the ad platforms and finance, so every sale is tied to the pound that bought it and judged on gross profit. Without a data team, building on tools you already own is often cheaper than buying an enterprise platform; the options are compared below.

What a business intelligence dashboard is

Business intelligence is the work of collecting data from the systems a business runs on, joining it, and presenting it so people decide faster and argue less. The dashboard is the part you see. Behind it sit three layers that decide whether it can be trusted:

• Connections that pull data from each system on a schedule, not by export. For marketing data integration, that means the ad platforms, analytics and the CRM at minimum.
• A data model that joins the data: brands, locations, channels, customers, enquiries, sales and costs. For marketing data this is often called a marketing data warehouse, though for a mid-sized business it can be a modest database.
• Definitions everyone has agreed: what a sale is, which date it belongs to, how gross profit is calculated.

A marketing dashboard is one set of views on top of this layer (see marketing dashboard examples by brand, location and channel). The business intelligence dashboard is wider: it is where marketing, sales and finance numbers meet and have to agree.

What it should join: four sources, one source of truth

A single source of truth is not one piece of software. It means each number has one agreed definition and one place it comes from, and the sources are reconciled against each other. For a business spending £5,000 or more a month on advertising, four sources do most of the work.

Web analytics, Search Console and reviews add context, but these four decide whether the dashboard can answer the question a finance director asks: what did that spend return? The third snag, platforms each claiming the same sale, is the subject of cross-channel attribution.

The evidence that joined data pays is mostly funded by platforms or vendors, so read it with that in mind. Research by BCG, commissioned by Google, found that companies linking all their first-party data sources can generate 1.5 times the incremental revenue from a single ad placement, and double the improvement in cost efficiency, compared with companies with more limited integration (BCG, 2021, 67 brands across Europe, the Middle East and Africa). Independent UK work points the same way: Nesta found firms in the top quarter for online data use were 13% more productive than those in the bottom quarter, other things being equal, while warning the link may partly reflect better management (Nesta, 2014, 500 UK firms).

The other reason is trust. Only 52% of senior marketing leaders told Gartner they could prove marketing's value and get credit for it, and they named CFOs as the executives most doubtful of that value (Gartner, 2024). Liz Kistruck, CFO of Motorway, told Marketing Week that when cost goes up and revenue does not, "I don't care what charts you show me" (Marketing Week, 2024). A dashboard that does not reconcile to the P&L will not survive its first board meeting.

Spreadsheets versus a built view

Spreadsheets are not the enemy. They are flexible, everyone can use them, and for one brand with two or three sources reviewed monthly, a well-kept workbook is enough. The trouble starts when a business adds brands, sites, agencies and ad accounts, and the workbook becomes the place where systems are joined by hand.

A worked example. Say a services group with four branches spends £15,000 a month on Google and Meta through two agencies. Every Monday an operations manager exports spend from three ad accounts, enquiries from the CRM and invoiced sales from finance, then matches them on campaign names. It takes most of a morning. One agency counts a call button click as a conversion, the other counts form fills, so their cost per lead figures cannot be compared. Finance closes monthly, so gross profit arrives weeks after the spend. The group ends up setting budgets on platform figures because they are the only current ones.

A note on business intelligence dashboards with real-time reporting: real time matters for stock, staffing and payments. For marketing and finance decisions made weekly, a daily refresh is enough, not least because ad platforms can credit a conversion days after the click that led to it.

Excel vs Power BI

Both are Microsoft products and share the same data-shaping engine, Power Query, so the choice is less about features than about how the numbers will be used.

Refresh limits are from Microsoft Learn; prices are from Microsoft's UK pricing page, checked on 3 October 2026.

The honest point: Power BI is business intelligence dashboard software, not the joined data. Owning it does not stop a business running the week from spreadsheets if nobody has built the connections to Meta, the CRM and finance, or agreed the definitions. Data Studio (formerly Looker Studio) is in the same position: free, strong on Google's own sources, and reliant on third-party connectors for the rest. Pick the tool your team will actually open. Whether it is useful depends on the data model underneath.

Build versus buy without a data team

Most businesses spending £5,000 or more a month on advertising have no data engineer. There are four realistic routes.

1. Buy a reporting tool with connectors. Quick and inexpensive. Good for marketing data, usually weak on CRM outcomes and finance, and it models your business the way the vendor's template does.
2. Hire an analyst. The Robert Half 2026 UK Salary Guide puts a mid-level data analyst at £48,250 and a business intelligence analyst at £42,500 (national 50th percentile, salary only, before employer costs). You gain someone who knows the business, and you depend on one person.
3. Bring in a business intelligence consultant for a fixed-scope build. Connections, data model and views are built in your own accounts and handed over with documentation. It needs a named owner afterwards and a tight scope so it does not drift.
4. Buy an enterprise platform with a data warehouse. Right when there is a data team to run it; oversized for most mid-sized businesses.

Five questions usually settle it:

• How many systems hold the numbers, and do you have admin access to each?
• Does the answer need finance joined, or only marketing data?
• Do you already use Power BI or Data Studio?
• Who will own the dashboard in a year, and will they have the time?
• Are budget decisions made monthly, or every week?

Owning a data store has a quieter benefit: history. In a standard GA4 property, the detailed data behind explorations and funnel reports is kept for two or 14 months (Google Analytics Help). A marketing data warehouse keeps raw history for as long as you need it, which matters when you want to compare this year's peak season with the last two.

Orion's business intelligence build is one version of the third route. It joins sales or bookings, CRM, ads and finance in your own accounts and delivers group, brand, location and channel views, automated reporting, alerts, a daily self-check and conversion feeds to Google and Meta. If you already own Power BI or Looker, it builds on that. It costs £3,400 after a £1,500 diagnostic, £4,900 in all, one-off or spread over six months (see investment).

How to create a business intelligence dashboard

1. Start from decisions. Write down the five to ten decisions it must support: weekly budget moves, which site needs attention, what goes to the board.
2. Map the sources. Every system, who owns it, and admin access to each.
3. Agree definitions. Sale, enquiry, gross profit, the date a sale belongs to, and how a sale is credited to a channel.
4. Connect and model. Pull each source on a schedule and join on stable keys: order IDs, CRM source fields, campaign names that carry brand and location.
5. Reconcile. Spend against invoices, platform conversions against CRM sales, revenue against finance. Flag mismatches; never hide them.
6. Build views by role. A one-page group view for the board, detail for the people who move budgets.
7. Automate the routine. Weekly and monthly reports and alerts, so the dashboard gets used rather than visited (see marketing reporting that runs itself).

The fifth step is the one that makes the rest worth having. Without it, a business intelligence dashboard is just a faster way to show figures nobody trusts.