BI dashboards people open without being told to
Metabase or Looker Studio, wired to real data
A dashboard nobody opens is a wasted build, usually because it shows what was easy to connect, not what the business actually decides on. We build BI dashboards starting from the three or four numbers your team checks every week, then connect the data behind them properly.
What it is
A BI dashboard turns raw data, from a store, a CRM, an ad account, a warehouse, into a small set of numbers a person checks regularly to make a decision: how is this week’s revenue tracking, which channel is actually profitable, is a churn number creeping up. The tool, Metabase or Looker Studio most often, matters less than the discipline behind it: starting from the question someone actually asks, then building the query and the chart to answer it, instead of connecting every available data source and letting a chart library decide what looks interesting.
When you need it (and when you do not)
You need this once decisions are being made on numbers someone remembers roughly or pulls manually each time, “I think conversion was around 3% last month,” is a sign the dashboard does not exist yet, or exists and nobody trusts it enough to check. It is also worth doing once more than one person needs the same numbers and currently gets them differently, a founder’s spreadsheet, a marketer’s ad platform screen, and a bookkeeper’s export, none of which agree.
You do not need a custom BI build if your main source already has a decent built-in reporting view and only one person ever looks at it, Shopify’s own analytics or a CRM’s native reports cover that case without a separate dashboard. The tell that you have outgrown the native view is wanting to see two sources side by side, which almost nothing’s built-in reporting does well.
How we build it
We start with a short conversation about the handful of numbers that actually drive a decision, not a wish list of every metric that exists, and build the dashboard around those first. Looker Studio is the better fit when the main sources are Google and Meta ad accounts or Google Sheets, it is free and its native connectors to those platforms are solid. Metabase, self-hosted on your infrastructure, handles SQL-backed sources and more complex filtering and joins better, and is the right choice once a data warehouse sits underneath the dashboard. Filters are built to match how people actually think about the data, by date range, channel, product line, not just whatever fields happen to exist in the source table. Access control means a sales rep sees their own pipeline, not the whole company’s margin data, from day one rather than as a later fix.
What to watch
The most common way a dashboard fails is silently: a metric’s definition changes upstream, a source stops syncing, and the dashboard keeps showing a number, just a stale or wrong one, with no visual difference from a correct one. We build a last-updated timestamp and basic freshness checks into every dashboard specifically so a stale number is visible, not hidden. The other risk is dashboard sprawl, once building one is easy, everyone wants a bespoke view, which fragments the shared definition of each metric right back into the problem the dashboard was meant to solve. We recommend a small number of well-maintained dashboards over many narrow ones, and say so before building a sixth view nobody asked for.
We also agree with you upfront on who owns a dashboard once it ships. A dashboard with no clear owner tends to drift out of sync with the business question it was built to answer within a couple of quarters, long before anyone notices the numbers no longer mean what they used to.
Price and timeline
| Scope | Price | Timeline |
|---|---|---|
| Single dashboard, 1-2 sources | from $1,200 | 1 to 2 weeks |
| Dashboard set across teams, shared warehouse | from $3,000 | 2 to 3 weeks |
Related
Built as part of analytics and custom development. Usually sits on top of an ETL pipeline and data warehouse and pairs with product analytics setup for behavioral data. See it in practice in an AI analyst across two brands and the Thailand D2C store audit and rebuild. Tell us which numbers your team checks every week: get in touch.
FAQ
How much do BI dashboards cost?
A single dashboard connected to one or two sources starts at $1,200. A fuller set, sales, marketing and operations each with their own view, pulling from a shared warehouse, runs $2,500 to $5,000.
How long does it take?
1 to 3 weeks for most dashboards: agreeing on the metrics that matter, connecting the data, and a short iteration period once the team actually starts using it and asks for a filter or a view nobody thought of upfront.
Metabase or Looker Studio, which one?
Looker Studio is free and connects well to Google and Meta ad data and Google Sheets, a good fit for marketing-heavy reporting. Metabase is self-hosted, handles SQL-backed sources and more complex joins better, and is the better choice once a warehouse is involved.
Do we need a data warehouse first?
Not always. If you have one or two sources, the dashboard can connect to them directly. Once you are joining three or more sources with different definitions of the same metric, a small warehouse underneath the dashboard saves more time than it costs.
Who owns the dashboard?
You do. Metabase runs on your infrastructure or a managed instance in your name; Looker Studio lives in your own Google account. Either way, access and the underlying queries are yours to keep or modify.