Try productized BI first
Before building custom, exhaust the productized options. They're cheap, fast, and usually good enough.
- Looker Studio (formerly Data Studio) — free, integrates with Google Analytics, Google Ads, BigQuery, Sheets. Best for marketing dashboards.
- Power BI — $14/user/month, integrates well with Microsoft stack. Best for finance/ops dashboards in MS-Office shops.
- Tableau — $75/user/month, most powerful, steepest learning curve.
- Domo / Sisense — enterprise pricing, all-in-one BI.
- Most small businesses cover 80% of their reporting needs with Looker Studio + a CRM/accounting integration.
Source data lives across systems BI can't integrate
BI tools have connectors for common SaaS (Google Ads, Stripe, HubSpot, QuickBooks). When your data lives in less-common systems — niche industry software, proprietary databases, legacy on-premise systems, browser-only portals — you have three options:
- ETL middleware (Fivetran, Stitch, Airbyte) — may or may not have connectors for your source.
- Manual export-import workflow — fragile, time-consuming, error-prone.
- Custom integration that pulls from each source and lands data where you can visualize it.
HG Oil Holdings case
HG Oil's inventory data lived across systems that no productized BI tool integrated with cleanly. The custom inventory dashboard didn't just visualize data — it built the data pipeline that made the visualization possible. That's why the productized path failed and custom paid off.
Calculation logic is too business-specific
BI tools handle SUM, AVG, percentages, time-series easily. They struggle with:
- Complex business rules ("profitability per load must account for tolls accrued by deadhead miles from the previous load").
- Calculations that span multiple data sources with non-trivial joins.
- Calculations that change based on context (driver settlement varies by load type, lane, season).
- Real-time forecast models that update as inputs change.
Real-time freshness matters
Most productized BI refreshes nightly (or on a schedule). For some operational use cases — dispatch, inventory, oil/gas wellpads, manufacturing floor — that's not fast enough. Custom dashboards backed by real-time databases (or change-data-capture pipelines) update within seconds of source-system changes.
Important caveat: real-time isn't always better. If decisions get made daily or weekly, nightly refresh is fine and cheaper.
Operational actions launch from the dashboard
BI tools are read-only by design. They show you what's happening; you go elsewhere to act on it. Custom dashboards can include action buttons — "approve this purchase order," "send this notification to the driver," "flag this customer for follow-up" — without context-switching to another system. That speeds up operations meaningfully.
What custom dashboards cost
A useful framing:
- Small custom dashboard (single data source, basic charts, one-off): $4,000-$8,000.
- Mid-size custom dashboard (multiple sources, real-time, some action buttons): $10,000-$25,000.
- Full custom inventory/operations dashboard (HG Oil scope): $35,000-$60,000.
- Ongoing maintenance: $200-$1,000/month depending on data-source volatility and feature roadmap.
