Asking “how much is a data platform” first does not get a useful answer, any more than asking “how much is a plant.” Until the scope is set, every quote is a guess. The dimensions below help you put budget in a range you can check, instead of being pulled by the cheapest wall screen or the most expensive platform.
| Type | Typical cycle | Where the money actually goes |
|---|---|---|
| Operations dashboard, phase one (3–5 KPIs, drill-down, T+1) | 6–10 weeks | Definition sign-off, extract reconciliation, permissions, turning off parallel Excel |
| Add alerts + a floor or store screen | Another 3–6 weeks | Thresholds, role subscriptions, station fit, master-data mapping |
| Multi-system subject warehouse / lightweight platform | 3–6 months | API contracts, scheduling, quality rules, audit—not another batch of charts |
These numbers are experience ranges, not a quote. Off-the-shelf BI can light a screen faster, but definitions and extracts still take people. Compare Off-the-shelf BI vs a custom dashboard: what actually differs. Mid-size companies do not need the third row as an entry ticket; see Do mid-size companies have to buy a data platform.
Why can two projects both called “build a dashboard” differ several times in price?
The gap is KPI count, source-system count, whether you drill to documents, whether you need alerts and a floor screen, how messy master data is, and who signs definitions. Hidden items: whether existing Excel can serve as a check, and whether anyone will support the APIs. Customer time on definitions often outlasts writing SQL. A decorative wall screen looks cheap; if you cannot open it at acceptance, you did nothing. See Buying BI as decoration is the same as doing nothing.
Can the cycle be squeezed to a two-week light-up?
What can light up is usually Excel poured in, or one table connected directly. Definitions, reconciliation, permissions, and drill-down do not get solid in two weeks. Fine for a tour; the weekly meeting still brings the old sheet back. Prefer three to five KPIs you can actually ask into. See Why an operations dashboard is not optional and If a number cannot open a document, what is the dashboard for.
Who writes the definitions? Will that slip the schedule?
It will. This is delay number one. In the proposal phase, decide who in finance and operations signs the same number. A project with no owner stalls in “let’s reconcile once more.” How one definition cuts argument: After one definition, weekly meetings argue less. When quality is poor, wash the key codes first; see Data quality is poor. Build the dashboard first, or clean first.
How do you control budget without shipping a decorative screen?
Phase one only builds KPIs that support the weekly meeting. Floor screens and a platform come on demand. Writing drill-down and turning off parallel Excel into acceptance is worth more than filling twenty empty charts. Standardize APIs first so the next chart is not quoted again; see Once APIs are standardized, the next chart is faster. The full path: From brief to launch: how a data project actually runs.
DaXi gives scope, phases, and an acceptance list after a KPI inventory—not an unverifiable lump sum first. To make the KPIs you must watch and the systems they live in explicit, submit on the Data service page. Rebuild or extend an old platform: Rebuild the old reporting platform, or extend it.
Count hidden cost, or a low price doubles later
Outside the quote: business interview time, hours checking against Excel, master-data merge, night-shift alert validation, licenses, and wall-screen hardware. If those are not in the plan, they become “why is it still not done.” Conversely, cutting the leadership globe and second-level real-time that nobody opens usually saves a development round without hurting decisions. Maintenance, failed-job handling, and definition revisions should be line items, so you do not discover in year two that nobody dares change a formula.
When comparing two vendors, ask them to quote the same KPI list and write whether definition docs, reconciliation, training, and turning off the old sheet are included. Comparing only totals compares who is willing to omit items. Phasing can prove the weekly meeting first, then decide on a platform. That is less risk than signing everything once.