Embedded Dashboard Scoping (Onboarding)

Embedded Dashboard Scoping (Onboarding)

1. 01

   Inventory the client's existing data sources before proposing any dashboard

   List every system the client already pays for (product database, billing, support desk, marketing warehouse) and confirm which ones expose an API or a scheduled export. Luzmo ships 40+ native connectors including Snowflake, BigQuery, and PostgreSQL, so most stacks are covered, but a missing connector changes the build estimate.

2. 02

   Separate metrics the client's end users will self-serve from metrics only the agency will maintain

   Self-service metrics need clean dimensions and documented definitions; agency-maintained metrics can stay in a scheduled report. This split is what keeps the retainer from turning into an open-ended BI help desk.

3. 03

   Confirm the tenancy model and row-level access rules in writing

   Ask whether each client account sees only its own rows or a shared view, and who inside the client organization approves access changes. Multi-tenant security is a configuration decision, not a default, and getting it wrong after launch means a rebuild.

4. 04

   Cap the custom visualization count at a fixed number per phase

   Over-customization is the named margin risk in this category. Agree on a dashboard count and a chart-type palette up front, then treat anything beyond it as a change order with its own line item.

5. 05

   Decide where natural-language querying sits in the client experience

   Qrvey and Sisense both ship conversational querying, and clients will ask for it once they see it. Decide whether it is in scope for launch or a phase-two upsell, because it changes both the data model requirements and the support surface.

6. 06

   Test brand inheritance against the client's design system before the build sprint

   Reveal integrates at the SDK layer so dashboards inherit the host application's colors, fonts, and component behavior, while CSS theming and iFrame approaches do not. Confirm which integration path the client's front end can support before committing to a timeline.

7. 07

   Write the data refresh expectation into the statement of work

   State the refresh cadence in plain language (hourly, nightly, on demand) and name the person responsible for the upstream pipeline. Ambiguity here produces the support tickets that erode the margin on an embedded analytics retainer.