Done-for-you · The Build Guide
Build it, click by click
The literal handoff for your GTM engineer, RevOps admin, or IT team. Set the four tools up in order — Salesforce first (the foundation), then Gong, Momentum, and Slack. Every use case is a numbered recipe with exactly what to click. Validate each step against your own instance and build in a sandbox first.
Foundation first — the source of truth. Salesforce holds the fields every other tool writes into, and the integration identity everything authenticates as. Build these before connecting anything. Role: System Administrator. Time: ~45–60 min.
How to build the reporting layer
You already defined the metrics and the data architecture. This is how you actually stand up the reporting on top of them — in the order a RevOps engineer would build it, so the numbers reconcile and people trust them.
The reporting build sequence
Write the metric dictionary first
Before a single query, define every metric in one place — name, exact formula, grain, source object, and the single authoritative system that owns it. Reporting dies when “pipeline” means created, open, and weighted on three different dashboards. One definition, one owner, one system per metric.
Instrument the source
The data has to exist before you can report on it. In the CRM, lock stage definitions and exit criteria, add required-for-stage validation, and structure the fields (picklists, not free text). Turn on activity capture (Gong / Clari) so calls and engagement land as data. If a metric has no field behind it, build the field now.
Pipe to the warehouse and model it
Sync CRM, activity, product, and support sources into the warehouse (Snowflake / BigQuery) on a schedule. Model it in a transformation / semantic layer (dbt) into clean marts — one row-grain per fact: an opportunity mart, an activity mart, and a pipeline-snapshot mart. Formulas computed once, read everywhere.
Build dashboards where they belong
Operational / real-time → native CRM reporting (Salesforce), where reps act inside the workflow. Analytical / trended → BI (Tableau / Looker / Sigma) on the modeled marts. The CRM can't trend what it overwrites — so anything historical or cross-org lives in BI.
Reverse-ETL the answers back
Computed scores and segments are useless sitting in the warehouse. Push them back (Census / Hightouch) onto the Salesforce records — health tier, fit score, segment — so reps see them in context and the loops act on them. The analysis becomes a field; the field drives an action.
Govern it
Every dashboard gets a named owner, a refresh cadence / SLA with a visible “last refreshed” stamp, and a data-quality check that runs before the number is trusted — row counts, null checks, and a CRM-to-BI reconciliation on the headline metric. A number with no owner and no freshness SLA is a rumor.
The dashboards to build
The Monday Cockpit
Per rep/team · weekly · sales leaders · CRM + Slack
Forecast by deal health, the three coaching moments with clips, and the one funnel stage leaking this week. Lives where the action is; sources health + forecast from the modeled layer.
The Forecast Roll-Up
Per opp → mgr → region → board · weekly · BI
Snapshotted weekly so you see the call change. The load-bearing column is the commit-vs-evidence delta, surfaced as a manager coaching prompt. New-business + renewal commit roll into one number.
The VP Scorecards
Per role/metric · weekly refresh · BI
One scorecard per owner (Sales, RevOps, CS, Marketing, Partners), each metric red/amber/green against its agreed target — plus the leading indicators that predict it 30–60 days out.
Pipeline Velocity & Coverage
Per segment · weekly · off the snapshot mart · BI
Velocity (opps × win rate × deal size ÷ cycle), coverage vs quota (3–5×), and stage aging — trended. Requires the pipeline-snapshot mart; movement is the whole point.
Attribution & Segment
Per opp + source/segment · monthly · BI
State the model (W-shaped, defined lookback) before you present it. Marketing-sourced reported separately so it's never double-counted; everything split by segment, channel, region.
Don't forget the snapshots — the one thing teams can't get back
Salesforce stores the current value of every opportunity. Change a close date or amount and the old value is gone. The day you decide to measure velocity, slippage, or week-over-week forecast change, you find you have no history — and you can't backfill it. Capture it from day one: a scheduled snapshot job in the warehouse (one row per open opp per day/week), Salesforce historical trending, or Clari's time-series auto-capture. Turn it on before you need it.
Build gotchas · acceptance criteria
One authoritative source per metric — no number computed two ways in two places.
Every metric has a written definition — name, formula, grain, source in the dictionary.
Numbers reconcile across CRM and BI — a reconciliation check runs before anyone trusts it.
Every dashboard has an owner and a refresh SLA — a named human and a visible timestamp.
A metric nobody looks at gets cut — judged by the decisions it changes, not the charts it holds.