Green Light Offer wanted dashboards they could trust. We fixed the data first.

When a sales floor says its dashboards are wrong, the tempting move is to rebuild the dashboards. We did the opposite. A dashboard is only ever as honest as the fields underneath it, and three of Green Light Offer's most important fields were structurally broken: the original lead owner was being overwritten on every reassignment, a twenty-second dial counted the same as a real conversation, and there was no way to tell how complete a deal record was. Rebuild the charts on top of that and you get prettier lies. So the first phase was entirely below the surface.

Lead ownership was the subtle one. Because leads are created under intake users' accounts, the "owner" field pointed at whoever typed the lead in, and every round-robin reassignment wiped the last real value. We built a before-save flow that stamps the first genuine acquisitions agent assigned to a lead, skips intake users and queues entirely, and never overwrites itself afterward. It passed in the sandbox and failed in production, which turned out to be its own lesson: a second, older active flow was firing first and blocking ours. Finding that conflict was as much of the job as writing the flow.

The completeness engine was built to survive change. Rather than a static formula that breaks every time the team adds a field, we drove it with custom metadata, so an administrator adds a field to the calculation by creating a record, not by calling us. The opportunity version runs as a bulk-safe trigger that only recalculates when a field actually flips between blank and filled, so a value merely changing from six to eight costs nothing. Meanwhile the contact-qualification rule was reduced to one hard number, 120 seconds, pointed at the call-duration field the client's phone system had genuinely been writing to for years rather than the Salesforce default.

Only then did we rebuild the reporting layer: twenty reports and department dashboards for acquisitions, dispositions, and title and closing, a consolidated company view, and a role-aware home page where each user sees their own numbers and their department's. Because the data underneath was finally true, the KPIs the team had asked for for months - Speed to Lead, Meaningful Contact Rate, Projected versus Actual Profit - were numbers leadership could act on rather than argue about.