Which channels produced the opportunities you closed?

Salesforce can attribute properly, and in a well-run org it does. The mechanism is campaign influence: every touch has to exist as a campaign member on the contact, and the model then distributes credit across them. When the data is complete the output is good.

The data is rarely complete, and the incompleteness is not random. Digital touches land automatically if the integration is configured. A conference conversation, a partner introduction, an inbound call to a regional office, a customer referral — these arrive as a person, not as a campaign member, and somebody has to remember to load them. The channels that get left out are systematically the offline ones.

So the report tells you about the channels that were easy to instrument, and is silent on the ones that were not. That silence reads as absence, and budget follows the report.

Instead of modeling how much credit each touch deserves, this asks whether the person on a closed-won opportunity appears in a given source list. It is a coarser question, and that is the point: it has a definite answer and the answer can be checked by opening two files.

High-confidence matches — an exact phone or email after normalization — count automatically. A phone that agrees only once its country code is set aside waits for a person to accept or reject rather than being counted. And opportunities that appear in no source list at all are reported as unattributed, however large that number turns out to be.

Because the whole thing runs on exports, it needs no admin project, no new object, no field on the opportunity and no change to how the sales team works.

Large Salesforce orgs usually run several businesses through one instance: new business and renewals, direct and channel, several product lines with wildly different deal sizes. A single cost per closed-won opportunity across all of them is an average nobody owns.

A record type or business unit column splits the report along the lines the company is actually managed on. It is the difference between a chart that gets shown once and a chart a business unit leader asks for monthly.

Channel-led revenue is notoriously hard to attribute, because the partner owns the relationship and the end customer may never touch your website. Deal registration gives you a record of the partner, not of the marketing that produced the end customer.

Adding partner registrations to the lead file, under the partner program's own source name, puts them on the same revenue axis as direct demand generation: what the channel program produced against what paid search produced, on a single scale.

In an organization where the marketing number is contested, the important property is not accuracy — it is auditability. A figure somebody can challenge and you can trace to two rows in two exports is worth more than a more sophisticated figure nobody can open.

Every match here is inspectable and reversible. High-confidence matches on a normalized phone or email count automatically; a phone that matches only without its country code is queued for a person to accept or reject, and that decision is recorded. Growth and Enterprise carry more than one seat, so finance and marketing can look at the same queue.

That is a deliberately unglamorous design. It is also the reason the number survives the meeting where somebody asks where it came from.