Sf Campaigns

Salesforce Campaign Analysis

Marketing analyst for Salesforce campaigns. Turn Campaign and CampaignMember data into ranked performance comparisons and invest/pause recommendations — with the math shown.

Dispatch

Initialize the org connection first (org_init convention — see references/execution-modes.md) — unless running in Demo mode (below), which has no org to connect to.

Portfolio Report

1. Query active campaigns:

SELECT Id, Name, Type, Status, StartDate, EndDate, NumberOfLeads,
       NumberOfContacts, NumberOfResponses, NumberOfConvertedLeads,
       NumberOfOpportunities, NumberOfWonOpportunities,
       AmountAllOpportunities, AmountWonOpportunities,
       BudgetedCost, ActualCost
FROM Campaign WHERE IsActive = true ORDER BY StartDate DESC LIMIT 50

2. Compute per campaign (guard every division against null/zero):

3. Present ranked by ROI (or pipeline when costs are unrecorded — say which and why). Pipeline-based ROI flatters campaigns that never close — when any ranked campaign has zero wins, also compute won-ROI ((AmountWonOpportunities − ActualCost) / ActualCost) and show both columns so the flattery is visible. Then close with recommendations: increase / maintain / pause / restructure per campaign, grounded in the numbers.

Campaigns with no ActualCost make ROI meaningless — flag them as a data hygiene finding rather than silently ranking them last.

Single-Campaign Deep Dive

Pull member detail and work the funnel:

SELECT Id, LeadOrContactId, Status, HasResponded, FirstRespondedDate,
       Lead.Name, Lead.Company, Lead.LeadSource, Lead.IsConverted,
       Contact.Name, Contact.Account.Name
FROM CampaignMember WHERE CampaignId = '<id>' LIMIT 2000

Report: funnel (added → responded → converted → opportunity → won), response timing distribution (FirstRespondedDate − campaign StartDate), member-status breakdown, and the influenced-opportunity list (SELECT ... FROM Opportunity WHERE CampaignId = '<id>' plus OpportunityContactRole paths when campaign influence matters). Note the attribution model in use — primary campaign source vs influence — because the same campaign can look brilliant under one and invisible under the other.

Comparison & Ranking

Same metrics across the named campaigns (or by Type across the portfolio), plus pattern analysis: which LeadSource values convert best, seasonal/timing effects when StartDates span quarters, and campaign-type benchmarks (webinar vs event vs listing vs nurture). Rank on the metric that matches the user's goal — pipeline efficiency (ROI) and volume (members) rarely agree.

Output

Tables for rankings and funnels, per the sf-audit report template's §7 instincts — and when the user wants a document deliverable, follow the sf-audit skill's report-template.md (from this skill: ../sf-audit/references/report-template.md) §7–8 (visualizations, single-file animated HTML) rather than dumping tables into prose.

Cross-skill handoffs

• Query tuning or bulk exports → sf-data
• Campaign-sourced leads needing enrichment before scoring → sf-leads
• Full org marketing-data hygiene → sf-audit (data quality section)
• Fixing the campaign/member/lead records themselves → sf-records; the shared record-writing canon it applies is ../../shared/standards/record-data-quality.md

Custom-field discernment (customized orgs)

Standard fields aren't always where the truth lives. Before computing metrics in an unfamiliar org:

1. Describe Campaign (and Opportunity when computing pipeline) and list populated custom fields whose names shadow the standard metrics — Actual_Spend__c beside an empty ActualCost, Total_Contract_Value__c beside a stale Amount, custom member-status fields beside Status.
2. Detect shadowing with a sampling query: compare populated-rate and recency of the standard field vs the candidate (SELECT COUNT(Id), COUNT(ActualCost), COUNT(Actual_Spend__c) FROM Campaign).
3. When a shadow candidate wins on population, ask the user which field is authoritative before computing — a confident ROI from the wrong field is worse than a question. Record the choice in the report's methodology note so the numbers are auditable.

Demo mode (no org)

The repo ships synthetic data at sample-data/ (campaigns.csv, campaign_members.csv). When no Salesforce MCP server is connected — or the user asks for a demo — run the same workflows against those CSVs and say so in the output. Skip org_init in demo mode — there is no org to initialize. The dataset contains deliberate hygiene findings (missing costs, zero-win campaigns) worth surfacing.