These come from my work as the in-house architect for a national nonprofit running six Salesforce clouds. The organization stays unnamed. The work is mine.
Data Cloud
Duplicate donor profiles cut from 20% to 5%
Supporter data came from the core Salesforce org and several outside platforms, and about one in five profiles was a duplicate of someone already on file. I implemented Data Cloud, brought each source in as a data stream mapped to the standard data model, and designed the identity resolution match and reconciliation rules that decide when two records are the same person. Duplicate profiles fell from 20% to 5%, so fundraising and marketing now work from one view of each supporter.
Integration
A siloed donation platform, brought into Salesforce
Online gifts were processed in a separate donation platform, so Salesforce never had the full picture of who gave and when. I built a webhook integration that sends each gift to Salesforce as it happens, where Apex matches the donor to an existing record or creates one and records the gift in Nonprofit Cloud. Gifts now land in Salesforce linked to the right donor, and fundraising reports include online giving without manual imports.
Integration
Slack alerts from Salesforce, two ways
Staff needed to hear about important moments in Salesforce as they happened, without watching dashboards. I built two patterns. Platform Events, published from Apex and Flow, carry alerts that more than one system cares about: the transaction never waits on Slack, and a Slack outage can't roll back Salesforce work. For simple one-to-one alerts, an Apex HTTP callout posts straight to a Slack webhook through a Named Credential, keeping the URL out of code.
Data quality
Donor addresses fixed in the backlog and at the source
Addresses arrived from the online giving platform exactly as donors typed them, and the donor base was due for a change-of-address pass. I processed an NCOA/PCOA file into about 18,900 address updates, 3,300 do-not-mail suppressions, and 1,400 deceased flags, then added validation to the gift pipeline with an Apex Queueable and the Google Address Validation API. Existing records were corrected in bulk, and new addresses are standardized as they arrive.
Nonprofit Cloud
A household model that covers every donor
The org runs Nonprofit Cloud on Person Accounts, and households existed only where two people shared a last name and address, leaving about 168,000 household accounts with no Party Relationship Group behind them. I helped the team decide to give every person a household, corrected about 2,900 existing group records, loaded the missing 168,000, and designed an Apex matching service so accounts created by integrations join the right household automatically.
DevOps
Releases and sign-in that don't depend on one laptop
Six clouds, several inbound integrations, and three Salesforce releases a year, with deployments tied to stored credentials. I moved the CI/CD pipeline from stored sfdx auth URLs to JWT-based authentication in GitHub Actions, rolled out Okta single sign-on, and run a release impact review across every cloud ahead of each seasonal release. Deployments now run from source control with no long-lived credentials, and every release is reviewed before it lands.
