One Normalized Lead Intake for Every Source

The problem

Leads arrived from paid ads, the website, and referrals through different paths, each mapped a little differently. Some never made it into the CRM at all. Older deals carried no record of the campaign that produced them, so there was no way to connect ad spend to revenue.

What I built

• A single intake. Every source posts to one secured webhook, which normalizes the answers and creates or updates the CRM record with full field mapping.
• A reconciliation. I compared the ad partner’s lead list against the CRM, record by record, to find leads lost in transit, then replaced the path that dropped them.
• Restored attribution. Campaign, ad set, and ad were backfilled onto older deals wherever the data existed, and corrected where it was wrong.
• A write-back. The CRM record ID goes back to the website, so later form activity lands on the same person instead of creating a duplicate.
• Locked-down sensitive fields. Only the roles that need sensitive answers can see them.

How it works

The website form posts to a webhook. Zapier normalizes the payload and upserts the lead in Salesforce by a stable external ID, so a resubmission updates the record instead of duplicating it. A monitor posts every failed run to a Slack channel, so a broken path gets noticed in minutes, not weeks.

The reconciliation is the step most teams skip. Until you compare the source list with the CRM line by line, you do not know how many leads you paid for and never saw.

Stack

Website form webhooks, Zapier, Salesforce, Slack error alerts.

Where it stands

Built in August 2026 and running. Everything downstream depends on it: the scoring model, the lead queue, and campaign reporting all assume leads arrive complete and on time.