Almost every Salesforce org that is more than a few years old ends up here. Someone added a field for a campaign in 2021. An admin left and took the knowledge of why that validation rule exists with them. A consultancy built automation on top of other automation. Nothing is broken enough to fix, so it never gets fixed, and meanwhile the team quietly stops trusting the reports.
Signs this is you
- Several fields mean almost the same thing and different teams use different ones
- Reports disagree with each other and people keep a spreadsheet instead
- Duplicate accounts and contacts, and no agreement on which one is real
- Automation fires in an order nobody can explain, or stopped firing and nobody noticed
- New hires are trained on workarounds rather than on the process
- Every small change feels risky, so changes stop being made
What a cleanup covers
Data model
Which objects and fields are actually used, which are filled in for fewer than a handful of records, which duplicate each other. Unused fields are not harmless — they slow page loads, confuse users and make every report harder to build. The output is a clear list of what to keep, merge and retire, with the risk of each.
Duplicates and data quality
Matching rules and duplicate rules configured so new duplicates stop being created, then a merge plan for the existing ones. Deduplication without the prevention step means you do it again in a year.
Automation
An inventory of every flow, process, workflow rule, validation rule and trigger per object, and what order they run in. This is where most strange behaviour turns out to live. Legacy Workflow Rules and Process Builder get consolidated into Flow, which Salesforce now requires anyway.
Permissions and setup
Profiles and permission sets that accumulated by copying whoever joined before, page layouts with fields nobody uses, record types created for a process that no longer exists.
What you get
An audit document you can read without being technical: what is wrong, what it is costing you, what to do about it, in priority order with effort against impact. Then the work itself, in stages — the items with the highest ratio of relief to risk first, each tested in a sandbox before production. Nothing gets deleted because it looks unused; anything irreversible is confirmed with you first.
Why this is worth doing before anything else
Every other project costs more on a messy org. Integrations have to map around fields that mean three different things. Custom development has to defend against data that should not exist. AI features read from a data model nobody trusts and produce output nobody believes. A cleanup is usually the cheapest project you will run and the one that makes the next three cheaper.
How this works
A free 20-minute call, then the org health check at a fixed $690 — a full audit within five working days, so you know what you are dealing with before committing to any remediation. You can take the audit and do the work yourself. That is a legitimate outcome and I will hand it over cleanly if that is what you decide.
Common questions
How long does an org audit take?
For a small to mid-size org, typically a few days to produce the audit document. Remediation is scoped separately once you have seen what is in it, so you are never committing to unknown work.
Will you delete things we still need?
No. Nothing irreversible happens without your explicit sign-off, everything is tested in a sandbox first, and fields that look unused get checked against actual data before anything is proposed for removal.
Can we just get the audit and fix things ourselves?
Yes, and for teams with a capable admin that is often the right call. The audit is written to be actionable by someone else, not to create a dependency on me.