Cloudastick Systems

Admin Chaos Playbook

Insights & Updates from the Cloudastick Team

Your follow-up guide to building the kind of Salesforce org that guarantees sleepless nights, finger-pointing, and dashboards that cry for help.

If Part 1 showed you how to sabotage your CRM through data structure, Part 2 takes the chaos a step further. We now enter the realm of administration, access, maintenance—or the lack thereof.

These are the kinds of decisions that don’t just annoy users…

They create legacy problems so legendary that future admins will write stories about them.

Let’s continue our journey of “What You Should Absolutely Never Do If You Want to Love Your CRM.”


7. Give everyone System Admin permissions

The Wild West Permissions Model

Why thoughtfully design roles and profiles when you can give everybody God-Mode?

Let every user, intern, and visiting consultant wield the power to delete objects on a whim.

What could possibly go wrong?

Real Example of Chaos

  • A sales rep accidentally deactivates key validation rules “because they were annoying.”
  • Someone edits the Opportunity Stage picklist in production at 4 PM on a quarter-closing day.
  • Someone clones the standard “Admin” profile… and names it “Sales Team.”

What You Should Do Instead

Follow least-privilege design.

Define profiles, permission sets, and permission set groups with surgical precision.

Admin access should be earned—not freely distributed like stickers.


8. Have the whole team share a single login

The Single Shared Account Nightmare

Why bother with user accounts when one generic login can serve all departments?

Audit trails? Compliance? Accountability?

Trivial concerns—until something breaks and nobody knows who modified what.

Real Example of Chaos

  • A workflow rule disappears. Five team members swear “it wasn’t me.”
  • Security review fails instantly because the login history shows 27 users “logging in” from different cities at once.
  • Password changes become group events.

What You Should Do Instead

Every user needs their own account.

Always. No exceptions.

Your org (and your sanity) depends on it.


9. Avoid the Sandbox at all costs. Ship everything directly in Production.

The Production-Only Development Method

Testing? Who needs testing?

Why validate solutions in a safe environment when you can deploy directly into the live system and see what breaks in real time?

Real Example of Chaos

  • A trigger loops endlessly and locks every Opportunity.
  • You try to roll back—but your rollback is also broken because it was never tested.
  • Management wants to know why the system is down. You point at the mirror.

What You Should Do Instead

Use sandboxes.

Use them deeply.

Use them often.

Treat production like a museum exhibit: look, don’t touch.


10. Never clean it up now—always promise to do it later

The Perpetual Deferral of Maintenance

Technical debt is like actual debt: the longer you ignore it, the more painful it becomes.

But sure—skip data cleansing, rule audits, and archival efforts.

Let future you handle it… or let future admins curse your name.

Real Example of Chaos

  • 98,000 contacts with missing emails.
  • 17 different lead sources, including “Other,” “Others,” “Other Source,” and “othr.”
  • A validation rule last updated “5 years ago” that no one remembers writing.

What You Should Do Instead

Schedule periodic cleansing, implement data stewardship processes, and review automation annually.

Maintenance isn’t optional; it’s survival.


11. Don’t build any CRM health dashboards

The Ignorance Is Bliss Approach

If you never track data quality issues…

Then those issues technically don’t exist, right?

If you don’t measure orphaned records, missing owners, broken automation, or stalled approvals…

Then users can’t blame you for knowing about them.

Real Example of Chaos

  • Thousands of Opportunities stuck in “Closed Lost” with no closed reason.
  • Accounts without owners… for years.
  • A dashboard showing 0 problems—because there is no dashboard.

What You Should Do Instead

Build data health dashboards.

Monitor field completeness, record ownership, activity logging, and process exceptions.

Trust me: you want to see problems before your CEO does.


12. Ignore user training and documentation

The “They’ll Figure It Out” Philosophy

Why train users when you can assume they’ll magically understand Salesforce?

Why create guides when tribal knowledge is so much more… mysterious?

Real Example of Chaos

  • Users create Opportunities without Accounts.
  • Someone asks, “What’s a report type?”
  • A team member tries to delete a dashboard because they thought it was “just clutter.”

What You Should Do Instead

Train your users. Document your processes. Create accessible learning paths.

CRM adoption depends on understanding—not hope.