Salesforce Implementation Checklist for Growing Teams

Key takeaways

• Most failed Salesforce rollouts were designed around features, not around how your team sells. Map the sales process on paper before you configure anything.
• Keep phase one small: leads, accounts, contacts, opportunities and the two or three reports leadership actually reads.
• Adoption is part of the build. Plan training, a named owner and a 30-day review before go-live, not after.

Picture a 35-person sales team that has outgrown spreadsheets and a lightweight CRM. They’ve signed for Salesforce, and someone asks, “So what do we need to do before kickoff?”

That question decides most of the outcome. A Salesforce implementation usually fails on process, data and adoption, rarely on the software. This checklist walks through the stages we use, in order, with the mistakes to avoid at each one.

Start with the sales process, not the platform

Before anyone opens Setup, write down how a deal moves from first touch to closed-won. Use a whiteboard or a doc.

Capture the stages, who owns each handoff, what has to be true to move forward, and what you report on. If your reps disagree about what “qualified” means, Salesforce will not settle it for you. It will just store the disagreement in a picklist.

Stage 1: discovery and scope

Name one business owner and one technical owner. “The team” isn’t an owner.

Then agree on scope in writing:

• Which clouds you need on day one (usually Sales Cloud, sometimes Service Cloud).
• Which teams go live first.
• Which systems must connect: marketing automation, billing, support, email and calendar.
• What you will not do in phase one.

That last line protects the project. Every growing team has a wish list, and without a “not yet” list the build never ends.

Stage 2: design the data model

Keep it boring. Standard objects (leads, accounts, contacts, opportunities) cover most of what a growing team needs.

Custom objects are fine when the business really has something Salesforce doesn’t model. They’re expensive when they exist because someone didn’t want to learn the standard one.

Decide these early, because they’re painful to change later:

• How you define an account, and how parent-child hierarchies work.
• Whether you use leads and contacts separately, or convert quickly.
• Which fields are required, and at which stage.
• Record types and page layouts per team.
• Roles, profiles and permission sets, built around least access.

Fewer fields are better. A form with 60 fields gets filled with junk.

Stage 3: clean and migrate the data

If you’re coming from another CRM or spreadsheets, migration is half the project. Our CRM migration checklist covers it in detail, and the short version is: decide what to leave behind, dedupe before you move, and run at least two trial imports into a sandbox.

If you expect messy accounts, read our guide to Salesforce duplicate management before you load anything. Setting up matching rules after a bad import is much harder than before.

Stage 4: build automation carefully

Flow is powerful, which is exactly why orgs end up with a dozen overlapping automations nobody can explain.

Some rules that keep an org healthy:

• One automation per object per trigger where you can. Consolidate instead of stacking.
• Name every flow so a stranger understands it: object, trigger, purpose.
• Document each automation in one line: what it does and who asked for it.
• Build lead routing and assignment as its own piece, and test it with real examples. Our lead routing best practices post shows what to cover.

Here’s an illustration (not a client example): a team adds a flow to assign new leads by region, and a second flow later to assign by company size. Nobody removes the first. Now a lead gets reassigned twice, and the rep who called it first loses the record. That’s automation sprawl, and it starts in month three.

Stage 5: integrations and reporting

Connect only what you need on day one. Email and calendar sync, your marketing platform and billing are common first picks.

For reporting, start with the views leadership opens every week: pipeline by stage, pipeline by rep, forecast and lead conversion. Build those first, test them against a spreadsheet you trust, and add more later. Our pipeline management best practices post explains what a healthy pipeline report shows.

Stage 6: test, train, go live

Test in a sandbox with real users, not just admins. Give each tester a short script: create a lead, convert it, log a call, move an opportunity, run a report. Write down what breaks.

Train by role. Reps need to know where to log activity and update stages. Managers need dashboards and approvals. Ten minutes of “here’s what changes for you” beats a two-hour tour of every tab.

Common mistakes that sink Salesforce rollouts

Mistake	What happens	Better move
Copying the old CRM field for field	Old mess, new price	Rebuild around today’s process
Too many custom objects	Hard to report, hard to maintain	Use standard objects first
No admin after go-live	Automation sprawl in a year	Assign an owner or a managed service
Training after launch	Reps build their own workarounds	Train before, review at 30 days
Adding agents too soon	AI acts on unstable data	Stabilise the basics, then add Agentforce

On that last point: if Agentforce is on your roadmap, design for it now. Our CRM AI readiness checklist lists what agents need from your data and permissions, and it’s much easier to build that in than to retrofit it.

The bottom line

A good Salesforce implementation is mostly decisions made early: a clear process, a small phase one, clean data and an owner after go-live. The configuration is the easy part.

If you want a second opinion on scope or sequence, we’re happy to look at your setup. A free 30-minute consultation is the fastest way to find out what your first phase should be.