CRM implementations rarely fail technically. The data imports, the pipeline gets built, the integration works. They fail three weeks later, when the person who was excited about it gets busy and everyone quietly returns to their inbox and a spreadsheet.
A good onboarding plan is designed around that risk rather than around features.
Weeks 1–2: decide how you sell
Before anyone touches the software, write down the sales process you actually run. Not the aspirational one. What happens when an enquiry arrives, who owns it, what has to be true to move it forward, and what you need to know at each point.
This is the step people skip, and skipping it is why so many CRMs end up configured around the software’s defaults rather than the business. Our guide to defining pipeline stages covers how to do it so the result is useful.
Also decide now what “good data” means. Which fields are mandatory, who owns a record, and what happens to a deal that goes quiet.
Weeks 3–4: build the minimum
Set up the pipeline, the fields you agreed, and nothing else. Resist every optional feature.
Import contacts – after cleaning them, not before. Deduplicate, drop anyone you will never contact again, and fix the obvious errors. Importing a messy spreadsheet creates a messy CRM and erodes trust in it from day one.
Connect the website so enquiries arrive automatically with their source attached. If you do nothing else technical, do this one; it is what makes the CRM worth consulting later.
Weeks 5–6: get people using it
Train in short sessions on the specific things each person has to do, not a tour of the platform. A salesperson needs four actions: log a conversation, move a deal, set a follow-up, find a record.
Then make it the only place. If the weekly sales meeting is run from the CRM, people update it. If it is run from a spreadsheet someone compiled, the CRM is already dead – it just has not been announced yet.
Expect a fortnight of grumbling. This is normal and it passes, provided someone keeps insisting.
Weeks 7–9: automate the annoying parts
Only now, once you can see how people actually work, add automation. The usual first candidates: acknowledging enquiries, assigning owners, reminding someone when a deal has gone quiet, and notifying the right person when something important happens.
Each automation should remove a task someone finds irritating. Automation that adds steps gets worked around.
Weeks 10–12: reporting, and the honest review
Build the three or four reports that answer questions you will ask monthly – enquiries by source, conversion by stage, average deal size, time to close.
Then review honestly. Is everyone actually using it? Is the data good enough to make decisions from? If not, the answer is almost never more training or more features. It is usually that the process does not match how people work, and the process needs to change.
Why rollouts stall
No internal owner. Someone must be responsible for the CRM being accurate. Without a name against it, it degrades.
Too much at once. Pipeline, marketing automation, service tickets and reporting in one go means nothing is bedded in before the next thing arrives.
Configured for management, not for sellers. If the CRM only exists to produce reports for the board, the people entering data get nothing and will stop.
Leadership not using it. If the MD asks for an update in a meeting rather than looking at the pipeline, everyone learns what really matters.
Frequently asked questions
Should we import all our historical data?
Usually not all of it. Bring across live opportunities, current clients and anyone contacted in the last year or two. Archive the rest somewhere retrievable. Old data makes the CRM feel unreliable without adding much.
Do we need paid onboarding from the vendor?
It helps with complex migrations and integrations. For a straightforward setup, the bigger constraint is usually deciding your own process — which no vendor can do for you.
How do we get a reluctant sales team to adopt it?
Make it genuinely easier than what they do now. If updating a deal takes thirty seconds and saves them writing a weekly report, adoption follows. If it is pure admin overhead, it will not.
When should we expect a return?
Better follow-up shows within weeks. Reliable reporting takes a quarter, because you need enough clean data to trust it.
Set it up around the business, not the software
We implement CRMs the way described here — process first, minimum build, adoption before automation. See our lead generation and sales service, or get in touch.
