CRM playbook
How to Build a Sales Process Inside Your CRM
A CRM cannot invent a sales process. It can only make the process you already have visible, coachable, and measurable. Teams that start with software end up with stage names that sound like a textbook and behavior that looks like the week before. Start with how buyers actually buy. Then put that path into the tool.
Interview the last ten wins
Before you draw a funnel, write the last ten won deals as timelines. Who showed up first? What evidence made you believe budget existed? When did legal appear? What almost killed the deal? You will see a pattern that is more honest than a vendor template.
Do the same for five losses. Losses tell you which stages are fake. If deals jump from demo to closed without a commercial conversation, you do not have a process. You have luck and a board.
Turn moments into stages
Each stage is a moment the buyer has already passed, not a task you hope to do. Discovery is complete when the problem and impact are documented. Qualification is complete when the path to money and a decision is known. Solution fit is complete when they have seen their workflow, not a generic tour.
Write exit criteria in buyer language. “We completed our discovery checklist” is about you. “The economic buyer confirmed a quarter and a range” is about them. Coach to the second sentence.
Required evidence, not required bureaucracy
You may require a next step, an amount, and a close date to leave the first stage. You may require a named decision process to enter commit. You should not require twenty fields on create. Sellers will then create nothing, or they will type nonsense.
Attach a note template to stage changes: what we learned, who we met, what risk remains. That template is the playbook in disguise. New hires learn by reading real deals, not by memorizing a slide.
Cadence is the process
A weekly pipeline review with a fixed agenda is part of the process. So is a same-day recap after a demo. So is a rule that silent deals age out. Software reminders help, but only if the meeting still happens when the quarter is busy.
Managers should inspect stage quality, not just totals. Pull five deals at random. If exit criteria are missing, the process is not installed. Training is cheaper than another dashboard.
When the product or market changes
If you add a low-touch product, do not force it through enterprise stages. Give it a shorter path. If you enter a regulated industry, add a validation stage instead of pretending security review is “negotiation.” The CRM should change when the buyer journey changes, not when a new VP wants different labels.
If two experienced sellers cannot agree which stage a deal belongs in, you do not have a process disagreement. You have undefined words. Fix the words first.
30-day process install
- Document ten wins and five losses as timelines.
- Draft five to seven stages with buyer-side exit criteria.
- Socialize the draft with sellers and revise once.
- Rebuild the pipeline view to match the new stages.
- Add a short note template on stage change.
- Run two weekly reviews using only the new definitions.
- Inspect ten random deals for evidence quality.
- Publish a one-page playbook next to the CRM bookmark.
Frequently asked questions
Can we copy another company's process?
You can copy the idea of exit criteria. You cannot copy their stage names and expect their conversion rates.
Should marketing define top-of-funnel stages?
Marketing can own lead statuses. Sales stages start when a human is working a commercial path. Keep the handoff explicit.
How rigid should this be?
Rigid on definitions, flexible on tactics. The stage meaning should not change by rep. The talk track can.
What if we sell two very different products?
Two processes, two pipelines. One compromised path serves neither buyer.
Does a process slow down great sellers?
A light process saves them from archaeology and from surprises in forecast week. Heavy forms do slow them down. Cut fields.
Operating notes for installing the sales process
Write stage exit criteria as buyer evidence. “Discovery complete” is weaker than “problem, impact, decision process, and next meeting confirmed.” Evidence makes stage movement coachable and reduces the temptation to advance deals because a rep feels optimistic. Keep criteria short enough to use during a live pipeline review.
Test the process with recent wins and losses before rolling it out. Reconstruct five to ten deals and ask where each would have sat at important moments. If the model cannot describe real history without exceptions, revise it before training the team. A process should fit how customers buy while still forcing useful discipline.
Use required fields sparingly. Require information at the moment it becomes necessary, not when an opportunity is first created. Too many gates encourage fake values and destroy trust in reporting. Reserve hard requirements for data that changes routing, forecasting, approvals, or the next team’s ability to do its work.
Inspect stage aging every week during the first month. Long dwell time can reveal unclear criteria, missing enablement, or a stage that combines two different buyer events. Adjust definitions before adding automation. Once the process is stable, automate reminders and hygiene around it rather than using automation to compensate for a process nobody understands.