How to Build a B2B Sales Process From Scratch

Most companies think they have a sales process. What they actually have is a CRM with a dropdown of stages someone invented on a Tuesday: "Contacted," "Interested," "Negotiation," "Closing." Those stages describe how the rep feels, not what the buyer did. That is why the forecast is always wrong and every deal is somehow "90% there" until it vanishes.

A real process is built around buyer actions, not rep optimism. Here is the skeleton.

Start with the ICP, not the stages

Before you draw a single stage, define who you are actually selling to. Not "mid-market SaaS." The specific buyer: their role, the trigger that makes them act now, the pain that justifies spending money, and who else sits at the table. If you cannot describe the person and the moment, no process will save you, because you will fill the pipeline with deals that were never going to close. Getting this right is the core of any go-to-market strategy worth the name.

Define stages by buyer actions with exit criteria

This is the part almost everyone gets wrong. A stage is not a feeling. It is a verifiable thing the buyer has done. Each stage needs an exit criterion: the specific event that lets a deal advance. For example:

• Qualified: the buyer confirmed a real pain, a rough budget, and a timeline. Not "seemed keen."
• Evaluating: the buyer agreed to a next step that costs them something, a demo with their team, a trial, access to data.
• Selecting: you know the decision criteria, the other options, and who signs.
• Committed: verbal yes and the paperwork is moving.

If a deal has not met the exit criterion, it does not move, no matter how good the last call felt. This one rule kills most of the fantasy in a pipeline.

Build a forecast you can actually trust

Once stages mean something, the forecast stops being a guess. Because each stage is defined by buyer behavior, you can measure the real conversion rate between stages and apply it honestly. A deal in "Evaluating" is not worth its full value times optimism. It is worth its value times the actual historical rate that "Evaluating" deals close.

Then separate what you commit to from your best case. Commit is what you would bet the quarter on. Best case is upside if things break your way. Blending the two is how founders lie to their board and themselves.

Build these three layers, ICP, buyer-based stages with exit criteria, and an honest forecast, and you have a process instead of a wish. If you want it built and installed correctly the first time, that is exactly the work a fractional CRO does before you scale spend behind it.