B2B sales pipeline: how to structure it properly?

What is a sales pipeline?

A sales pipeline allows you to track sales opportunities between their creation and their closing.

Each opportunity occupies a stage corresponding to its level of progress: scheduled diagnosis, meeting held, proposal sent, decision pending, or any other stage that is genuinely relevant to your process.

The goal is not just to visualize potential sales. A good pipeline should allow you to quickly answer several questions:

• which opportunities are actually active?

• where does each deal stand?

• what is the next action?

• which opportunities are stalling?

• what potential revenue can reasonably be expected?

• where in the sales process are the most opportunities being lost?

However, we must distinguish between several concepts that are often mixed up.

This distinction matters particularly when Marketing and Sales use the same CRM. A contact can, for example, become a MQL and then a SQL without necessarily becoming an opportunity immediately. Our MQL vs SQL guide details this handoff logic between Marketing and Sales precisely.

When should an opportunity enter the pipeline?

Not all contacts present in the CRM should appear in the sales pipeline.

Let's imagine a company with:

300 contacts

↓

40 leads actually matching its target

↓

15 engaged sales conversations

↓

8 projects concrete enough to be tracked

It can make perfect sense to have only 8 opportunities in the pipeline.

In some organizations, an opportunity is created as soon as a sales rep agrees to work a lead. In others, it appears only when a sufficiently concrete need, context, and project have been confirmed.

There is no universal threshold.

Rather, the important rule is to be able to unambiguously answer this question:

What event triggers the creation of an opportunity in our CRM?

Without this rule, the pipeline often ends up filled with prospects met once, hypothetical requests, and deals that never really existed.

B2B lead qualification comes into play beforehand to determine which prospects deserve active handling and at what point a project can become a real opportunity.

How to define the right stages of your sales pipeline?

A common mistake is to open your CRM, look at the default stages, and then adapt your process around them.

It is better to do the reverse:

sales process → real events → pipeline stages → CRM configuration

A pipeline stage should ideally correspond to an observable change in the state of the opportunity.

For example:

"Interested"

is a stage that is difficult to interpret. Two sales reps can have very different definitions of what an interested prospect is.

Conversely:

"Diagnosis completed"

describes a verifiable event.

For each stage, three questions must have a clear answer.

What is the entry criterion?

What must be true for an opportunity to reach this stage?

What is the exit criterion?

What event allows the opportunity to move forward?

What is the next expected action?

What must the team do to trigger this event?

This logic avoids having a pipeline made up of subjective statuses like "hot", "interested" or "to watch".

It also allows multiple sales reps to interpret the stages in the same way.

The number of stages must then remain proportionate to the actual process. A simple sales cycle does not need twelve stages just to give an impression of precision.

The more complicated the pipeline becomes to maintain, the more likely the teams are to bypass it.

Example of a sales pipeline for a B2B service company

Let's take a consulting firm or an agency whose sales cycle relies on an initial discussion, a diagnosis, and then a commercial proposal.

A pipeline could look like this:

This pipeline is not a template to be copied.

A company offering a low-ticket standard offer will probably have a much shorter cycle. Conversely, a complex sale involving multiple decision-makers, a technical phase, or a formalized purchasing process will require other stages.

The CRM architecture must start from the real process. This is also the principle we apply when we design or rebuild a CRM: simplify the operations before configuring the objects, stages, views, and automations.

What data should you track for each opportunity?

Creating all the properties offered by the CRM does not make the pipeline more reliable.

Each piece of information should serve either to make a decision, trigger an action, or analyze the process.

For a B2B organization, a few data points are often enough to build a solid initial foundation:

Not all of these are mandatory in every company.

An agency selling highly variable projects may need an estimated amount as early as the diagnosis. A business whose price is only determined at the time of the proposal can fill it in later.

The same principle applies to the many qualification fields, probabilities, or scoring options:

Data that doesn't help anyone decide or act generally ends up becoming poorly maintained data.

CRM quality depends less on the number of available fields than on the reliability of the few pieces of information actually used.

How to avoid a pipeline filled with dead deals?

A pipeline quickly becomes unusable when opportunities that are no longer moving forward remain mixed in with genuinely active deals.

A proposal sent three months ago with no response is not necessarily an open opportunity anymore.

It is useful to distinguish between at least several situations.

Active

A next action exists and the opportunity is genuinely moving forward.

Stale

The opportunity should be moving forward, but there is no recent activity or credible next action.

Nurturing

The need potentially exists, but the timing is not close enough to keep the deal in the active pipeline.

Lost

The project will not go through in the current sales cycle.

Therefore, nurturing should not become a catch-all stage used to avoid closing an opportunity.

In the same way, "no response" is generally not a sales stage.

One of the simplest checks consists of regularly identifying:

open opportunities with no scheduled next action.

This information is often much more useful to a sales rep than a dashboard containing twenty indicators.

The CRM must facilitate daily action before producing reporting.

What KPIs should you use to manage your sales pipeline?

Once the stages are sufficiently reliable, the pipeline allows you to analyze the actual performance of the sales process.

A few indicators are particularly useful.

Number and value of active opportunities

They give a view of the sales volume currently in progress.

Conversion rate between stages

It identifies the places where opportunities drop off.

For example, a high volume of diagnoses but very few proposals can signal a problem with qualification, the offer, or the process.

Time spent in each stage

An opportunity staying twice as long as normal in a stage probably deserves special attention.

Sales cycle length

Measuring the time between the creation of an opportunity and its closing helps to better anticipate signings and sales capacity.

Win rate and reasons for loss

A conversion rate alone explains very little.

Associating lost deals with sufficiently structured reasons — budget, timing, competition, lack of priority, bad fit — allows you to identify more actionable trends.

Closing probabilities and forecast

These can also become interesting, but you should avoid arbitrary rules like:

Qualification = 20%
Proposal = 50%
Decision = 80%

These probabilities become truly useful when they are progressively calibrated using the company's historical sales data.

The pipeline is then a building block of a broader revenue management system, which aligns directly with the RevOps logic: aligning processes, data, tools, and teams rather than treating the CRM as a simple contact database.

What can be automated in your sales pipeline?

Once the process is stabilized, several administrative tasks can be automated without making the system more complex.

When an opportunity moves to Proposal sent, the CRM can, for example:

• automatically record the date sent;

• create a next follow-up task;

• alert the owner if no activity has occurred after a certain period.

When an opportunity becomes Won, the system can:

• record the close date;

• pass the necessary information to the delivery team;

• create an onboarding;

• notify the relevant teams.

AI can also be relevant when information needs to be interpreted rather than simply tested.

For example, a meeting debrief might contain:

The project is approved in principle, but the proposal must be reviewed with the finance department in early September.

An AI can help extract the context, the date, or the next action. A classic rule, however, remains more reliable for handling:

Stage = Won → launch onboarding.

So the principle remains:

Use a rule when a rule is enough. Use AI when information actually needs to be interpreted.

Modern CRMs now integrate a growing portion of this logic directly into their automation engine. Our guide on Attio Workflows shows, for example, how far a sales process can stay within the CRM before a layer like Make, n8n, or a dedicated integration becomes relevant.

But automation must come after pipeline design.

Automating a poorly defined process usually just executes bad rules faster.

Which CRM should you use to manage your sales pipeline?

The best CRM depends less on the number of features available than on the process it needs to represent.

Pipedrive remains particularly suitable for teams looking for a sales-focused CRM centered around the pipeline and sales activities.

Attio brings more flexibility to design the data model, relationships, and workflows around customized B2B processes.

HubSpot becomes highly relevant when sales CRM, marketing, and customer service need to operate within the same ecosystem.

Airtable can also serve as a CRM when the process is very specific or heavily connected to operations and business tools.

However, the tool comes after defining the process. Migrating to a more sophisticated CRM will not fix ambiguous stages or opportunities that are never updated.

How to build a genuinely usable sales pipeline?

The right method can remain simple.

1. Observe the actual sales cycle

Identify how an opportunity progresses today, and not how it should theoretically progress.

2. Define precisely when an opportunity exists

Avoid turning every contact or conversation into a deal.

3. Identify the events that actually move a sale forward

Meeting held, need validated, proposal sent, approval obtained…

4. Turn only these useful events into stages

The pipeline must remain simple enough to be understood without constant documentation.

5. Define the necessary data and next actions

Each stage must help the team know what to do next.

6. Implement this logic in the CRM

Only now do the objects, fields, views, and permissions come in.

7. Automate sufficiently stable rules

Nothing more.

8. Monitor usage and evolve the system

A pipeline is not set in stone. Data and feedback from sales reps allow you to progressively remove useless stages, clarify certain rules, or improve management.

Therefore, a good sales pipeline is not the one with the most stages, fields, or automations.

It is the one that the team can maintain easily enough so that it remains true to reality.

Frequently asked questions about the B2B sales pipeline