Implementation
How to design HubSpot sales pipelines that reflect how you sell
Stages that reflect customer commitments, properties required at each step and rules that keep the forecast reliable. With a complete worked example.
A sales pipeline in HubSpot should be designed around what the customer does, not what the sales rep does: each stage should correspond to a verifiable commitment from the buyer and have a clear exit criterion. Once that is sorted, the probabilities, required properties and reports almost fall into place on their own.
In this guide you will see how to define the stages, when to create more than one pipeline, what to ask for at each step and which rules to switch on so the data is reliable.
What a pipeline stage is (and is not)
The temptation is to copy the internal process: "call made", "email sent", "proposal being prepared". Those are tasks, and tasks belong in the deal's activities. A stage answers a different question: where is the customer in their decision?
The test is simple. If two reps look at the same deal, they should put it in the same stage without talking to each other. If they do not, the stage is badly defined, and everything calculated on top of it (forecast, stage-to-stage conversion, time in stage) will be noise.
How to define the stages step by step
- Reconstruct three or four recent sales, won and lost, with the rep who handled them. Note the moments when the customer did something: accepted a meeting, shared information, asked for a price, brought in another person.
- Group those moments into five to seven open stages, plus the closed won and closed lost stages HubSpot includes by default.
- Write the exit criterion for each stage as one verifiable sentence, such as "the customer has confirmed budget and decision date".
- Assign a probability. If you have history, calculate the real conversion from each stage; if not, start with estimates and review them after six months.
- Decide which properties each stage requires before a deal can enter it.
An example for a B2B services company with a consultative sale:
| Stage | Exit criterion | Properties on entry |
|---|---|---|
| Discovery | Meeting held with someone from the affected area | Main need, source |
| Qualified | Budget, timeline and decision-maker identified | Estimated amount, close date, associated decision-maker |
| Proposal sent | The customer has the proposal and there is a review date | Associated quote, amount |
| Negotiation | The customer has asked for changes or terms | Competitors, discount requested |
| Verbal commitment | Accepted, pending signature | Expected signature date |
- 01Discovery
- 02Qualified
- 03Proposal sent
- 04Negotiation
- 05Verbal commitment
- 06Won or lost
Probabilities: the number nobody reviews
Each stage has a closing probability, and HubSpot multiplies it by the amount to calculate the pipeline's weighted amount. If the probabilities are the defaults or whatever someone entered on day one, the forecast based on them is not worth much.
Once you have a few months of history, calculate what percentage of the deals that passed through each stage ended up won and adjust. Do it per pipeline: a renewal in negotiation does not have the same probability as a new sale in the same stage. And if the team also uses forecast categories (pipeline, best case, commit), make it clear that those are the rep's judgement and the stage probability is the statistical figure. Comparing the two is one of the most useful conversations in the weekly sales meeting.
How many pipelines do you need in HubSpot?
Fewer than you think. A new pipeline is justified when the process genuinely changes: different stages, very different sales cycles or separate teams. Selling two products with the same process does not justify it; a "product line" property and filtered reports are enough for that.
Cases where it does make sense to separate: new business versus renewals or upsells, direct sales versus resellers, or a post-sale pipeline (onboarding, delivery) run by another team. When there are several pipelines, keep stage names and meanings as similar as possible so consolidated reports make sense.
What to ask for at each stage without slowing reps down
In the pipeline settings you can mark properties that the user must fill in when moving a deal into a specific stage. It is the most effective tool for clean data and the easiest one to ruin: if every stage asks for eight fields, the rep will type anything just to move forward.
Our rule is to ask for two or three properties per stage at most, and only those that feed a decision or a report. Amount and close date, as soon as the deal is qualified. Lost reason, always, as a dropdown with an "other" option that forces an explanation. Everything else can live on the record and be filled in when the time comes.
Pipeline rules and automation by stage
With Professional or Enterprise you can switch on pipeline rules. The most useful ones:
- Prevent skipping stages, across all of them or only the critical ones. Closing as won or lost is still allowed.
- Prevent moving backwards after a given stage, so the forecast does not jump around.
- Limit who can edit deals in specific stages, for example once they have passed to finance.
- Deal approvals before moving forward, available in Sales Hub Enterprise.
In the pipeline's automation tab you can create tasks or send notifications when a deal enters a stage. Use it for the simple things (creating the "send proposal" task on entering Qualified) and keep workflows for logic with conditions. Winning a deal usually triggers the most valuable steps: creating the invoice in the ERP, alerting operations or opening the project. At several companies that invoice with Holded, a Spanish invoicing tool, for example, the won deal creates the invoice or pro forma with the quote's line items and sends the number and PDF back to HubSpot.
Leads and deals: where the pipeline starts
A common mistake is creating a deal for every contact who fills in a form. The pipeline fills up with opportunities that are not opportunities, the forecast inflates and conversion rates stop meaning anything. The earlier phase (contacting and qualifying) has its place in the Leads object, with its own pipeline of stages such as new, attempting, connected and qualified, or in lead status if you do not use it.
The deal is created when there is a real opportunity. With Sales Hub Professional or Enterprise you can even require a deal to be created when a lead becomes qualified, so none get lost in the handover.
Typical mistakes when designing pipelines
- Stages that are tasks ("call", "send email").
- An "On hold" stage where deals go to die without ever being closed as lost.
- The same probability on every intermediate stage because nobody reviewed it.
- One pipeline per rep, instead of one pipeline with owners.
- Changing stages with hundreds of deals in them without reviewing the workflows and reports that depend on them.
That last one deserves attention: before deleting or renaming a stage, check which workflows, lists and reports use it. The same applies to properties, which break dependent automations just as easily. And if you want to make the most of stages in reporting, HubSpot dashboards for leadership shows how to use time in stage and stage-to-stage conversion.
How to tell whether your pipeline works
Three indicators tell you early on. Average time in each stage: if one accumulates deals for months, its exit criterion is unclear. Conversion between consecutive stages: if one lets almost everything through, it may be surplus. And the percentage of won deals that skipped stages: if it is high, the stages do not reflect how your customers buy.
Review the design at three and six months with the sales team. Designing pipelines is part of every HubSpot implementation we do at Vexco and, when you are coming from another CRM, part of the migration too, because it is the best moment to reorganise stages before moving the history.
