HubSpot pipeline stages for an agency: why the default seven are wrong

The setup builds you a seven-stage sales pipeline even when you tell it you work alone. Here's what an agency pipeline should actually contain, and why fewer stages hold up better.

When I set up a HubSpot account and told it I was one person running agency work, it built me a pipeline with seven stages: Prospecting, Initial Contact, Qualification, Proposal Sent, Negotiation, Closed Won, Closed Lost.

It’s a competent pipeline. It’s also a sales-team pipeline, handed to someone who had just said there was no sales team.

That gap is worth understanding before you accept any default, because the pipeline you inherit on day one is the one you’ll still be working around a year later.

What a pipeline stage is actually for

A stage exists to answer one question: what has to happen next, and who does it.

That’s the whole job. If two stages have the same next action, they’re one stage. If a stage has no action attached — if it just describes a feeling about the deal — it isn’t a stage, it’s a note.

This matters more for a small team than a large one. A sales floor can afford fine-grained stages because moving cards is someone’s job. When the person moving the card is also delivering the work, every extra stage is admin that competes with billable time — and loses.

Why seven stages breaks for a small agency

Look at what the default asks you to distinguish:

Prospecting vs Initial Contact. The difference is whether you’ve reached out. For a one-person shop with a handful of live conversations, you know this without a column.

Qualification vs Negotiation. Real agency deals don’t move cleanly between these. You qualify during negotiation and renegotiate after qualifying. Cards end up parked in whichever one you touched last, which makes both meaningless.

Proposal Sent is the one stage in the default that genuinely earns its place. Agencies do formalise before negotiating, and “proposal is out, waiting” is a real state with a real next action: follow up on a date.

The failure mode isn’t that seven stages is wrong in theory. It’s that nobody maintains seven stages while also doing the work, so cards stop moving, and a pipeline where cards don’t move tells you nothing. A stale pipeline is worse than no pipeline, because you still trust it.

What to use instead

Four stages, each with an unambiguous next action:

The two additions that matter more than any extra stage:

A cutoff for silence. Decide what turns a live deal into Lost — three weeks with no reply is a common one — and apply it. Without this, your pipeline value inflates with deals that died months ago, and every forecast built on it is fiction.

A reason on Lost. Price, timing, scope, went with someone else. Four options, one click. After twenty deals this is the most useful data in the whole CRM, and it costs nothing to collect.

Two things the setup does that you should undo

Delete the sample deal. The onboarding created one for $1,000 assigned to Brian Halligan — HubSpot’s co-founder — sitting in the production pipeline. Harmless, but it skews the first numbers you look at.

Check the stage probabilities. Each stage carries a win-probability used for weighted forecasting, and the defaults are template guesses, not your history. If you plan to use the forecast for anything, either set them from your own closed deals or ignore the weighted column entirely. A number that looks measured but isn’t is worse than no number.

Before you build it

You get one pipeline on the free plan. Not one per client, not one for new business and one for delivery — one. That constraint should shape the stages you pick more than any best-practice list, including this one.

If separating clients or separating sales from delivery is the point, check the pipeline count on the tier you can actually afford before you design anything.

Related: every hard limit on HubSpot’s free tier and what a CRM’s setup reveals before you commit.