Structuring pipeline stages properly: a template for SMEs

Most pipelines fail not because of the number of stages but because of the cut: the boundaries between stages are blurry, and suddenly every move becomes a judgement call. Structuring pipeline stages properly means placing every stage boundary at a verifiable customer commitment – not at one of your internal activities. This guide covers the craft behind the cut, with a template your SME can adopt as it stands.

How do you recognise a well-cut stage?

By four traits: the transition is verifiable, it hangs on the customer's behaviour, the stages do not overlap, and each stage carries its own win probability. If two team members would independently place the same deal in the same stage, the cut is good.

The most important trait is the customer commitment: "proposal sent" is your activity – it says nothing about where the customer stands. "Customer discussed the proposal in a meeting" is a commitment. Stages that hang on customer behaviour turn your pipeline into an early-warning system; stages that hang on your activities turn it into an activity report.

Which cutting mistakes do SMEs make most often?

Four patterns keep reappearing – all four are avoidable when you re-cut. First: activity stages like "proposal created" or "email sent", which measure your effort instead of the customer's progress. Second: parking stages like "On hold" or "Later", where deals moulder instead of being decided – the lost archive with a follow-up reminder exists for exactly that.

Third: salami stages like "Negotiation 1", "Negotiation 2", "Negotiation 3" – they fake a precision nobody maintains consistently. Fourth: the first stage as a contact dump. If every trade-fair contact lands in "New", the conversion of the whole pipeline gets diluted. A pipeline holds deals – concrete sales opportunities. Contacts without an opportunity stay in contact management.

What does the stage template for SMEs look like?

Five stages are enough for most Swiss SMEs in B2B. Each stage gets a definition plus an exit criterion that must be met before the deal moves on.

1. Incoming: a concrete sales opportunity is identified. Exit: a first conversation has taken place; need and contact person are known.
2. Qualified: the deal is worth pursuing. Exit: budget range discussed, decision path known, next meeting booked.
3. Proposal discussed: not just sent – discussed. Exit: the customer has gone through the proposal with you and given feedback.
4. Decision: negotiation over price, scope, contract. Exit: verbal commitment or a definitive no.
5. Close: won (contract signed) or lost – both recorded with a reason, because lost deals are your best source of learning.

Notice what is missing: no "Follow-up" stage (that is a task, not a stage), no parking stage, no activity stage. Every boundary hangs on something the customer did.

How do you adapt the template to your business?

Change the template where your buying process demonstrably differs – and only there. If you sell explanation-heavy projects, a "Needs analysis" stage between "Qualified" and "Proposal" often pays off, with the exit criterion "requirements confirmed in writing". With short cycles and standard products you can merge "Decision" and "Proposal".

The ground rule from defining your sales process applies: stages follow your customers' buying process, not your preferred structure. And if you run very different lines of business – say projects and maintenance contracts – separate them into multiple pipelines rather than building one universal pipeline made of compromise stages.

Why does the stage cut decide your forecast?

Because every stage carries its own win probability – and the weighted forecast is built directly from it. A CHF 20,000 deal in "Qualified" at 25 percent counts as CHF 5,000; the same deal in "Decision" at 75 percent counts as CHF 15,000. If the stage boundaries are blurry, those percentages are fantasy – and so is your entire forecast.

You do not have to guess the probabilities either: after a few months of a cleanly run pipeline you read them from your own history – how many deals that reached "Proposal discussed" were actually won? That is why a clean cut pays off twice: it improves day-to-day management and it makes your planning credible. A blurry cut, by contrast, produces numbers nobody trusts month after month – and a forecast nobody trusts is a forecast nobody uses.

How do you migrate to the new cut?

In three steps, on one quiet afternoon: first create the new stages with their exit criteria in the CRM, then sort every open deal individually – strictly by criterion, not by feeling – and finally note the criteria where the team can see them. Deals that fit no stage are usually no longer deals at all: archive them instead of dragging them along.

Then hold steady: run six weeks, change nothing. Only a few weeks of data show whether the cut holds – conversion rates per stage reveal where deals get stuck and whether a boundary is blurry. In Advanzo you adjust stages yourself at any time; existing deals keep their history.

How does AI keep the stages current?

The best cut is worthless if deals are not kept up to date – and in daily business that is the first thing to slip. Autopilot in Advanzo analyses incoming email on open deals and, among other things, proposes stage changes: if a customer confirms the proposal meeting by email, a "Change stage" suggestion card appears with AI reasoning and a confidence level. You approve, edit or dismiss – with one click.

Daily scans additionally flag deals that have passed their close date or have sat unusually long in one stage – often the first signal of a blurry boundary or a sleeping deal. Autopilot is part of the AI add-on (CHF 9 per user/month) and never sends emails automatically; every step needs your approval. Details are on the pricing page.

Frequently asked questions (FAQ)

What is the difference between a stage and a task?

A stage describes where the customer stands in the buying process; a task describes what you do next. "Send follow-up" is therefore never a stage but a task within a stage. Mixing the two fills your pipeline with activity noise.

Do "Won" and "Lost" belong in the pipeline as stages?

They are end states, not working stages – which is why most CRMs track them separately. What matters is that both are recorded with a reason. One quarter's loss reasons are worth more than any sales training.

How strictly should exit criteria be enforced?

Strictly enough that the numbers hold, loosely enough that nobody games the system. A proven rule: moving a deal is allowed as soon as the criterion is met – and the weekly review spot-checks samples rather than policing every click.

What do I do with paused deals if there is no on-hold stage?

Mark them lost with the reason "postponed" and set a follow-up task for the right moment. The pipeline stays a picture of your genuinely open opportunities, and the contact is still not forgotten – the reminder brings it back.

Do different products need different stages?

Only if the buying process genuinely differs. If only price or duration differ, one pipeline is enough. If the customer's steps differ – say project business with tenders versus a standard product bought directly – separate pipelines are cleaner.

How often should I revise the stage cut?

At most once or twice a year, and only with data behind it – say a stage that is constantly skipped or whose conversion has been out of line for months. Every rebuild costs comparability: after a re-cut, your stage statistics start again from zero. Small corrections to the exit criteria, on the other hand, are fine at any time, as long as the stage boundaries themselves stay put.

A clean stage cut turns your pipeline into a measuring instrument instead of a collection of opinions. Advanzo brings everything you need: freely adjustable stages on a Kanban board, conversion numbers per stage and an Autopilot that proposes stage changes before deals go stale. Swiss hosting, FADP/GDPR compliant, free up to 25 deals. Build your pipeline with the template – start for free.