A SaaS sales pipeline is the sequence of states a buyer moves through between the moment you notice them and the moment they pay you every month. We run it in six stages: signal, conversation, meeting booked, qualified opportunity, proposal, signed and live. In our experience most SaaS pipelines do not break at the close, they break at stages one to three, which is why this guide spends most of its time there, with benchmarks from our analysis of 1,000,000 B2B cold calls in France.

Why SaaS pipelines break at the top, not at the close

SaaS has three properties that change how a pipeline behaves. Revenue is recurring, so a bad-fit deal costs you twice: once to close, once to churn. Cycles are longer, so a stage definition that is a feeling rather than an event lets deals sit for a quarter. And buying groups are wide, so a single contact is a fragile deal.

Put those together and late-stage fixes become expensive. A proposal that dies in legal was usually lost at qualification. A demo that goes nowhere was usually booked on a pitch-first opener to a persona who never had the problem.

The cheapest place to fix a SaaS pipeline is therefore the entry. Two numbers from our data show how much room there is: reps use only about 20% of the signals already in their CRM, and 50% of Sonik's own pipeline comes from signals that were already sitting there.

The 6 stages of a SaaS sales pipeline, with a benchmark for each

Each stage below has one exit event. If the event has not happened, the deal has not moved, whatever the rep feels.

Stage 1: Signal

We do not start the pipeline at "prospecting", we start it at a reason to call. A website visit from a target account, a champion who changed jobs, a closed-lost deal passing its six-month mark, a lead that went quiet after a demo. The exit event is an account on a call list with the trigger written next to the name. Benchmark: in our experience most teams act on roughly one signal in five.

Stage 2: Conversation

A conversation is a live exchange with the decision-maker, not a gatekeeper or a voicemail. Exit event: the person picked up and you spoke. French B2B average connect rate is 11.3%, with a range of 11-20% depending on persona and time slot. At 4pm it is 15.8%, at 8am 9.8%. In our experience, most connects happen within the first three attempts. Our north star for this stage is 20 conversations per SDR per day.

Stage 3: Meeting booked

Exit event: a calendar invite accepted, with a date. "Interested, will send times" is still stage 2. Benchmark: trigger-based openers reach a 7-10% meeting rate per conversation; pitch-first openers get 2-3%. In our data 65% of reps still open by explaining what they sell.

Stage 4: Qualified opportunity

The meeting was held, the problem is confirmed, there is a next step with a date. Exit event: the prospect answered the "how" questions and agreed to the next step. Discovery questions that start with "how" lift conversion by 190% in our benchmark, which is why this stage is where we coach hardest.

Stage 5: Proposal and second contact

Pricing is on the table and at least two people on the buyer side have seen it. Exit event: a second stakeholder joined a call. In our experience a single-threaded SaaS proposal is a coin flip; a multi-threaded one stalls far less often, because objections surface before the proposal, not after.

Stage 6: Signed and live

Signed is not the end for a subscription business. The stage closes when the first user logs in and the kickoff is done. A deal that never goes live churns at renewal and should count against the rep who closed it. Closed-lost deals leave this pipeline and re-enter stage 1 as a signal six months later.

Stage exit criteria keep a pipeline honest

The single biggest improvement we have seen in SaaS pipelines is not a tool, it is replacing stage definitions written as feelings ("engaged", "warm", "likely") with definitions written as events you can see in the CRM or the dialer. A meeting is in the calendar or it is not. A second contact was on the call or was not.

Three side effects follow. Forecasts stop depending on rep optimism. Pipeline reviews get shorter because there is nothing to debate. And stalled deals become visible, because "days since last exit event" is a number you can sort by.

9 fixes, one per bottleneck

1. Stage 1, empty signal list: export website visitors, job changes and dormant leads from the CRM every Monday and dial them before any cold list.
2. Stage 2, connect rate under 11%: move dial blocks to 10am and 4pm. Founders pick up around 10am, marketing around 4pm, HR around 1pm.
3. Stage 2, too few conversations: parallel dial. It gives reps about 5x more conversations per hour, which is how you reach 20 a day without hiring.
4. Stage 3, under 3% meeting rate per conversation: rewrite the opener around the trigger, not the product.
5. Stage 3, "send me an email": book the slot while the prospect is still on the line and send the invite before you hang up.
6. Stage 4, demos that go nowhere: replace yes/no qualification with "how" questions and ask "What do you mean?" before answering any objection (+197% in our data).
7. Stage 4, no-shows: confirm the day before with the reason the prospect gave you on the call, in their words.
8. Stage 5, proposals that stall: no proposal leaves without a second contact having joined a call.
9. Stage 6 and closed-lost, leaking revenue: call closed-lost and dormant contacts with "How have you been?" rather than a cold open. That line reaches about 10% success versus 1.5% for a standard cold open.

How Sonik approaches SaaS pipeline stages

Sonik is built around stages one to three. Signals pull website visitors, job changes, closed-lost anniversaries and dormant leads out of the CRM into a daily call list, so stage 1 fills itself instead of depending on a rep's spare afternoon.

The parallel dialer handles stage 2: several lines at once, voicemails skipped, a live conversation handed to the rep with the trigger and the last CRM notes on screen. That is how a rep gets to 20 conversations a day.

Stage 3 is the rep's job, but AI call prep writes the trigger-based opener before the call, and coaching flags pitch-first openers and one-line objection handling after it. The dashboard reports conversion from each stage to the next, so you see where the funnel narrows before the quarter tells you.

Conclusion

A SaaS pipeline is only as predictable as its stage definitions and only as full as its signal list. Fix the top three stages first: signals in, conversations at the right slot, meetings booked trigger-first. This week, rewrite each stage of your pipeline as a single observable exit event and sort your open deals by days since that event. The deals at the top of that list are your real pipeline problem.