Sales stages are the tasks a conversation has to finish, in order, from the first hello to an agreement about a specific next action. The scheme exists for one reason: so that at any second of the call the seller knows where they are and what has to happen next. Without that, a call becomes a pleasant chat that ends in “let me think about it.”
Below: the classic five-step scheme and the extended seven, what a rep actually does at each step, the questions that move a deal, how the conversation map differs from the deal statuses in your CRM, and the one place calls break more often than anywhere else.
Where this frame comes from
Ten years of running call centres, up to fifteen hundred live operators at peak, across France, Mexico and beyond. The five- and seven-step schemes themselves are the common ones, written down the way phone sales has settled on them over decades. What is ours is everything about how they fail: the done-signals, the tells you can hear on a recording, and the timings further down. None of that is research and none of it is an industry standard — it is what survived reviewing those floors’ calls.
What sales stages actually are
Every conversation about buying travels the same road. First the person decides whether you are worth listening to. Then they tell you what they need. Then they measure the offer against what they expected, and they doubt. Only after that do they agree or refuse. The scheme gives each piece of that road a name and a job.
The practical value is not in the names. It is in the done-signals. Every step has an outcome you can observe, and the outcome is what tells you whether you may move on. The commonest failure in phone sales: a rep describes the product without ever finding out what the person needs. He talks for ten minutes, misses, and gets “thanks, I’ll think about it.” On paper all five steps happened. In fact the call fell apart at the second one.
Treat it as a map rather than a text. The words are different every time; the set of tasks is the same. A strong rep walks the steps invisibly and would not name them if you asked. A new one needs the map literally, or he drifts into a monologue about the product inside the first minute.
The scheme goes by several names — the five steps, the classic five, the stages of the sales technique. The differences are cosmetic. What follows is the most widespread version of it.
A step is not finished because you said your part. It is finished when the other person did theirs.
The five stages of a sales call
Each step has three properties: a task, a done-signal, and a typical mistake. The mistakes are the more useful column, because they are where a conversation loses momentum and because they are audible on a recording.
- Contact. The task is to earn the right to the conversation. Done when the other person answers your first question on the substance of it. The mistake is a long opening about your company and its achievements.
- Needs. The task is to understand the request in the customer’s own words. Done when you can repeat back what they need with nothing invented added. The mistake is two closed questions in place of a conversation.
- Presentation. The task is to show how the offer closes what the customer named themselves. Done when they start asking about terms, timing or price. The mistake is a tour of the entire product.
- Objections. The task is to find out what sits behind “too expensive” or “I need to think about it,” and to answer that. Done when the doubt has been named concretely. The mistake is arguing with the first wording you hear.
- Close. The task is to agree a specific action with a date and a time on it. Done when the agreement has a date. The mistake is an ending along the lines of “I’ll give you a ring sometime.”
The fifth step is usually pictured as a signature on a contract. On the phone the close is smaller than that: agreement to a survey, a meeting, an invoice, a trial order. Each of those moves the deal forward, and each of them gets talked through to a date and a time rather than left as “sometime next week.”
What the rep actually does at each stage
Everyone knows the theory. The difference between a strong seller and a weak one appears in the specific actions, so here is the minimum set of them.
- Contact. Name yourself, the company, the reason for this call to this person, and ask permission for a couple of questions. “Ivan, this is Anna from N. You left a request for a quote yesterday. I’ll ask two questions so I don’t offer you the wrong thing — all right?”
- Needs. Ask open questions and write the customer’s wording down verbatim. Those same words come back in the presentation, which is most of why they are worth writing down.
- Presentation. Name two or three benefits, strictly against what the person said themselves. Check after each one: “Is that the part that matters to you?”
- Objections. Establish the meaning before you answer. “Expensive compared to what?” gets further than a prepared argument about quality.
- Close. Offer a specific step with a date in it: “I’ll hold a specialist for Thursday at three — shall I confirm it?” An open ending, of the “have a think and I’ll call back” kind, is a lost deal, and it is worth counting as one in the numbers.
About the order. Going back is a normal part of the job. If a doubt surfaces during the presentation that you did not expect, the needs step was not finished, and the right move is to return to the questions. What looks bad is the other thing entirely: mechanical progress down the list while the previous task sits unsolved.
Telling the floor to “work all five stages” changes nothing
The instruction cannot be broken in any way anybody can point at, which is why it survives so many sales meetings. What changes behaviour is one recording played back with two marks on it: the second where the rep jumped into the presentation, and the question he should have asked before it. That is a conversation about evidence rather than about effort, and it is short.
Stage one: earning the right to the conversation
The first seconds decide whether the rest of the call happens at all. In that window the person is answering one question for themselves: is this spam or is this business. Everything you say afterwards is heard through that filter, including the parts you were proud of.
A working opening has four elements: who you are, what company, why you are calling this person specifically, and a request for a couple of questions. The budget for it is 10 to 15 seconds, hello to the reason for the call — a prescription from the same frame as the timings further down, not a measurement of anything. The reason has to be concrete. “You left a request on the site yesterday afternoon” works, because the person recalls their own action. “We offer services for business” fits every number in the country, which is exactly why it does nothing.
- “Is now a good time?” at the very start. It hands the other person a ready-made reason to end the call before they have any idea what it is about.
- “Am I interrupting you?” The same trap, more politely worded.
- Half a minute of company history and partner logos. By the time the point arrives the line is already dead.
- Reading off a page. Flat intonation with no pauses in it is audible instantly, and it reads as an autodialler rather than as a person.
If the timing genuinely is wrong, agree something exact: “Right — I’ll call at ten past six, does that work?” A precise time gets answered far more often than a vague “I’ll try later.” On inbound the step is shorter, because the caller already chose to ring: name yourself and confirm they have reached the right place. And there is only one done-signal for this step — the other person answered your question in their own words. A one-syllable “uh-huh” does not count as an answer.
The questions that move a deal
Put the stages and the questions for each of them together and you have a frame you can work from without a memorised text. These are the ones that most often get a deal moving off the spot.
- About the situation. “How are you handling this at the moment?” “Who else is involved in the decision?”
- About the problem. “What doesn’t work in the current setup?” “What happens if it all stays exactly as it is?”
- About the criteria. “How will you know an option is the right one?” “Which matters more here, the timing or the price?”
- About dates and money. “By what date do you need it?” “What sort of budget are you working to?”
- About the next step. “If the terms suit, what has to happen for us to start?”
Three rules, without which the questions work at half strength. One question at a time: ask two in a row and the other person merges them and answers on the convenient one. A pause after the question, because silence pulls out the more useful half of the answer, and reps fill that silence because it is uncomfortable rather than because it helps. And give the customer’s own wording back to them — a person who hears their own words is readier to move on.
Two or three questions is usually the whole of the needs stage, provided they are questions that change what gets offered afterwards. That count is a prescription we work to, taken from our own floors rather than from research, and the real test is not arithmetic: the step is finished when you can repeat the request back in the customer’s language.
Deals are moved by questions. Arguments only fasten down what the customer already said.
Seven stages: the extended scheme
The extended scheme adds one step at each end of the five — one before the call, one after the deal. That makes seven.
- Preparation. Before dialling, look at the record: where the lead came from, what they looked at, how the last conversation ended. A couple of minutes here makes the reason for the call precise and shortens the first step.
- Contact. The right to the conversation.
- Needs. The request in the customer’s own words.
- Presentation. A solution against the request that was actually named.
- Objections. Working through whatever stands in the way of agreeing.
- Close. An agreement with a date on it.
- After the deal. Confirming the order, reminding about the date, checking it went through, the repeat sale and the request for a referral. This is where the second and third deals with the same customer come from.
You will meet versions that put negotiation over terms, or the upsell, in seventh place. The argument about the correct number is not worth having: count as many steps as there are tasks your deal cycle actually solves. If installation, training and renewal follow the payment, seven describes your process more honestly than five does.
How the stages relate to your CRM funnel
Funnel stages are the deal statuses in the CRM: new lead, qualified, proposal sent, invoiced, paid. They describe a customer journey measured in days and weeks. The conversation steps live inside a single call and take minutes. One is nested inside the other: for a deal to move honestly into “proposal sent,” contact and needs have to be closed in the conversation first.
The most expensive breakage in a sales team
Statuses move while the conversation tasks stay unsolved. In the CRM the deal looks warm and sits at “proposal sent.” On the recording there was no needs stage at all — the rep emailed a price list the first time it was asked for. Two weeks later the deal hangs with no explicable reason, and the forecast that was built on it was never real in the first place.
The check is simple. Take the deals that have sat at one status longer than the rest and listen to the last conversation on each. You are not listening for tone or for politeness. You are listening for whether the step the status claims is finished ever happened.
Where the call breaks most often
Listen to recordings back to back and two narrow places show up. The first is the move from the greeting to the questions: the person still has not worked out why they are being called, and hangs up. The second is a skipped needs stage — the rep hears the first hint of interest, goes into the presentation, and talks blind from there to the end.
On a recording it is the same handful of tells every time:
- the rep holds most of the talking time;
- one question in the whole call, or none at all;
- the answer to “too expensive” arrives within a second, with no attempt to find out what it means;
- the customer asks the same question twice;
- the call ends with no date and no next step in it.
The catch is that all of it lives in the audio. A team lead listens to a few calls a week, the sample is whatever was to hand, and a systematic hole at the second step does not show up in a sample like that. In our own operations under 5% of recorded calls are ever reviewed by a human, and that is the honest ceiling of manual quality control rather than a comment on anyone’s diligence. Nobody has the hours.
While calls are sampled, a sales team gets repaired from random examples.
Why a neuro-agent runs the same map
Speech analytics is what makes the stages measurable instead of arguable. Locator listens to every call and scores it on 30+ parameters — how much of the talking time the rep took, whether open questions were asked, whether an objection was raised and how it was answered, whether the agreement had a date in it. That is 100% of calls, against a handful per rep per month under manual review. It runs at 0.10 per minute ex VAT and starts from the recordings you already have, so there is nothing to install before the first report.
What changes when the sample is everything
Under manual quality control, whether a rep works the needs stage is an opinion argued from two recordings, and the rep’s counter-opinion is just as well supported. At 100% coverage on 30+ parameters it is a number per rep and per scenario. The coaching conversation then starts from the same page for both people in the room, which is most of why it goes anywhere.
Some of the conversations can be handed over outright. A neuro-agent calls in a voice and walks the same map: contact, needs, offer, dated next step. Three shapes of it, at published rates ex VAT. Bene Hotline answers inbound and informs, at 0.20 a minute. Bene Qualifier runs the first stages, finds out whether there is interest and passes a warm customer to a person, at 0.25. Bene Sales carries the conversation through to an order, at 0.30. The point of the arrangement is not that a machine sells better than your best rep — it is that the routine first steps stop consuming the hours of the people who are good at the fourth step and the fifth.
The frame the agent runs is the same one worth prescribing to the floor. A cold call that is working gets from hello to a dated next step in under four minutes, and one recovery attempt past a first refusal is the limit before the number is burnt for good.
What to do this week
- Write the five steps down with a done-signal against each one, in the words your product actually uses. One page. If two people on the team would write different done-signals for the same step, that disagreement is the finding.
- Pull the deals that have sat longest at a single status and listen to the last call on each. Mark the step where the conversation stopped doing its job.
- Count the questions in each of those calls. Not the words — the questions.
- Repair one thing only: the transition from greeting to first question. It is the cheapest step to fix and the one most calls die at.
- Decide now how you will know next month whether anything changed. Hand-sampling more recordings is not an answer, because you already know what a sample that size can and cannot see.
The scheme itself is not the point. A map is worth having because it lets you say where a conversation stopped, in a sentence both people in the room accept — and once that sentence is available, a sales team stops being fixed by anecdote. Everything above is a prescription drawn from our own floors and offered as a starting frame, not as a standard. Keep the steps whose done-signals your team can genuinely observe, and drop the rest.