The B2B sales glossary

Marketing that targets named accounts rather than volume.

ABM picks a list of companies and builds tailored messages and actions for each, in close coordination with sales. It relies on good knowledge of each account's buying committee.

How a customer first came to you.

Referral, trade show, search, prospecting, partner: knowing the channel of each signed customer shows where to invest. Customers often say it themselves in the first conversation, which is more reliable than a CRM field filled in at random.

The average yearly value of a contract.

ACV brings a contract down to its amount over one year, usually excluding one-off fees. It helps compare deals of different lengths and size the sales effort: a high ACV justifies a longer, more hands-on cycle.

Regulation (EU) 2024/1689 governing AI systems.

The AI Act classifies AI uses by risk. Some are banned, others, such as assessing employees, are high risk and subject to strong obligations. For a sales analysis tool, how it is used therefore matters as much as the tool itself.

A programmable access point to a piece of software.

An API lets two pieces of software exchange data with no human step, for instance to send meeting notes to the CRM. It requires development, where a connector or MCP can be used directly.

Subscription revenue over a year.

ARR adds up the yearly value of all active subscriptions. It is the core metric of SaaS companies: it grows with new customers and expansions, and falls with cancellations and downgrades.

A typical portrait of the people who buy.

A persona groups customers who share a role, a kind of company and a buying context. It is useful when it rests on what those customers actually said, their objections and their criteria, rather than on a profile imagined in a workshop.

A four question qualification grid.

BANT checks whether the prospect has a budget, whether the contact can decide, whether the need is real and whether there is a deadline. Easy to remember, it mostly helps sort opportunities quickly early in the cycle. It says little about how the customer will actually decide.

A short card to answer against a competitor.

A battlecard sums up, for one competitor, its strengths, its weaknesses, the objections it raises and the arguments that work. The best ones reflect what customers said about that competitor, not what its homepage says.

Everyone who weighs on a purchase decision.

In B2B, a decision rarely rests with one person: users, managers, finance, IT, legal or procurement each have a say. Mapping this committee, who decides, who influences, who can block, avoids discovering an opponent at signing time.

A sign the customer is moving towards a decision.

A question about rollout, timing or the contract, a new person invited to the meeting, a request for references. Spotted in the conversations, these signals tell where a deal really stands, better than the stage declared in the CRM.

What it costs on average to win a new customer.

CAC adds up a period's marketing and sales spend and divides it by the number of new customers. Computed by acquisition channel, it shows which ones really pay off.

Keeping the audio or video of a conversation.

A recording lets you replay a passage, quote the customer word for word and pass a conversation to a colleague. It involves personal data: participants must be informed and retention limited to what was announced.

The rules for recording a customer conversation.

Recording a call is processing personal data. It rests on a legal basis, requires informing participants before recording, and limits retention and use to what was announced. Participants must be able to object.

Turning the audio of a conversation into text.

Transcription converts speech into text, usually with each speaker's name and a timestamp on every turn. It is the basis of any analysis: without reliable text there is no search, no exact quote and no counting.

An approach where the rep brings a point of view.

From the book by Matthew Dixon and Brent Adamson, the Challenger approach asks the rep to teach the customer something they did not know about their own market, to tailor the message to each stakeholder and to keep control of how the sale unfolds. It requires knowing well what comparable customers go through.

The internal ally who defends your solution at the customer.

A champion has influence, a personal stake in the project succeeding and access to the decision makers. They sell for you when you are not in the room. An enthusiastic contact with no influence is not a champion, and a deal that depends on one stays fragile.

Customers or revenue lost over a period.

Churn is measured in customers lost or revenue lost. The early signs often show in conversations well before the cancellation: a contact who changes, usage that drops, repeated dissatisfaction, a competitor coming back into the conversation.

The moment the customer commits.

Closing means the signature, but it is prepared throughout the cycle: a customer signs when their criteria are covered, their objections answered and their decision process complete. A hard close often reveals a step skipped earlier.

The total revenue expected from a customer over the relationship.

CLV estimates what a customer brings in from first to last day, given their average lifetime and expansions. Compared with acquisition cost, it says how much you can spend to win a customer.

Contacting someone who did not ask for it.

A call or email to someone who does not know the company yet. Its response rate depends mostly on how relevant the message is to that specific person. In Europe, email prospecting is governed by the GDPR and by national rules on commercial communication.

What each side promised to do, and by when.

"I'll send you the quote on Friday", "we'll approve internally by the end of the month": the commitments made during a conversation, on both sides, are easy to forget and costly to miss. Capturing them with their due date lets you follow up at the right time and keep your word.

Knowing what competitors do and what your customers think of them.

Competitive intelligence tracks competitors' offers, prices and announcements. The most telling source remains what customers say about them in conversations: who they are looking at, what attracts them, why they left or stayed.

Analysing customer conversations to extract facts.

These tools record or import calls, meetings and emails, transcribe them, then extract objections, competitors named, commitments or decision criteria. Their value lies in what you can then find and count, customer by customer and across all conversations.

The software where the team tracks contacts, accounts and deals.

A CRM centralises contacts, companies, opportunities and the history of actions. It is only reliable when kept up to date, which costs reps time: hence the value of feeding it automatically with what comes out of conversations.

Selling another product to an existing customer.

Cross-selling offers a product that complements what was already bought. It works when it answers a need the customer expressed, not when it just follows the catalogue.

A summary indicator of an account's risk or strength.

It combines signals such as usage, support tickets, payments and the quality of the relationship. A score with no explanation is hard to act on: it should point to the facts that move it, including what the customer said recently.

The team that makes customers succeed after signature.

Customer success supports rollout, tracks usage, anticipates churn risk and spots expansion opportunities. Its work starts from what sales learned and promised: without a solid handover, it starts from scratch.

Automatically completing contact and company records.

Industry, size, role, contact details: enrichment fills the CRM from external databases. Such information about people is still personal data under the GDPR, including when it comes from public sources.

The contract governing data entrusted to a provider.

When a vendor processes data on behalf of its customers, Article 28 of the GDPR requires a contract setting out the purpose, the security measures, the authorised sub-processors and what happens to the data when the contract ends.

The country where data is stored and processed.

Hosting in the European Union simplifies compliance, but it is not enough: processing by a provider outside the EU is still a transfer, governed by the GDPR. You need to look at where data is stored and where it is processed.

One sale in progress with a customer, from opening to outcome.

One customer can carry several deals over time: a first purchase, an expansion, a renewal. Each deal has its value, its stage and its outcome, won, lost or no decision, and telling these sales apart avoids confusing the relationship with a customer with one particular deal.

What the customer will compare solutions on.

Price, ease of rollout, integrations, security, quality of support: each customer weights their criteria their own way. Knowing them early, in the customer's own words, lets you build the demo and the proposal around what will actually be assessed.

The steps the customer goes through to say yes.

Who approves what, in which order, with which meetings and timelines. A known decision process allows a credible signing date. An unknown one explains a good share of deals that slip from quarter to quarter.

Showing the solution applied to the customer's problems.

A good demo does not tour every feature: it takes the problems heard in discovery and shows how they get solved. It is also when the objections and questions of the other attendees come up.

Knowing who speaks, and when.

Diarization splits a recording by speaker. Without it, a transcript is a block of text where what the customer said can no longer be told apart from what the rep said, which makes any analysis of objections or commitments unreliable.

The first real conversation, to understand before proposing.

The rep asks about the customer's context, problems, stakes and way of deciding. They talk little and listen a lot. What is said there feeds the whole rest of the cycle: qualification, demo, proposal.

The person who can say yes to the budget.

The economic buyer has the final word on the spend, even if they never use the product. They are not always in the conversations: knowing who they are, what matters to them and whether someone carries the case to them is part of any serious qualification.

Representing a text as a list of numbers.

An embedding turns a passage into a list of numbers such that two passages with similar meaning give similar lists. It is what makes semantic search and automatic grouping of passages that say the same thing possible.

The message that follows a conversation or a silence.

A useful follow-up brings something: a recap of what was said, the answer to an open question, the promised document. It is all the quicker to write when the previous conversation was well recorded.

The European regulation on personal data.

The GDPR applies whenever data about people in the European Union is processed, including call recordings, emails or CRM records. It requires a legal basis, clear information, limited retention and rights for the people concerned.

A qualification grid built around the customer's goals.

Popularised by HubSpot, this grid starts from the customer's goals and plans before budget and decision, then has them weigh what happens if nothing changes. It suits sales where the need is not yet stated when the conversation starts.

The kind of company that gets the most value from your offer.

An ICP describes a company, not a person: industry, size, organisation, buying context. It is built from the customers who sign, stay and grow, and it decides where prospecting should focus. A persona, in turn, describes the people inside those companies.

Customers who come to you.

Content, search, referrals, trade shows: inbound attracts prospects who reach out themselves. They often arrive with a need already stated, which changes how discovery should be run.

A strategic customer managed in a dedicated way.

A key account weighs heavily in revenue or reputation, involves many contacts and justifies its own account plan. The challenge is keeping an overall view of what is said with each of the customer's teams, often spread across several people on your side.

A person or company that could become a customer.

A lead can come from a form, a trade show, a referral or prospecting. It is only a starting point: qualification will tell whether it becomes an opportunity.

Scoring leads to know which to handle first.

The score combines fit with the ICP (size, industry, role) and engagement (visits, downloads, replies). It should be recalibrated regularly against the leads that actually signed.

The legal justification for processing data.

The GDPR provides six, including consent, performance of a contract and legitimate interest. Each processing must have one, chosen before it starts and stated to the people concerned.

The kind of AI model that reads and writes text.

An LLM generates text from an instruction: summary, answer, classification, drafting. It can also produce false but plausible statements. Tying it to specific sources, and asking it to cite them, makes its answers checkable.

A standard to connect an AI assistant to your tools.

MCP lets an assistant such as Claude or ChatGPT query a data source directly, with the user's permissions. Applied to sales conversations, it lets you ask your usual assistant a question and get an answer grounded in real customer conversations.

A qualification method for complex sales.

MEDDIC checks six points: the expected measurable gains, the person who signs, the selection criteria, the decision process, the pain to solve and the internal champion. Each point rests on what the customer said, not on the rep's intuition. Conversation records are therefore its raw material.

MEDDIC plus the contract path and the competition.

MEDDPICC adds two points to MEDDIC: the administrative path to signature (legal, procurement, security) and the competitors in the running. Both additions address deals lost late in the cycle, when everything looked settled but the contract stalled or another solution moved ahead.

An automated attendee that joins the video call to record it.

The bot shows up in the attendee list and records the meeting from the video platform. Other tools capture audio from the browser or the computer, with no attendee added. Either way, participants must be informed of the recording.

The written record of what was said and decided.

Useful notes fit in a few lines: the context, what the customer said, what was decided, who does what and by when. Filed in the CRM, they let a colleague pick up the account without replaying the conversation.

A lead marketing considers engaged enough to pass on.

An MQL has shown measurable interest: a download, a sign-up, repeated visits. The definition must be shared with sales, or the two teams are not talking about the same leads.

One month of subscription revenue.

MRR is the monthly version of ARR. It is often split into new MRR, expansion MRR, and MRR lost to cancellations and downgrades, to see where growth comes from and where it leaks.

A calendar shared with the customer up to signature.

A mutual action plan lists the remaining steps, on both sides, with dates and owners: technical validation, legal review, committee presentation. Built with the customer, it makes delays visible and gives a signing date grounded in facts.

Adjusting price, scope and terms until there is an agreement.

Negotiation covers price, but also term, scope, timelines, payment terms or support. Each concession is best traded for something in return, and written down: what was promised here must be honoured after signature.

The dated action a conversation ends on.

A useful next step has a date, an owner and a content: "demo to the finance team on the 12th", not "let's keep in touch". A deal with no dated next step is often a deal that stops without anyone deciding it.

A deal lost to inaction, not to a competitor.

The customer picks no one and keeps their current situation. It happens when the pain is not strong enough, when the risk of change seems too high or when nobody carries the project internally. It is an outcome in its own right, to analyse like a loss.

A measure of how likely customers are to recommend you.

NPS asks one question scored from 0 to 10 and subtracts the share of detractors (0 to 6) from the share of promoters (9 and 10). It shows a trend but not its cause: the free comment and the conversations with the customer tell you why.

What a cohort's revenue becomes after a year.

NRR compares the recurring revenue of a group of customers a year apart, including expansions and minus cancellations and downgrades. Above 100%, the existing base grows on its own, without a single new customer.

A reservation the customer voices before committing.

Price, timing, an incumbent competitor, the risk of change, lack of resources. An objection is not a no: it tells you what is left to address. Grouped across all conversations, objections show what the market really holds against an offer.

A customer's first weeks, up to first value.

Onboarding covers setup, training and first use. It succeeds when it aims at the goals the customer stated during the sale, not at a generic list of features to switch on.

Customers you go out and find.

Outbound starts from the company: calls, emails and messages to chosen accounts. It lets you target the ICP precisely, at the cost of a harder first contact than inbound.

The concrete problem that pushes a customer to change.

Lost time, a cost, a risk, team frustration. The pain is what makes a project a priority over all the others. Recorded in the customer's exact words, it shapes the proposal, prepares the demo and later tells you why the customer signed.

The administrative path of a contract, from agreement to signature.

Legal review, security questionnaire, vendor onboarding, purchase order: the paper process often starts after the agreement in principle and can take weeks. Anticipating it during qualification keeps a deal won on substance from getting lost in delays.

All open opportunities, by stage.

The pipeline shows how many deals are open, at which stage and for what value. It is only as good as the stages declared: a deal sitting in "proposal" for three months with no conversation is not really in the pipeline any more.

Pipeline value relative to the target.

A coverage of three means the pipeline is worth three times the period's target. The right level depends on the team's actual conversion rate: the lower it is, the higher the coverage needed.

The regular meeting where the team goes through deals.

Manager and reps check, deal by deal, the real stage, the next action and the risks. It is shorter and more accurate when everyone brings what the customer actually said rather than an impression.

A framed trial to prove the solution delivers.

A POC tests the solution on a limited scope, with success criteria set in advance. Without those written criteria it drags on and proves nothing. Well framed, it turns a sales promise into a result the customer saw.

The place your offer holds in customers' minds.

Positioning says who the offer is for, against which alternatives and why it is different. It is verified in how customers describe and compare you during conversations.

What you draw from a conversation once it is over.

Summary, objections, commitments, next step, competitors mentioned, deal progress: post-call analysis turns a conversation into information filed under the right customer. It has time to draw on the account's whole history, which live analysis cannot.

The document that sets out the offer, its scope and its price.

A convincing proposal restates the customer's stakes in their own words, describes what will be delivered and prices the investment. It often travels without the rep, to people absent from the conversations, so it must stand on its own.

Finding and contacting new potential customers.

Prospecting covers choosing the accounts to target, finding the right contacts and making the first contact, by phone, email, social network or event. It works better the closer the targets are to the customers who already signed.

The quarterly meeting with a customer on results achieved.

A QBR reviews what was achieved, the difficulties met and the goals ahead. It is prepared by rereading the quarter's conversations, to come back to what the customer asked for and show what was done about it.

Checking that an opportunity deserves the time it will take.

Qualifying means testing an opportunity against a grid (need, budget, decision, timeline, competition) to decide whether to pursue it, park it or drop it. Reliable qualification rests on what was said in the conversations, not on the impression left by the last meeting.

What a rep or team must sign in a period.

A quota is expressed in revenue, number of deals or ARR. Set without looking at the real pipeline and the historical win rate, it encourages inflated forecasts rather than accurate ones.

Find the right passages first, then write the answer.

With RAG, the system first retrieves the relevant excerpts from a base, then asks the model to answer from those excerpts only. The answer can then cite its sources, and what the base does not hold stays unanswered rather than invented.

Hints shown to the rep during the call.

The tool listens to the conversation and suggests, as a topic comes up, an answer to an objection, a competitor card or a forgotten question. Useful when suggestions arrive fast and stay discreet, it can also pull attention away from the customer.

Extending a contract that reaches its end date.

A renewal is decided well before its date: on the value delivered, the quality of the relationship and what the customer said through the year. It is a sale in its own right, with its objections and sometimes a competitive bid.

How long a piece of data may be kept.

The GDPR asks that data be kept only as long as its purpose requires. For recordings and conversation records, this period is set in advance, stated to the people concerned and actually applied.

Combining conversations, CRM and activity to steer revenue.

Revenue intelligence extends conversation analysis to forecasting and pipeline tracking: it ties what is said with customers to the state of the deals. These platforms mostly target large sales teams and their managers.

The function that aligns tools, data and processes for revenue teams.

RevOps brings together marketing, sales and customer success operations: CRM, stage definitions, reporting, tools. Its job is that everyone reads the same numbers with the same definitions.

Helping a rep improve from real conversations.

Coaching rests on specific moments: a poorly handled objection, a discovery that was too short, a forgotten next step. Replaying the passage together beats general advice. It remains support, distinct from assessing people.

The time and steps between first contact and signature.

Its length varies with price, the number of decision makers and the risk the customer perceives. Measuring it by segment, and spotting at which stage deals slow down, shows where to focus effort.

Giving reps what they need to sell.

Training, content, competitor cards, objection answers, onboarding new hires: sales enablement prepares the teams. It is all the more accurate when it starts from real conversations, what customers ask and what makes them hesitate.

The tool that runs email and call sequences.

These platforms schedule outreach sequences, track replies and remind reps of tasks. They handle the volume of first contacts, while conversation analysis sheds light on what happens once the conversation has started.

The estimate of what will be signed in a period.

A forecast combines reps' judgement, pipeline stages and conversion history. It is mostly wrong on deals whose decision process is unknown or whose last conversation is long past.

What the team knows about its customers, organised and searchable.

It brings together each customer's conversations, objections, commitments and decisions, instead of leaving them in people's memories. A good one answers questions ("who mentioned this competitor this quarter?") and links back to the original passages.

A shared framework to run and assess a sale.

MEDDIC, SPIN, Challenger or BANT give a team a shared vocabulary and a list of points to cover. Their value depends on how they are actually applied: checking, conversation by conversation, which criteria were covered and which are still missing.

A sales team's guide to good practice.

A playbook gathers the targets, the cycle stages, the discovery questions, the answers to objections and the competitor cards. It ages fast: updating it from what customers actually say keeps it useful.

An outline of what to say at a given stage.

A script is a support for a first call, a discovery or an answer to an objection. Read word for word, it shows. It works better as a list of questions and key messages than as a text to recite.

Where a deal stands in the cycle.

Discovery, demo, proposal, negotiation, signature: each stage should have an observable exit criterion, such as "the decision maker attended the demo". Without it, stages reflect everyone's optimism more than real progress.

Passing a signed customer to the team that will support them.

At handover, the customer success team needs to know why the customer signed, what was promised, who their contacts are and which reservations they voiced. A handover done from memory makes the customer answer questions they already answered.

The revenue the pipeline produces per unit of time.

It is the number of opportunities times the average deal value times the win rate, divided by the cycle length. It shows which lever, more opportunities, bigger deals, a better win rate or a shorter cycle, weighs most.

Searching by meaning rather than exact words.

Semantic search finds "that's too expensive for us" when you look for price objections, even without the word "price". Combined with keyword search, it covers both paraphrases and proper names.

Estimating whether a text sounds positive, neutral or negative.

Applied to sales conversations, it gives a sense of mood, but it reads irony, surface politeness and industry codes poorly. An exact customer quote is usually more telling than a sentiment score.

A deal whose signing date keeps moving back.

A deal slips when it was expected in one period and moves to the next. One slip happens; repeated slips often signal a project that is no longer a priority for the customer, or a decision maker who was never met.

A four step questioning method.

Described by Neil Rackham, SPIN chains four kinds of questions: about the current situation, the problems met, their consequences, then the value of a solution. The idea is to let the customer state the stakes in their own words rather than having them presented.

A lead sales has accepted to work on.

An SQL has passed a first sales filter, often a qualification conversation, and justifies opening an opportunity. The MQL to SQL conversion is one of the most watched indicators of marketing and sales alignment.

Anyone at the customer the project concerns.

User, manager, buyer, lawyer, security lead: each stakeholder has their expectations and can move the project forward or block it. Recording who was met, in which role and what they said avoids asking the same questions twice and overlooking a voice that matters.

A provider your vendor hands part of the processing to.

Hosting provider, transcription service, AI model provider: each sub-processor sees part of the data. Their list, and the country where they process the data, are part of what to ask before choosing a tool.

How much of the time the rep talks versus the customer.

This metric compares talking time on each side. In discovery, a rep who talks most of the time learns little. The figure alone does not say whether the conversation went well: it is an invitation to reread what was said.

Automatically spotting a topic across all conversations.

A topic tracker flags each passage where a subject appears: a competitor, a price objection, a requested feature. Keyword based, it misses paraphrases. Meaning based, it also finds what is said differently.

Selling more of the same product to an existing customer.

More users, a higher plan, more volume. The opportunity often shows in account conversations: a new team interested, a limit reached, a growing need.

What the customer gets, and why from you rather than elsewhere.

A clear value proposition ties a specific problem to a result. It is tested in conversations: if customers restate it in their own words, it lands. If they talk about something else, it needs rework.

A customer's exact words, quoted as they were said.

A verbatim is neither rephrased nor summarised. It gives weight to an analysis, helps write a pitch in the customer's own language and lets anyone check where a conclusion came from.

What customers say about their needs and your offer.

Voice of customer gathers what customers express, in surveys but above all in sales and account conversations. Used by product, marketing and sales teams, it is only worth something if it stays tied to the quotes it came from.

The share of deals won among those that reached an outcome.

It is usually computed as deals won over deals won and lost. Whether deals with no decision are counted changes the result a lot, so say which one you use.

Understanding why deals are won or lost.

It compares win and loss reasons at scale: price, a missing feature, a competitor, timing, no decision. It is more reliable when it starts from what customers said during the cycle than from explanations given after the fact.