How companies actually buy: a B2B sales course in ten modules
First published
Updated
10 modules · templates, worked numbers, self-checks
A company does not buy the way a person buys. A committee justifies a decision its members are slightly afraid of making, and the seller's real job is lowering their perceived risk. This note starts with the survey evidence on how buying groups actually behave, then gives you the complete course: ten modules with email templates, qualification frameworks compared honestly, worked pipeline numbers and a self-check at the end of every module.
A survey of disagreement
Gartner's May 7, 2025 release reports that 74% of B2B buying teams demonstrated what it calls unhealthy conflict during the decision process. The figure is a survey finding from a particular research organization, not a census of every business purchase. [1]
The survey covered 632 B2B buyers in August and September 2024. The fieldwork date belongs beside the headline number. Calling it a 2026 measurement because an article is published in 2026 would misstate when the observations were collected. [1]
Gartner defines unhealthy conflict through conflicting objectives, disagreement about the best course or being overruled by external decision-makers. That is more specific than saying buyers dislike one another. The definition is part of what the percentage measures. [1]
The release describes buying groups of five to 16 people across as many as four functions. These reported group sizes explain the relevance of consensus, but they should not be treated as a mandatory structure for every procurement decision. [1]
An organization is not a single preference
A seller can speak to one interested person without knowing whether the wider buying group agrees. A conversation, a qualification step and a purchase commitment are different stages.
Conflicting objectives need not imply irrationality. As an analytical example, one function might focus on operating value while another examines implementation burdens. This example is not a measured distribution in the survey; it illustrates why individual enthusiasm and organizational agreement can diverge.
An external decision-maker can also change a team's apparent conclusion. Gartner includes that circumstance in its conflict definition. A report should therefore distinguish internal debate from an approval process controlled elsewhere in the organization. [1]
A stalled deal does not establish which kind of disagreement occurred. A salesperson's account, a buyer interview and a survey category provide different kinds of evidence. An article should not diagnose a specific company's motives without its own reporting.
What consensus was associated with
In the survey, groups reaching consensus were 2.5 times more likely to report a high-quality deal. The wording report matters. It describes the surveyed buyers' account of deal quality, not an independently audited profit or productivity result for every transaction. [1]
The association does not prove that forcing agreement causes a better purchase. Groups might differ in other ways, and a public press release does not supply the full methodological detail needed to resolve those alternatives. The result is informative without becoming a causal law.
Gartner also reports that buyers experiencing group relevance were three times more likely to report a high-quality deal. That is another survey relationship. It is not a forecast that a particular piece of sales content will triple a particular company's results. [1]
A useful article therefore preserves the outcome definition and the unit of analysis. Buyers, groups, deals and organizations are not the same denominator. Switching between them can make a survey statistic sound more general than it is.
Personalization can point in different directions
The release distinguishes content relevant to the group from content relevant to one individual. It argues that group-level relevance can help members understand one another's viewpoints. That distinction complicates a simple claim that more personalization is always better. [1]
Gartner reports a positive 20% impact on consensus for group-relevant tailoring and a negative 59% impact for individual-level relevance. The release does not provide enough public model detail here to translate those into a promised percentage-point change for a real sales team. [1]
Its explanation is that individually tailored content can reinforce existing perspectives and confirmation bias. This is Gartner's interpretation of the research. A reported article should attribute the explanation instead of describing it as independently observed in every buyer group. [1]
The guide below covers much more than consensus: building lists, sales steps and the path to a close. Keeping it intact places the survey in context rather than presenting a single statistic as the whole practice of B2B sales.
Evidence for a fuller article
A stronger reported expansion would add buyer interviews and documented decision processes. It would ask what objectives differed, who held approval authority and how the final outcome was evaluated. Those are reporting questions.
It would also distinguish a healthy challenge to assumptions from conflict that prevents a decision. The survey's term unhealthy should not imply that every objection is a problem. Disagreement can surface costs or implementation risks that enthusiastic advocates overlook.
The source is both the producer of the research and a commercial research business. That provenance does not invalidate its findings, but it belongs in the evidence assessment. The public release gives selected results, not the complete dataset or questionnaire.
The defensible conclusion is that a buying group introduces an agreement problem separate from individual persuasion. The survey makes that problem visible; the complete guide below remains the wider material. No universal close-rate prediction or automatic-sales recommendation is inferred.
The complete course
How to use this course
Each module stands alone, but they are sequenced the way a deal is: who to sell to, how to reach them, how to qualify, how the account decides, how money moves, and how the whole thing becomes a system instead of an improvisation. Frameworks (BANT, SPIN, MEDDIC, Challenger) are presented with their origins and their honest failure modes, because each one is a tool for a specific situation, not a religion. Benchmarks are sourced where they exist and labeled as rules of thumb where they do not; your own funnel data replaces every benchmark here within a quarter or two of disciplined measurement.
- 1. How B2B buying actually works - the mental model everything else hangs on
- 2. ICP and list-building - who to sell to, and how to build the list
- 3. Outbound that earns a reply - email, LinkedIn, cold calls, sequencing, deliverability
- 4. Discovery - BANT, SPIN, and MEDDIC compared honestly
- 5. Multi-threading - how a committee actually decides, and how to sell the whole account
- 6. Selling into pharma, biotech, and CROs - compliance gates, procurement, long cycles
- 7. Pricing and negotiation - anchors, concessions, and the procurement endgame
- 8. Pipeline math - conversion benchmarks, coverage, velocity
- 9. CRM and forecast discipline - the operating system of a sales motion
- 10. Founder-led sales to first team - making the sale repeatable, then handing it off
Module 1 · How B2B buying actually works
In consumer sales you convince a person to want something. In B2B you help a group of people justify a decision they are already slightly afraid of making. That difference drives everything else in this course.
Three facts anchor the model. First, the buyer is a committee: Gartner's B2B buying research puts the typical enterprise buying group at roughly 5 to 11 stakeholders spanning about 5 distinct functions, and 77% of buyers in their survey described their last purchase as very complex or difficult. Second, the most common outcome is not losing to a competitor but losing to nothing: research behind The JOLT Effect (Dixon and Adamson, drawing on 2.5 million recorded sales conversations) attributes 40-60% of lost deals to customer indecision - the deal dies because the buyer's organization cannot get itself to act, not because a rival won. Third, your real competitor is therefore the status quo, and the status quo's greatest ally is the buyer's fear of messing up: nobody gets fired for not buying, people do get blamed for a purchase that fails.
So the seller's job is not persuasion in the consumer sense. It is de-risking. Every artifact you produce - the first email, the demo, the business case, the security documentation, the reference call - exists to lower one of the buyer's perceived risks: the risk that the product does not work, the risk that implementation fails, the risk that the champion looks foolish internally, the risk that the money is wasted, the risk that a regulator or auditor objects later.
The question running through every buying committee is not "is this good?" but "what happens to me if this goes wrong?" Sell to that question.
A useful corollary: buying is a job people do alongside their actual job, usually resentfully. The person you email did not wake up wanting to evaluate vendors. Anything that reduces the effort of buying - clear next steps, materials they can forward internally, a point of view on their problem - is worth more than anything that increases the volume of your persuasion.
The informal group whose consent a purchase requires. Users, budget owners, technical evaluators, security, procurement, and an executive sponsor. Rarely listed anywhere; you discover it.
The prospect's current way of handling the problem, including doing nothing. Your real competitor in most deals.
A deal that ends with the prospect buying nothing. Usually a consensus or risk problem, not a product problem.
Fear of messing up: the buyer's personal career risk attached to a failed purchase. The JOLT Effect research names this as the dominant driver of indecision.
Someone inside the account who wants the deal to happen and will sell for you when you are not in the room. Module 5 is largely about them.
Capture: buyers already looking (inbound, RFPs). Creation: buyers with the problem who are not looking yet (most outbound). Early-stage companies live on creation.
- Treating the deal as a conversation with one person. One enthusiastic contact is not a deal; it is a lead with a spokesperson.
- Selling benefits when the buyer is weighing risk. Features answer "what does it do"; buyers are asking "what happens if it fails".
- Reading politeness as progress. Committees are polite; momentum is measured in actions they take (meetings with colleagues, data shared, security review started), not tone.
- Assuming silence means a competitor won. Most silence is internal drift, which you can influence with structure and deadlines.
- Can you explain why "we lost on price" is usually a misdiagnosis of a no-decision loss?
- Can you name the personal risk your champion takes on by advocating for you, and one thing you can give them to reduce it?
- Can you restate your product's value as three risks it removes for the buyer, rather than three features it has?
Module 2 · ICP and list-building
Outbound fails at the list, not the email. A mediocre message to exactly the right fifty accounts outperforms a brilliant message to five thousand loosely-matching ones. The ICP (ideal customer profile) is the filter that decides who earns a place on the list.
Building the ICP
An ICP has three layers. Firmographics: industry, size, geography, business model (for a clinical-trials data platform: mid-size CROs, or biotech sponsors running 2+ concurrent trials). Triggers: events that make the problem urgent now - a new study starting up, a submission deadline, a funding round, a VP of data management just hired, an FDA information request. Exclusions: who looks like a fit but is not - too small to have the problem, locked into a competitor contract, building in-house. Exclusions are what make an ICP useful; a profile that excludes nothing filters nothing.
With existing customers, derive the ICP by pattern-matching your best deals: fastest closes, highest retention, loudest champions. With no customers yet, derive it from the problem: who feels the pain most acutely, who has budget for it, and who can act fastest. Write it as a testable hypothesis ("mid-size CROs running oncology trials, where data management is the bottleneck on database lock") and expect to revise it every quarter from win/loss evidence.
Building the list
Sources, roughly in order of signal quality: your own traffic and inbound (they raised a hand), trigger-event monitoring (clinicaltrials.gov registrations, funding announcements, job postings, conference agendas), curated databases (Sales Navigator, Apollo, Crunchbase), and purchased lists (lowest quality, treat as raw ore). For clinical trials specifically, clinicaltrials.gov is an underused goldmine: every registered study names the sponsor and often the CRO, with phase, indication, and timelines - that is a targeting database your competitors' generic tools cannot replicate.
Then find the people: typically 2-4 roles per account (the economic buyer, the operational owner of the pain, a technical evaluator). Research on cold outreach consistently shows multi-contacting an account outperforms single-contacting - one large study of millions of cold emails found that reaching several people at one company meaningfully lifted reply rates, because you are increasing the chance of hitting the person who actually owns the problem.
Data quality is a deliverability issue, not just a targeting issue: bought or scraped emails bounce, bounces burn your sending domain, and a burned domain means even your good emails land in spam. Verify addresses before sending, cap bounce rates (keep them under roughly 3-5%), and treat list hygiene as weekly maintenance.
Ideal customer profile: the firmographic and situational description of accounts worth pursuing. Filters the list before any messaging happens.
The role-level description of the people inside an ICP account you will actually talk to (their goals, fears, vocabulary). ICP picks the company; persona picks the humans.
Something that changed at the account making the problem urgent: funding, a new study, a regulatory deadline, a new executive. The difference between "relevant someday" and "relevant this quarter".
Total addressable, serviceable addressable, and serviceable obtainable market. Useful for sizing; dangerous as an excuse for a broad list.
Share of emails rejected by the receiving server. Hard bounces (bad addresses) above a few percent damage sender reputation fast.
The trust score mailbox providers assign your sending domain and IP, built from bounces, spam complaints, engagement, and authentication (SPF, DKIM, DMARC). Slow to build, fast to destroy.
- Defining the ICP by who could theoretically buy instead of who is likely to buy this quarter. Theoretical fit is how pipelines fill with zombies.
- Building a 5,000-row list before testing a 50-row one. Run a small batch, read the replies, then scale.
- One contact per account. If your one person is on leave, in a different function, or simply not feeling the pain, the account reads as dead when it was never really tried.
- Skipping exclusions. "Anyone in pharma" is not an ICP; it is a directory.
- Can you state your ICP in one sentence that includes at least one firmographic, one trigger, and one exclusion?
- Can you name three observable trigger events for your product and where you would monitor each one?
- If you pulled your current list, could you defend every row as matching the ICP, with the trigger noted?
Module 3 · Outbound that earns a reply
Outbound is interruption by definition. The recipient did not ask to hear from you, so the entire game is earning the next ten seconds of attention, three times in a row: the subject line earns the open, the first line earns the read, the ask earns the reply. Everything about craft follows from that.
What the numbers say
Benchmarks from large sending studies (Belkins and Reply across millions of B2B emails; Instantly's platform data): average cold email reply rates sit around 3-5% and have been declining year over year as inboxes crowd; average open rates hover near 28%; plain text beats templates with images and attachments (attachments and graphics roughly halve reply rates); short emails beat long ones; and 2-4 touches capture most of the replies a sequence will ever get, though well-built sequences of 6-8 touches still add incremental meetings. Treat these as calibration, not targets: tightly targeted, trigger-led campaigns to a fresh list can clear 10%, and a sub-2% reply rate on a verified list usually means a deliverability or targeting problem, not a copywriting one.
The anatomy of a cold email
- Subject line: 2-5 words, lower-case or sentence case, specific to them, no hype. "ADaM timelines at [CRO]" beats "Revolutionize your data pipeline". The goal is to look like internal mail, not marketing.
- First line: proves this email is for them alone - a real observation about their company, study, hiring, or news. If the first line could be sent to 500 people unchanged, it is not personalization, it is mail merge with extra steps.
- The bridge: one sentence connecting the observation to a problem hypothesis. "Teams in your position usually..." is honest; pretending you know their internals is not.
- The value: one sentence on what you do, framed as an outcome with proof (a number, a named-seeming case, a mechanism). Not a feature list.
- The ask: small, specific, easy to say yes to. "Worth 15 minutes?" outperforms "Let me know your thoughts". Interest-based asks ("Is this a problem on your radar?") outperform meeting asks when the buyer is senior or the brand is unknown.
Subject: ADaM timelines at Helio Research Hi Dana - saw Helio is running the Phase II oncology readout for the Meridian program. Teams managing that many SDTM-to-ADaM mapping decisions usually lose weeks to lineage reconciliation before database lock. We turn trial data into submission-ready ADaM datasets with a full audit trail, so your biostats reviewers spend their time on outputs, not provenance. One mid-size CRO cut their database-lock prep from 6 weeks to 8 days. Is faster lock prep on your radar for the current studies? Iris
Why it works: First line is about them, not you. The problem is stated as a hypothesis ('usually'), which is honest and invites correction. One outcome with one concrete number. The ask is interest-based and answerable in five seconds.
Subject: Congrats on the Series B Hi Marcus - saw the raise last week. Scaling from 4 to a dozen concurrent studies is exactly when spreadsheet-based data management starts breaking in ways that show up at the worst time (usually two weeks before a readout). We help sponsors your size get submission-ready datasets without adding headcount - audit trail included, which your future diligence teams will thank you for. Worth a quick look while the hiring plan is still flexible? Iris
Why it works: The trigger does the relevance work. It names a specific, time-bound risk (the raise creates a scaling problem now), and the ask is tied to their planning window, which gives a reason to reply this week rather than someday.
Subject: How a CRO your size cut lock prep time Hi Priya - a data management lead at a CRO your size was spending most of each study's final month reconciling dataset lineage by hand. They moved to our platform and got that month back - submission-ready ADaM with the audit trail built in. Happy to send a 2-page writeup of what they did. Want it? Iris
Why it works: A peer story about a company like theirs is the strongest social proof in a risk-averse market. The ask asks permission to send something, which is trivially easy to grant and starts a conversation without demanding calendar time from a stranger.
Subject: Closing the loop Hi Dana - I've sent a couple of notes about ADaM prep for the Meridian readout and haven't heard back, so I'll stop here. If database-lock prep is ever the bottleneck, you know where to find me. And if there's a better person on your team for this, a name is all I need. Either way, good luck with the readout. Iris
Why it works: Break-up emails routinely produce the highest reply rate of any touch. Giving explicit permission to say no is respectful, and the referral ask converts a dead end into a map of the account.
The proof points in these templates (the 6-weeks-to-8-days case, the Series B trigger) are illustrative placeholders. Swap in your own verified numbers and real observations before anything goes out - a claim you cannot back up, once sent, costs the account.
LinkedIn as a channel
LinkedIn works as a warm-up layer, not a pitch channel. The pattern that performs: follow the person, engage with one or two of their posts substantively (a comment that adds information, not "Great post!"), then connect with no note or a short one that mentions nothing about your product, then - only after they accept - a short message following the same anatomy as the email. Voice notes and short video messages get outsized reply rates precisely because they cannot be automated well; they signal a human spent effort. Pitching in the connection request is the fastest way to get ignored and reported.
Cold calls
Cold calls are not dead; they are just filtered. In compliance-heavy industries many senior people will never answer an unknown number, but operations and data-management folks often do. The call's only goal is to sell the next step, never the product. A permission-based opener consistently outperforms fake familiarity:
"Hi Dana, this is Iris from Yesod. I know I'm interrupting - can I take 30 seconds to tell you why I called, and you tell me whether it makes sense to keep talking?" [If yes] "We help CROs get submission-ready ADaM datasets without the manual lineage work. I saw you're running the Meridian oncology program and guessed lock prep might be a pain point. Is that anywhere near your world?"
Why it works: Asking permission respects the interruption and gets a micro-yes. The 30-second promise is kept, which earns the next minute. The question at the end is about their world, not your product.
Sequencing
Single touches lose to sequences. A working default: 6-8 touches over 3-4 weeks, mixing channels, each touch adding something new (a case, a number, a question) rather than "bumping" the previous email. A sample cadence:
Deliverability: the plumbing that decides everything
None of the above matters if you land in spam. The rules: authenticate the domain (SPF, DKIM, DMARC) before anything else; send cold outreach from a separate domain or subdomain so a burned reputation never touches your corporate mail; warm up new domains over 2-4 weeks with low volume; cap daily sends per mailbox (roughly 20-50 cold emails per mailbox per day is a sane ceiling); verify every address; never buy a list and blast it; keep plain text (no tracking-pixel-heavy HTML templates, no attachments in cold mail). Watch bounce rate, spam complaint rate, and reply rate weekly - they are the vital signs.
- Personalization theater: "I loved your recent post!" with no substance is worse than no personalization, because it advertises that you are faking relevance.
- Bumping: "Following up" as the entire content of touches 2 through 6 trains the recipient to ignore you. Every touch must carry new value.
- Pitching in the LinkedIn connection request.
- One email, no reply, account marked dead. Most first replies come on touch 2-4; the break-up email converts a surprising tail.
- Blasting 500/day from a fresh domain and concluding outbound does not work when you actually concluded that spam folders exist.
- Can you write a first line for your top ten accounts that could not be sent to anyone else?
- Can you state your sequence's touches, channels, and spacing from memory, and what each touch adds?
- Do you know your current reply rate, bounce rate, and sends-per-mailbox-per-day? If not, that is the gap to close first.
Module 4 · Discovery: BANT, SPIN, MEDDIC compared honestly
Discovery has two jobs that most reps collapse into one. Job one: decide whether this deal deserves your time - disqualification is the goal, and a fast no is a good outcome. Job two: learn how this specific account would actually buy, so you can help them do it. The frameworks below are lenses for those two jobs, not scripts.
The three frameworks
BANT (Budget, Authority, Need, Timeline) came out of IBM in the 1960s as a lead-scoring shorthand. It is fast and crude: does money exist, can this person spend it, is the pain real, is there a date. Its flaw is that modern budgets are created for good projects rather than pre-allocated, and "do you have budget?" asked early mostly teaches the buyer to lie or leave. Use BANT thinking at the top of funnel; do not use it as a meeting agenda.
SPIN (Situation, Problem, Implication, Need-payoff) came from Neil Rackham's research at Huthwaite in the 1980s - 35,000 observed sales calls, the finding that in large sales, the questions that matter are the ones that make the buyer articulate the cost of their problem in their own words. Situation questions gather facts (keep them few; you should have done homework). Problem questions surface the pain. Implication questions make it expensive: "when lock prep slips two weeks, what does that do to the readout timeline and the team?" Need-payoff questions let the buyer state the value of solving it: "what would it be worth to get that month back?" SPIN's insight is durable: buyers believe their own conclusions more than yours.
MEDDIC (Metrics, Economic buyer, Decision criteria, Decision process, Identify pain, Champion - later extended to MEDDPICC with Paper process and Competition) came out of PTC in the 1990s and is the closest thing enterprise sales has to a standard. Its core idea: a deal is real only to the extent you can produce evidence for each letter. It is less a questioning technique than a deal inspection checklist.
The quantified outcome the buyer will be judged on. "Cut lock prep from 6 weeks to 2" - a number the champion can put in front of their boss, not a vibe.
The person who can actually release the money. Have you met them? If everything you know comes from someone two levels down, you do not have an E yet.
The written or unwritten scorecard: price, validation burden, integration, audit trail. Ask what they will compare, and what would disqualify you.
The steps, people, and dates between now and signature, including legal, security, and quality review. Get it in their words, then verify it against reality.
The specific operational wound, sized. Not "data is hard" but "two FTEs for six weeks per study on lineage reconciliation, with an audit finding last year".
A person with power and a personal win attached to your success, who sells for you internally and passes the tests in Module 5.
MSA, SOW, security review, vendor qualification, procurement portal. In pharma this alone can take a quarter; knowing it early is the difference between a forecast and a fantasy.
Who else is being evaluated, including "do nothing" and "build it ourselves". If you do not know, assume the status quo is winning.
An honest comparison
| BANT | SPIN | MEDDIC | |
|---|---|---|---|
| Origin | IBM, 1960s | Rackham / Huthwaite research, 1988 | PTC, 1990s |
| What it is | Qualification shorthand | Questioning method for building pain | Deal inspection standard |
| Best at | Fast top-of-funnel triage | Discovery conversations in complex sales | Forecasting enterprise deals honestly |
| Failure mode | Interrogation about budget that buyers dodge | Endless discovery with no qualification spine | Checklist theater: fields filled, evidence absent |
| Use it when | Inbound volume needs fast sorting | Running the discovery call itself | Any deal big enough to forecast |
Running the call
A good discovery call is 70% them talking. Open with an agenda and permission to be honest in both directions ("if this is not a fit, I will tell you and stop"). Ask about current process before product: how they do it today, what that costs, what happens when it breaks, what they have already tried, what happens if nothing changes. Listen for numbers and dates - a buyer who can quantify their pain is a buyer who can build an internal case. End every call with a concrete, dated next step that requires something from them (a colleague introduced, data shared, a review meeting scheduled). A next step that requires nothing from the buyer is not a next step.
The ongoing decision of whether a deal deserves your time. Done continuously, not once at the top.
A SPIN question that enlarges the stated problem into its business consequences, so the buyer feels the cost of inaction.
A SPIN question that has the buyer state the value of solving the problem in their own words.
The MEDDIC discipline: a field is "known" only if the buyer said it, in a verifiable way. Rep assumptions do not count.
Deliberately ending pursuit of a poor-fit deal. Frees time for winnable ones; a pipeline without disqualifies is a wish list.
- Checklist interrogation: firing BANT questions in order feels like a deposition and teaches buyers to give defensive answers.
- Accepting metrics without proof. "We would save a lot of time" is not a metric; "two FTEs for six weeks per study" is.
- Guessing the decision process from the last deal you closed. Every account's paper process is different; ask, then verify.
- Discovery that never ends. At some point you must propose; qualification that stretches past the third meeting is usually avoidance.
- Treating "no budget" as fatal in month one and "we have budget" as sufficient. Both are usually wrong.
- For your most advanced deal, can you fill every MEDDIC letter with buyer-stated evidence, and name which letters are still assumptions?
- Can you write three implication questions and three need-payoff questions specific to your product's pain?
- Can you explain when BANT is the right tool and why it fails as a discovery-call structure?
Module 5 · Multi-threading: how the account actually decides
Module 1 established that the buyer is a committee. Multi-threading is the practice of selling to all of it, deliberately, instead of relying on one contact to carry your case. Single-threaded deals are fragile in two directions: your one contact can leave, get overruled, or lose interest, and - more subtly - one person, however enthusiastic, cannot manufacture consensus among five to eleven stakeholders who each hold a veto of a different kind.
Who is on the committee
Has influence, has a personal win tied to your success, and sells for you in rooms you are not in. Test them: do they share internal information, do they arrange meetings, do they act on agreed steps? A contact who likes you but cannot move the org is a coach, not a champion - friendly, useful for intelligence, useless for closing.
Can say yes when everyone else says no, and vice versa. Usually a VP or above with budget authority. You need direct access to them before the deal can be real; a champion who will not broker that meeting is telling you something.
Security, IT, data engineering, QA. They cannot say yes, but each can say no. In regulated industries their veto is the strongest. Engage early, with documentation, before they feel surprised.
Enter late, measured on savings and risk terms, not on your product. Module 7 covers them. The mistake is meeting them for the first time when you think the deal is done.
The data managers who will live in the product daily. They rarely control budget, but their resistance surfaces in evaluation notes and pilots. Their enthusiasm is your champion's ammunition.
Anyone inside (or adjacent to) the account who tells you how things really work: who hates whom, what killed the last vendor, when the budget cycle turns. Coaches do not need to want you to win; they need to like you and like being right.
How to multi-thread without going around your champion
- Ask for the room, not around it. "Who else will weigh in on this? It would help me to hear their concerns directly - could we set up 30 minutes with you and [the QA lead] together?" Framed as making the champion's internal case easier, this is a gift, not a bypass.
- Give the champion a weapon, not a burden. A one-page internal brief they can forward: the problem in their company's words, the quantified outcome, the answer to the three objections you know are coming, the price and timeline. Most deals are sold in meetings you never attend; write for that meeting.
- Use the mutual action plan. A shared, dated list of every step to signature: evaluation, security review, vendor qualification, legal, signature, go-live. Co-authoring it with the champion surfaces the hidden steps ("oh, procurement will need a Coupa registration") and turns drift into visible slippage you can discuss.
- Escalate with cover. When the thread is stuck below the economic buyer, offer the champion air cover: "Would it help if our CEO reached out to your VP directly, peer to peer?" This flatters everyone and moves the deal upward with permission.
You do not close deals. Champions close deals, using materials you built, in meetings you were not invited to. Your job is to arm them and to make sure every veto-holder met you before they met your contract.
- Confusing a friendly contact with a champion. Test for action, not warmth.
- Going over your champion's head silently. You might win the meeting and lose the account's trust.
- Treating the technical gatekeepers as obstacles to route around. In regulated industries they are the decision.
- Discovering the paper process after verbal yes. Ask "what happens between you saying yes and a signed contract?" in the second meeting, not the tenth.
- For your biggest active deal, can you name every committee member, their stance, and the last date you spoke to each?
- Does your champion have a forwardable one-pager written for their internal audience, in their vocabulary?
- Is there a dated mutual action plan both sides have seen and edited?
Module 6 · Selling into pharma, biotech, and CROs
Everything in the previous modules applies here with the dial turned up. Life-science buying is ordinary enterprise buying plus three multipliers: institutional risk aversion backed by regulation, a validation burden that lands on vendors, and timelines governed by study milestones rather than fiscal quarters.
Why this market is different
The regulatory frame (GxP, and in the US 21 CFR Part 11 for electronic records and signatures) means your buyer is not just choosing software; they are accepting a system whose failure could surface in an FDA inspection or a submission deficiency letter. That is why their first questions are about audit trails, validation, access controls, and data lineage, and why "we move fast" is a terrifying pitch in this market. Speed is a benefit only after trust in the controls. Reframe: "submission-ready with a full audit trail, fast" - the compliance property is the headline, speed is the bonus it buys.
The validation burden is real work for the buyer: under GxP, a computerized system used in regulated activities must be validated for its intended use, and the sponsor remains responsible even when the vendor supplies documentation. This is why vendors with a ready validation package (IQ/OQ/PQ documentation, test evidence, release notes discipline, a Part 11 compliance matrix) get shortlisted and vendors who say "yes we\'re compliant" without artifacts get qualified out.
How the buying process runs
- Vendor qualification: before or alongside the commercial discussion, expect quality questionnaires, a possible audit (sometimes remote), security review, and a Part 11 assessment. These are gates with pass/fail outcomes, run by QA and IT - people you may never meet unless you ask. Have a qualification pack ready to send the day it is requested.
- The paper process: MSA plus per-study SOWs is the norm with CROs and sponsors; add confidentiality terms, data-processing terms, liability caps, insurance certificates, and often a procurement portal registration (Ariba, Coupa). From verbal yes to signature routinely takes 6-12 weeks in this market. Build it into every forecast.
- Timeline drivers: study milestones - protocol finalization, first patient in, interim readout, database lock, submission - create natural urgency windows that fiscal calendars do not. A vendor that arrives six weeks before database lock with a credible way to compress lock prep is solving a live emergency; the same vendor six months earlier is a nice-to-have.
- Budget reality: sponsors fund by program and trial budget; CROs price into bids and watch margin per study. A sponsor asks "does this reduce my risk and timeline?"; a CRO asks "does this make my bid more competitive or my delivery cheaper?". Same product, two different pitches - lead with risk and submission quality to sponsors, lead with margin and bid-win rate to CROs.
Turning the gates into the pitch
The compliance gates that slow competitors down are your differentiation if you industrialize them. Maintain a standing due-diligence room: Part 11 matrix, validation summary, security architecture, SOC 2 or equivalent if you have it, SOPs for change control and incident response, sample audit-trail exports. When a QA reviewer receives complete, inspection-grade documentation within a day of asking, the audit becomes a sales asset - you just demonstrated the exact competence they are buying. In a market where buyers are paid to be suspicious, documented competence is the whole pitch.
Land and expand
The natural deal shape in this space: pilot on one study or one therapeutic area (priced to be easy to approve, scoped to produce a measurable before/after), then expand across the portfolio at renewal. Design the pilot to generate the metrics your champion needs for the expansion case - weeks saved to lock, queries avoided, review hours cut. The pilot is not the deal; it is the evidence factory for the deal.
The family of good-practice regulations (GCP, GLP, GMP) governing regulated work. It is why "trust me" is not a sales strategy here.
The FDA rule for electronic records and signatures: audit trails, access controls, record integrity. The checklist your buyer's QA team will literally walk through.
The buyer's formal assessment of whether you are allowed to touch regulated work: questionnaires, documentation review, sometimes an audit. Pass/fail, and the QA team runs it.
Installation, operational, and performance qualification - the validation evidence a system does what it claims, in the buyer's environment, for its intended use.
Master services agreement (the standing legal terms) plus statements of work (per-study scope and price). Two documents, two clocks.
The moment a study's data is frozen for analysis. The deadline that concentrates all upstream data pain - and the urgency window for selling data tooling.
- Leading with speed. In a Part 11 world, unproven speed reads as uncontrolled. Lead with the audit trail; sell speed as what the audit trail makes safe.
- Treating QA and procurement as enemies. They are stakeholders with legitimate vetoes; vendors who equip them get through gates weeks faster.
- Being surprised by the paper process. Six to twelve weeks from yes to signature is normal here; forecasting signature next week because the champion is excited burns your credibility with your own founders.
- Piloting without a measurement plan. A pilot that ends with "it went well" instead of numbers is a missed expansion.
- Quoting consumer-style month-to-month pricing. Buyers here expect SOW-shaped commercial structures; matching that expectation removes friction.
- Could you send a complete vendor-qualification pack today, within one business day of a request?
- Can you pitch your product twice - once for a sponsor (risk, timeline) and once for a CRO (margin, bid competitiveness) - without reusing a sentence?
- Do you know the next three study milestones for each of your top five accounts, and which one creates urgency for you?
Module 7 · Pricing and negotiation
Pricing is positioning
Price communicates what kind of thing you are. A clinical data platform priced like a productivity app signals it is not built for regulated work; priced like enterprise infrastructure it invites the serious evaluation you want. Choose a value metric - the unit the price scales with - that grows with the value the customer gets: per study, per dataset, per submission. Avoid metrics that punish usage you want to encourage, and avoid per-seat pricing when seats are not where the value lives.
Anchor with structure, not a number alone. Presenting three tiers (pilot / standard / enterprise) does two jobs: it makes the middle option the natural choice, and it turns the negotiation from "how much less than X" into "which shape of engagement". Whatever you quote first frames everything after; quote from strength, with the reasoning attached.
Negotiation mechanics
- Trade, never give. Every concession purchases something: a discount buys a longer term, a case study, a faster signature, payment upfront, a reference call. "If we can do X, could you do Y?" is the only shape a discount should have. A concession given for nothing teaches procurement that your prices are opening bids.
- Know who you are negotiating with. Before commercial terms move, confirm you are talking to someone who can say yes. Discounting to a champion who then takes your new price to procurement for another round is negotiating against yourself.
- Concede slowly and in shrinking steps. 10%, then 4%, then 1.5% signals an approaching floor; three flat 10% cuts signal unlimited depth.
- Use silence. State the proposal, then stop talking. The urge to fill silence is where most self-inflicted discounts come from.
- Protect the value story. When pushed on price, return to the metric: "this pays for itself in one avoided week of lock-prep labor per study" reframes a cost discussion as a return discussion.
The procurement endgame
Procurement is a profession with its own incentives: savings against your quote, contract risk reduced, process followed. They are not the enemy; they are a buyer of a different product (terms). Expect: the first response is always that you are expensive; the request for a discount arrives attached to a deadline that is probably soft; liability caps, indemnities, IP clauses, and insurance requirements will consume more calendar than the price discussion. The counter-moves: keep the champion and economic buyer warm through the whole endgame (deals die in legal while sellers wait politely), give procurement a concession they can report as a win (budgeted for in your original quote), and hold one concession in reserve for the final week.
Everything is negotiable exactly once. Set the price so you can give something and still like the deal; whatever you give, get something dated and specific back.
- Discounting for speed at quarter-end. Buyers learn the pattern and wait for it; you train your market to never pay list.
- Negotiating against yourself: revising a proposal before the other side responds to it.
- Treating procurement's first "too expensive" as information. It is procedure. The information is in what they do next.
- Letting the relationship with the champion go cold while lawyers exchange redlines. The deal is still alive and still needs feeding.
- Can you state your value metric and why it grows with customer value, in one sentence?
- For your current quote, do you know the two concessions you would trade, and what you would ask for each?
- Do you know who at the account can actually sign - and have you heard them say yes to the shape of the deal?
Module 8 · Pipeline math
Pipeline math is what turns sales from a mood into a machine. Every stage has a conversion rate; multiply them together and the funnel tells you exactly how much input a closed deal requires - and therefore what next quarter's revenue needs you to do this week.
A worked funnel
A plausible outbound funnel for a mid-market B2B product, stated as assumptions you would replace with your own data: 1,000 verified contacts entered into sequences; a 4.5% reply rate yields 45 replies; 40% of replies convert to a held meeting (18); two-thirds of meetings qualify into real opportunities (12); a 20% close rate on qualified opportunities yields 2 to 3 customers. The point is not the specific numbers - it is that each arrow is a lever with an owner and a fix. Low replies: list or deliverability. Replies but no meetings: the ask or the value statement. Meetings but no opportunities: ICP or discovery. Opportunities but no closes: multi-threading, business case, or price-value.
Coverage and velocity
Two formulas run the operating rhythm. Pipeline coverage: open pipeline value divided by the target for the period. The common rule of thumb is 3-4x coverage - if the quarter's target is $300k, you want $900k-$1.2M of qualified pipeline, because win rates under 50% mean most pipeline will not close. Sales velocity: (number of opportunities x win rate x average contract value) / sales cycle length. It answers "how much revenue does this machine produce per month" and decomposes misses into their four levers. Improving velocity 20% can come from 20% more opportunities, a win rate from 20% to 24%, a contract size up 20%, or a cycle 17% shorter - the formula tells you which lever is cheapest to move, and cycle compression is usually the most neglected one.
Run the math backward from the goal: target revenue / ACV = deals needed; deals / close rate = opportunities needed; opportunities / meeting-to-opp rate = meetings needed; meetings / reply-to-meeting rate / reply rate = contacts to sequence. That final number, divided by working days, is the daily activity budget. When the required number exceeds what a human can do, the answer is not heroics - it is fixing a conversion rate or raising ACV.
- Counting unqualified deals as pipeline. Inflated pipeline feels safe and lies to the forecast; apply the MEDDIC evidence standard before a deal counts.
- Averaging across unlike segments. A 30% close rate that blends warm inbound with cold enterprise outbound describes neither.
- Optimizing the stage you enjoy instead of the stage that is broken. The funnel points at the bottleneck; believe it.
- Treating benchmarks as your numbers. Every rate above is a placeholder until your CRM has two quarters of honest stage conversions.
- Can you recite your funnel's stage-by-stage conversion rates from your own data (not this page)?
- Given a revenue target, can you compute the required weekly outbound activity in under five minutes?
- Do you know your current pipeline coverage multiple for next quarter?
Module 9 · CRM and forecast discipline
The CRM is not a reporting burden imposed by management; at a startup it is the company's memory and the only honest mirror the sales motion has. The test of a CRM setup: a stranger could read it and know, for every deal, what has been said, what happens next, and what could kill it.
The minimum viable setup
- Stages with exit criteria, not vibes. A deal moves to "Evaluation" when the buyer agreed to an evaluation, not when you sent materials. Write the exit criterion on each stage; forecasts made from vibe-stages are astrology.
- The next-step rule. No open opportunity exists without a dated next step involving the buyer. This single rule eliminates most pipeline rot.
- MEDDIC fields on the opportunity. Metrics, economic buyer, decision process, paper process, champion - each filled from buyer-stated evidence, each visibly empty when it is an assumption.
- Closed-lost reasons in the buyer's categories. "No decision / status quo", "lost to competitor", "no budget", "timing", "lost contact". Price and competitor get blamed by default; honest categories are how you learn that most losses were indecision (Module 1) and respond with better risk-reduction instead of discounts.
The weekly rhythm
A 45-minute weekly pipeline review, same time, covers: what closed, what slipped and why (slippage reasons are data), what has no next step (fix or close-lost it), what is stuck in a stage beyond its normal dwell time, and the forecast call for the period. Forecasting uses three buckets - commit (you will be surprised if it does not close), best case (real, but something must go right), pipeline (everything else that is qualified). The discipline is in the definitions: "commit" means the buyer has told you the yes and the paper process is in motion, not that the demo went well.
Happy ears - hearing enthusiasm as commitment - is the default failure of founder-led pipelines, because founders are optimists by selection. The antidote is the evidence standard: a forecast category is a claim about the buyer's behavior, and claims need buyer-stated evidence.
- CRM as afterthought: fields updated the night before the review produce a history of fiction.
- Zombie deals kept alive to make coverage look healthy. A deal with no buyer action in 30+ days is closed-lost until the buyer proves otherwise.
- Stage inflation: moving deals forward to feel momentum. The funnel math in Module 8 only works if stages mean something.
- Logging losses as "price" when the champion went dark. Mislabeled losses produce the wrong fixes.
- Does every open opportunity in your CRM have a dated, buyer-involving next step right now?
- Could you defend your current commit list to a skeptic using only buyer-stated evidence?
- Do your closed-lost reasons distinguish "no decision" from "lost to competitor"?
Module 10 · Founder-led sales to first team
Founders outsell every early hire, and it does not matter. Founder sales works because of things that do not transfer: authority to change the product mid-call, pattern-matching across every customer conversation ever held, and the credibility of the title. The actual job of founder-led sales is not revenue - it is discovering a repeatable sale: one ICP, one pain story, one pitch, one price shape, that a normal competent person could run. Until that exists, hiring salespeople just adds witnesses to improvisation.
Codifying the motion
The artifact that makes the sale transferable is the playbook, built from evidence rather than aspiration: recorded calls (what the founder actually says, not what they think they say), the discovery questions that produced the best deals, the three objections that recur and the answers that worked, the email sequences with their reply rates, the qualification bar a deal must clear to deserve founder time, and the closed-lost postmortems. A useful test: hand the playbook to a smart outsider and watch them run a mock call. Every place they stumble is a place the motion still lives in the founder's head.
The first sales hire
Timing: hire when the founder has closed enough deals the same way that the pattern is visible - commonly 5-15 customers, depending on deal size - and when the bottleneck is genuinely selling hours, not product gaps or positioning confusion. Profile: the first hire is a builder, not a manager - someone who has sold without brand, air cover, or marketing, who will write the playbook as much as follow it. Hiring a VP of Sales from a big company at this stage is the classic mistake: they know how to run a machine, and there is no machine yet.
Comp basics: OTE (on-target earnings) split roughly 50/50 between base and variable for a first AE, quota set at 4-5x OTE as a starting heuristic, accelerators above quota, and a ramp period (typically 2-4 months, longer for long cycles) with reduced quota expectations. In a 6-12 month sales cycle like clinical data, judge early hires on leading indicators - qualified meetings, evidence-filled opportunities, cycle hygiene - because closed revenue lags their work by two to three quarters.
What founders keep: the strategic accounts, the final pricing calls on unusual deals, and the discovery conversations that keep the product honest. Founder-led sales never fully ends at a small company; it narrows to the deals where the founder is genuinely the best seller.
- Hiring to escape selling. The first reps amplify whatever exists; if the motion is broken, they break louder.
- Expecting the playbook on day one from the hire. Codification is the founder's job; the hire's job is pressure-testing it.
- Judging a long-cycle hire on first-quarter revenue. You will fire good people and keep lucky ones.
- The founder staying in every deal. If nothing closes without you, you have consulting revenue with extra steps.
- Could a smart stranger run your next first call using only your written playbook?
- Can you name the bottleneck in your motion right now - selling hours, product gaps, or positioning clarity?
- Do you know which deals in your current pipeline genuinely require a founder, and why?
Sources and further reading
Statistics cited in this page, and the books behind the frameworks.
Frameworks: SPIN Selling (Rackham, 1988); MEDDIC (PTC, 1990s; MEDDICC, Whyte); The Challenger Sale and The JOLT Effect (Dixon and Adamson, 2011 and 2022); BANT (IBM, 1960s). Funnel conversion figures in Module 8 are illustrative assumptions, labeled as such - replace them with your own CRM data.
Compiled 28 Sep 2026. Framework attributions are historical record; benchmarks are cited to their studies above; all worked numbers are illustrative. Regulatory references (Part 11, GxP) are orientation for sellers, not legal or compliance advice.
Sources and further reading
- Gartner Sales Survey Finds 74% of B2B Buyer Teams Demonstrate Unhealthy Conflict During the Decision ProcessGartner newsroom, 7 May 2025
- Gartner B2B Buying ReportGartner
- Losing to customer indecision (The JOLT Effect)Challenger Inc.
- Part 11, Electronic Records; Electronic Signatures - Scope and ApplicationUS Food and Drug Administration
- 21 CFR Part 11eCFR