先雇两个你会买单的销售：Jason Lemkin 谈招人、薪酬和产品协同

Lemkin says only a truly self-serve product can avoid a sales team, and even Canva and Slack added one eventually. Canva waited until well past $500 million in revenue, while Slack's revenue was mostly enterprise by the time it went public. Most companies end up hybrid, and he says founders too often treat self-serve versus sales-led as an either-or choice.

The key advice is to be honest about how your first 10 to 20 customers actually arrive. If they came in with a credit card and never talked to anyone, lean into viral loops and B2C-style growth. If they said they would pay thousands if you solved a specific problem, that is not a self-serve motion, and ignoring it is a mistake. He tells of a portfolio company that fired its sales team because the founders disliked selling and fell from about $5 million in revenue to under $1 million.

On the question of who should close those first deals, Lemkin says 95 times out of 100 the founder must close the first 10 customers. Founders are what he calls "A+ middlers": even if they are uncomfortable opening or closing, they are better than most reps at the middle of the conversation, and customers love talking to the CEO. Founders then need to work up the courage to ask for next steps and for money.

Lemkin's trigger for hiring is that once more than 20% of the founder's time is booked with customers, they need leverage. He says a CEO should spend about 20% of their time on sales and 20% on recruiting. He insists on hiring two reps rather than one, because otherwise there is no way to A/B test humans.

His one cheat code is that the first reps must be people you would buy your own product from. Founders should expect to interview around 30 candidates: about 20 will be unprepared, eight okay, and one or two magicians who ask the right questions and explain the product and the customer's problem convincingly. He warns against being dazzled by resumes from Twilio or Cloudflare and acronym fluency, and looks for what he calls pirates and romantics who love the product.

He recalls his own first rep at EchoSign, who had been let go by a previous startup and was living in his brother's garage, yet described the problem better than any of 30 candidates and stayed for a decade. Early leads are precious, he says, so a polished hire who fumbles your three monthly leads can sink the company. Candidates should also have a couple of years of experience, ideally near your deal size, and enough maturity to be trusted with a customer.

Lemkin says you need two reps hitting quota before hiring a manager. A VP of sales takes you from three reps to 300 by scaling a script that is starting to work. Hiring earlier, he argues, is a Hail Mary that asks one person to find product-market fit, be the first and second rep, and scale everything at the same time. He says a company that raises money and then hires a VP of sales to fix stalled sales will often see that person leave in eight months with much of the money spent.

Early sales hires must also be product experts, he says, because a startup's product has gaps and a subtle 10x feature that good reps learn to sell. A VP who only wants to manage processes will not understand it. He also argues that today's VP of sales should still carry a bag or at least be in deals 20 to 30 hours a week when they start.

Lemkin also tells founders that before any VP arrives, they must be the "crummy first head of sales" themselves, managing reps and working in the field. Without a repeatable process he puts the odds of a VP succeeding near zero.

Lemkin says that after the turbulence of 2021 to 2023 he has interviewed about 50 VPs of sales in a year and found the majority no longer want to do sales. His first screening question is whether they want to sell. He views carrying a quota as proof that someone still cares about the craft, and says the best salespeople like money and also like honing the script and beating competitors.

He favors a simple "Columbo" question for VPs of sales or product: what would you do in your first two weeks? If the answer isn't to meet customers, he is out. The great product leaders he knows say "leave me alone, I'm going to go talk to 20 customers", while the wrong people say they will spend a month on process, Salesforce setup and territory planning.

Lemkin says he is empathetic about burnout but argues founders cannot hire these candidates. He acknowledges that at very large scale, such as Adobe's, sales can become dashboard-driven, but most startups are nowhere near that size. Even Salesforce, he notes, still has Marc Benioff meeting customers and prospects in person at Davos.

Lemkin's favorite test is a version of "sell me this pen", but with your app. He prefers it in a second interview, with time to prepare, and is surprised how many candidates haven't even watched the product's explainer video or webinar. The goal is not perfection but enough confidence in the core problem the product solves.

He says the mistake 99% of founders and reps make is thinking they sell in B2B when they are really solving problems. A good rep is like someone selling a Tesla Model 3 performance, not a used car: honest about weaknesses, expert on the answers, and willing to tell a prospect not to buy yet. He says the best reps have the confidence to say the product isn't the right fit today.

He thinks hiring discipline eroded after late 2020, when companies hired warm bodies and dropped reference checks and live demos. He argues a failed hire is the founder's fault, since the diligence should have caught it. He likes to take the last interview with a candidate who already has the job and give them a safe space to raise concerns so they don't leave three months later.

Lemkin says founders get comp wrong. Early on, the only thing that matters is whether a rep can close more than they take home. He suggests a ramp in which a new rep keeps 100% of what they close for a first quarter at most, while still being paid a market salary. A typical split is 50% base and 50% bonus, so a $140,000 OTE is really around $6,000 a month of fixed cash during a short experiment.

A rep must bring in multiples of their pay: roughly 3x for small business, 4x for mid-market and 5x for enterprise, with a $140,000 package needing to produce $500,000 to $600,000 or more. With reasonable marketing costs, reps are accretive, and negotiating $150,000 down to $130,000 barely matters. He says that if the plan is right you rarely have to revisit it, and too many tweaks just burn cash.

In the early days he prefers a few excellent reps to many mediocre ones, concentrating leads on the best closers. He tells of Pipedrive's first sales rep, who sorted self-serve customers by seats and called those with more than one. The rep sold AOL 100 seats after it had 20, took 20%, and for a time out-earned the founders, which Lemkin saw as a good sign.

Lemkin calls mishires the single most important issue in scaling and says a rep who fails should be moved on quickly, partly for their own sake. Sales performance is more binary than other functions: someone can be excellent in one environment and fail in another. So you must find the problem before hiring and root it out fast, because leads are too precious.

His top hint is to hire someone whose last product was harder to sell than yours. His first SDR, Sam Blond, came from Intacct, selling online financials when buyers distrusted the cloud for that data, and he found Lemkin's product much easier. People coming from easier products, he warns, often lack the skills to sell yours.

As heuristics, he says more competitive markets are positive: the number four in a crowded space may be great. A more technical sale is a risk, since many business-process sellers melt when selling to engineers, which he calls B2D. Going from a simple business process to a complex product with many integrations is another danger sign.

Lemkin describes the shape of a sales org as rules of eight: eight SDRs per manager, eight AEs per director, and eventually eight directors under a VP. There are no real efficiencies in sales, he says, so half of a sales-led company will be in sales at $10 million, $50 million and $100 million. After a period of pushing SDR teams to 12, he thinks everyone has gone back to eight.

For most founders, he says, the first head of sales should be a stretch VP who was a director or senior director. A stretch hire will be hungry, but may break if they cannot hire better managers beneath them. He recommends promoting 50% of managers from within and hiring 50% from outside, because an org of managers who were simply reps for two years cracks.

He also explains the terms: an SDR is an entry-level opener paid roughly $60,000 to $80,000 OTE, while an AE closes leads and might earn $90,000 to $200,000. The "full-stack AE" doing their own outbound is mostly a myth, he says, and few founders have the stamina to manage a team of SDRs. He notes there is no real senior-IC track in sales and warns against making the top AE the head of sales.

In the best B2B companies he has seen, the head of product is deeply involved in sales. He gives the example of a company approaching $60 million in revenue where the head of product, present in a big-customer meeting, recognized the customer only wanted parts of the product and saved the deal. Such a VP can commit on the roadmap in the room, which a sales leader may not be empowered to do.

For handling sales requests, Lemkin says customers can digest only quarterly releases at most. Each quarter, give your head of sales a fixed budget, say 10% of story points, to spend as they like. Without it, deal-by-deal swings from HubSpot to SAP to Salesforce burn out engineering. He also wants product to hold weekly alignment sessions and explain what is planned and what would be disrupted by a change.

When a junior PM is asked by a rep to add a feature, he says the right answer is to push the decision up to the VP level rather than decide. Tension between product and sales is a healthy sign, he says, but it needs structure. A weekly meeting where each side force-ranks priorities in their own team helps surface better ideas, such as discovering a feature can be built quickly or outsourced.

Asked whether product teams should own revenue targets, Lemkin says he needs another year to form a view, but warns about what happened to customer success. It went from optimizing happy customers to being measured on revenue, and he says it damaged relationships. He recounts a company that demanded $50,000 upfront from a customer paying $299 a month and later walked the threat back.

He wants product aligned with the revenue goal, especially senior leaders with equity, but does not want average PMs forced to deliver revenue from every experiment. He contrasts startups with big companies that don't fire people when a new initiative fails, which he finds liberating up to a point. He cites a checkout-upsell example as the kind of dark pattern that works but does not create loyal customers.

His broader point is that B2B revenue must compound, and customer-hostile moves anti-compound. For self-serve and PLG businesses he says the most important metric is churn, and he is out if churn is not top decile for a low-end product.

The one habit he would give founders and product leaders is to never leave a meeting without a next step. It need not be a check; it can be another demo or asking who else should see the product, since deals now involve several stakeholders. Founders often end the Zoom and wait in their inbox, but he says the best salespeople march deals forward step by step and only ask for money once the deal is in the red zone.

On resources, he points to the free SaaStr University course and the book From Impossible to Inevitable, and to the Pavilion community, while urging caution about the flood of paid sales courses. On free trials, he notes that 14-day trials arose partly because Salesforce's reps wanted to close deals within the month, with no evidence they were better for customers, while Slack, Canva and Zoom did well with very long free periods.

He calls annual contracts for SMBs bad advice from people who have never built products, since customers should pay the way they want. He argues for a "VP of Free", someone who champions the large free base that sales and growth teams are not motivated to nurture. Founders should also earn price increases with added value.

He closes the main interview with a challenge to product people: make this the year of the customer and ship three great things.

In the lightning round, Lemkin's life motto is "be kind". He says his biggest regrets are moments he was not kind, such as lashing out at a customer champion who switched to DocuSign. The top SaaS CEOs he has met are kind without being soft, and he says that when an employee fails, it is the founder's fault for hiring them.

On running SaaStr, he says it began as small meetups in 2012 and 2013 and grew by accident into a one-day event in 2015 with 3,000 people. Events have diseconomies of scale: free meetups cost nothing, a hotel event costs hundreds of thousands, and a mass-scale multi-day event is around $1,000 per attendee in the Bay Area or Las Vegas. He puts the Bay Area event at about $10 million to run and London at about $2 million.

His advice for founders is that every startup should bring its own customers together, even over a steak dinner, because customers sell to prospects. Among industry events, they should pick the one or two best, since the best people attend about two a year, and skip second-tier events. He also recalls SaaStr losing $10 million in a day when it was canceled in March 2020.