别把学习外包：创始人必须自己把销售闭环跑通

Pete Kazanjy did not come up through sales. His background is in product marketing and product management, starting at VMware, and he founded the recruiting software company TalentBin in 2009. He quickly found that B2B software does not sell itself, so he became the company's first seller, sales manager and sales leader. TalentBin was acquired by Monster Worldwide in 2014, and he later wrote *Founding Sales*, built the Modern Sales Pros community and started Atrium, a sales-analytics product.

Kazanjy says the bad reputation of sales comes from movies like *Glengarry Glen Ross* and the used-car stereotype. He prefers to see sellers as microeconomic agents who carry supply in their back pocket and go looking for people who have, or don't yet realize they have, the demand. That process matters for getting technology deployed in the first place.

What makes it "modern" is measurement. With CRMs and the digital trail left by email, calendar, phone and Zoom, behavior can be measured, managed and improved. He compares it to product analytics moving from licking a finger and testing the wind to Mixpanel and Amplitude, and to the pre- and post-*Moneyball* shift in baseball.

Kazanjy calls *Founding Sales* the sequel to *The Lean Startup* and Steve Blank's work on customer development. First you validate a real problem, then you build a minimum feature set, then you need people to pay for it reliably and at scale. This is a loop, not a handoff: a minimum viable product will probably be weak, and getting to something valuable means sitting in front of many customers, which is a sales behavior.

He argues it is easier for a founder to become minimally good at selling than for a third party to become as expert in the founder's subject. Using Atrium as an example, he says he is probably the world's expert on sales analytics, so trying to train someone else to go sell it early would have been a poor use of time. He ties this to Steve Blank's observation that startups often have to fire their first VP of sales, because the founder skipped this step and outsourced it.

Lenny summarizes three reasons to sell yourself: learn what to build, learn how to position and pitch, and work out what you will teach your first salesperson. Kazanjy agrees and adds why packaging matters. B2B companies scale mostly by adding salespeople, who are capped by the hours in a week, so you go from two reps to four, eight and sixteen only once each stage works.

Kazanjy's rule of thumb is a technical metaphor: when it runs reliably on your local machine, you test whether it reproduces somewhere else. In practice that means turning first meetings into customers at something like a 15% to 25% win rate, and doing so across enough at-bats, maybe 50 or 100, that the result feels statistically significant. Two wins out of ten arm's-length conversations is a good start, but he wants to see it repeat.

The next step is to move what is in the founder's head into slides, scripts and email templates and teach two more reps, not ten. If you jump stages, he warns, "you're hosed." He also advises running prospects in cohorts of about ten first meetings rather than 100 at once, so you can see how discovery questions, slides and demos land and iterate before the next cohort.

He treats the sales motion as software you constantly update. If a prospect asks a question you can't answer well, you build a slide that handles that objection and add it to the talk track, so the next loop goes better. Because only a fraction of deals ever close, he recommends tracking leading indicators such as the share of first meetings that become second and third meetings. In his example, seven of ten first meetings reach a second, and about 4.5 reach a third.

To the objection that a self-serve, freemium product won't need sales, Kazanjy says founders don't have to be good at sales, just non-zero. Many product-led motions stagnate because nobody layered sales on top, a topic he covered in a guest post on Lenny's newsletter about adding sales to a self-serve product. The most famous example he cites is Dropbox, which had superb sales leaders but, in his view, never put enough product investment behind selling across a whole organization.

His summary is "never mistake your lead gen for your business." Customers paying $19 or $29 a month are great, but the large ARR numbers come from $50,000, $100,000 or $250,000 contracts, and organizations want a human to navigate those. A bottoms-up product can win over a user who has no budget authority, but eventually you must talk to the person who holds the purse strings.

Asked whether essentially every B2B company ends up with a sales team, he says it is a question of when, not if. Atlassian had a sales organization even if it didn't call it that, and went far by pricing very low. Developer-tool companies like Datadog had meaningful sales teams early, and Snowflake has hundreds of salespeople.

Kazanjy says the hardest part for engineers and product managers is a shift in behavior. They usually talk to the same six to ten people each day, while selling means meeting multiple new people daily, having a couple of minutes to build rapport, and keeping many parallel conversations straight in a CRM. He compares this to Paul Graham's maker-versus-manager schedule: salespeople live on a manager's schedule of constant context switching, with perhaps two to five customer meetings a day.

Once you accept that, you can drill the behaviors. He calls one "turbo rapport": think about how quickly you could befriend a bartender, a flight attendant or a barista, who are used to people being unfriendly, and bring that to prospects. From rapport you can ask candid or provocative questions, follow-ups, and eventually ask for money and then stay quiet. He quotes a friend's comparison to a pianist playing scales: find the equivalent exercises for selling.

He also reframes selling away from convincing someone to buy, and the "ice to an Eskimo" idea. A seller, he says, is a consultant with a preference for one solution. If a company has one salesperson and Atrium's minimum customer profile is about 10 to 300, he would turn them away. When discovery reveals a costly problem and your product really fixes it, "everybody wins."

For a first step, he offers an easy version, which is to start conversations with strangers by complimenting a shirt, not the weather. The harder one is to be crisp about your ICP, the ideal customer profile. That covers both the account characteristics and the personas involved, such as the user, the technical stakeholder and the budget owner, so you spend your loops on people at the white-hot center of the problem.

Kazanjy advises against hiring a VP of sales first. You need 10, 20 or 30 customers yourself, and because "there is no GitHub for sales motions," the playbook lives in your head and documents. He recommends a couple of early-stage pioneer sellers instead. A head of sales who now manages managers hasn't sold in a while, and as Jason Lemkin joked in a tweet he mentions, an executive who already scaled a big company may wonder why they would do it again for a small startup.

Instead, look for early people from companies that sell to a similar persona at a similar price. For a recruiting-software startup, that might be early sellers at Lever or Greenhouse. For a design tool, he would look at earlier sellers at Figma, or a sales manager not far from selling, and he cites Marissa Fuhrer as an example profile. Such a candidate is more likely to join after you can show a roughly 25% win rate yourself and two sellers closing about $50,000 of bookings a month each.

If a hire isn't working, check first that you gave them the materials: the deck, written discovery questions and a demo script. If they have these and still show poor win or activity rates, that is a leading indicator of a bad fit. He suggests hiring two people to start, because one failure then tells you less about the process, and he thinks four is hard to manage early on.

On timing, Kazanjy suggests continuous monitoring rather than a deadline. Month one is onboarding and mock discovery, month two should bring 10 to 20 first meetings with about half reaching a second, and month three should bring proposals. If no first meetings are being booked after a month, the hire isn't working. He warns that waiting nine months to find out is the worst outcome, since the leading indicators would have shown it by month two.

Asked about remote work, Kazanjy says it mainly hurts junior salespeople. New sellers need to learn the motion and be audited, so the correction loops need to be fast. Sitting among three or four people, listening to their calls and giving feedback right after each one updates their "software" far faster than reviewing calls once a day. He advises early-stage founders to sit shoulder to shoulder with their sellers, and worries about a generation of 24-year-old SDRs who missed that kind of learning.

His closing advice for founders and product managers is not to be afraid of sales. People who say you have to be a born seller are, in his words, "just talking their book." Persuasion, communication and thinking about what's in it for the other side are useful for internal as well as external selling.

In the lightning round he recommends *The Goal* by Eli Goldratt and Bill Walsh's *The Score Takes Care of Itself*, both of which inspired Atrium. He also likes a written screening doc of about a dozen biographical questions for hiring; he says 57% of candidates never complete it. His final story is about a seller who ran three or four opportunities with GRIN before landing the account, which now uses Atrium with about 100 SDRs and 80 AEs.