Sales 101 with my ex-boss Guy Zerega (former Stack Overflow EVP)
Guy Zerega ran sales and marketing at Stack Overflow and once hired Jack. His first rule of sales: developers don't want to talk to you, they want to read a well-written email — because to a developer, language is the written word, not the spoken one.
Guy Zerega led sales and marketing at Stack Overflow, where he built the revenue org from scratch after founder Joel Spolsky — who didn't believe in sales at all — watched Atlassian eat Fog Creek's lunch with a worse product and a real sales team. He now leads sales at Cyborg, an end-to-end encrypted inference data startup, and he was Jack's old boss.
This is a 101 on what actually matters in sales, especially when you're selling to developers: why the channel you choose matters as much as the message, what the best salespeople do after a good first call, when a founder should stop doing it all themselves and hire a team, and why you should always hire three salespeople instead of two.
The Q&A below is experimental — extracted from the episode transcript with AI. If you spot a mistake, let us know at jack@scalingdevtools.com.
You used to hammer 'know your audience' into us constantly at Stack Overflow. Why is that the first rule of sales?
Our job is to communicate, and the way I speak to a five-year-old is different from how I speak to a 25-year-old to a 50-year-old, from a business person to a creative person to a technical person. What I need to understand is that the person in front of me dictates how I communicate. The idea that you want to treat people the way you want to be treated doesn't apply, because not everyone is like you. The real way to think about it is: how does Jack prefer to be spoken to? As a fallback, yes, treat people the way you want to be treated, but as you get to know them, treat them the way they want to be treated.
What's the number one complaint about salespeople, and what does it actually mean in practice?
The number one complaint is we don't listen. That means either you're talking and I'm not hearing what you're saying, or I haven't stopped talking to give you a chance to speak. When you're not being heard, it can strike a deep emotional trigger — it can make you feel like what you're saying isn't valued, and any salesperson creating that environment is never going to be successful. The best way to ensure you're listening is to ask good, open-ended questions and let the prospect describe and talk. If they never describe their challenges because you're talking too much, you make them feel defensive, and clients who feel defensive shut down and become nonresponsive.
What does 'know your audience' mean specifically when the audience is developers?
Developers code in a language, but to a developer, language is the written word, not the spoken word — you'll notice developers on Slack say 'I'm talking to Jack' when they're actually typing to Jack. That tells you developers prefer the written word to the spoken word, so if you think you're going to call someone up and talk them into your product, you're probably not choosing the right channel. The other side is that developers are trained to look for bugs in code, so if you send an email that's grammatically or factually incorrect, it has no value — you lose credibility instantly. You have to research the person and be thoughtful, because the more you know about who you're trying to connect with, the better your chances.
Say the first meeting goes well. What do the best salespeople do next?
I was a competitive athlete for a lot of years, and I had the good fortune of working with an Olympic coach who used to say you have to earn the right to win — I translate that as you have to earn the business. After a good call, you send a summary of what was discussed and what you agreed as next steps. That's the real meaning behind 'always be closing' — it's not closing the sale, it's always be getting commitment on next steps, because there's no such thing as a one-call close. The best salespeople are people who care about the well-being of their customers, who want to sell something worthwhile and meaningful — if you don't care about that, what chance do you have of being a good salesperson? Trust is what makes someone agree to buy: when people make you feel good, and are easy to talk to, and provide valuable insight, you begin to trust them, and no one buys from someone they don't trust.
When a founder has been doing founder-led sales and wants to hire their first salespeople, what do you tell them?
First, I ask: what's the commonality among the sales you've had? Is there a pattern? Because when you begin to scale, if there's no pattern, it's really hard to scale. You're trying to build an ideal candidate profile the same way you'd build an ideal client profile — what background is relevant, is it a technical sale. The other thing that's really important at an early-stage startup is finding someone who's actually worked at an early-stage startup before. People get enamored with hiring someone ex-Microsoft or ex-Google, but that person had every marketing tool and sales collateral at their disposal at a huge company — you put them in a startup and they're asking where's my sales collateral, where's the next focus group. Look for someone who's been successful at a small company instead.
Why do you say to hire three salespeople rather than two?
If you hire two salespeople and both are terrible, is it that both are bad salespeople, or is it the product? You can't tell. Hire three and you actually get a fair comparison — you'll have a top, middle, and bottom. I've seen companies hire two, where one has zero success and the other has huge success, and then every person they hire after that struggles too — because they hired someone who's a ridiculously good salesperson that no one else can replicate. That's not the worst thing, because someone experienced can sometimes break down what that person is doing and help you recreate it, but that's a different skill, and most successful salespeople can't articulate what they're actually doing.
How do you actually decide the deal size where self-serve stops working and you need a salesperson?
It's not purely a mathematical or financial question, it comes down to the product. At Stack, when I got there, we sold a single thirty-day job listing self-serve, and we found out fast that nobody hires somebody in thirty days, so it typically wasn't successful. We moved to selling five months' worth of job listings as a minimum, and then found that having a salesperson reach out roughly doubled the amount of revenue a client was willing to spend on self-serve, just by educating them on what's required to actually be successful in that market. I've worked in environments with a $9,000 average deal closing ten a month, and others at $5,500 average closing twenty a month — both can be fine. It always starts with: what is the market, what is the product.
You mentioned Vercel hiring technical people instead of typical salespeople. What do you make of that?
Vercel understood that in order to get the momentum they needed to move into the enterprise, they had to have the support of the developer community, and you don't get that by putting slick salespeople in front of developers — you put technical people in front of the community instead. The other side is that my understanding, and I'm not technical at all, is that Vercel's products are so intuitive and easy to use that they're a pleasure at every turn, and something like that is going to be way more successful on the self-serve front than something that isn't. Developers are much more comfortable at their keyboards than in a dialogue, so a product where everything's in the docs and foolproof is the ideal product for that audience.
Has the rise of AI changed how much word-of-mouth and referrals matter in sales?
More today than a year ago, the testimonial or referral of a colleague is more important than what you're reading online, partly because of trustworthiness but also because of the sheer volume of noise out there. A salesperson used to be the word-of-mouth, the expert, but now information is at everyone's fingertips, clients are more educated about the product than ever, and it becomes information overload — people just want to know who in their network has had this same problem and solved it. What you're really relying on now is the value of the network, of having strong relationships with the people you've worked with.
If you had to give one piece of advice to a DevTools founder about their sales process, what would it be?
We used to call it honeypotting — every sales call is an experiment. If you approach it like a scientific experiment, you ask: what are the controls, what are the things we can actually control? That's the best advice for any founder: double down on your success. Get some repeatable actions and repeatable behaviors in terms of those controls, and as you hire salespeople, they'll find the market for you. Then hire experienced salespeople — an experienced sales leader who's been there, done that goes a long way.
Links & references
This episode is brought to you by WorkOS.