How Slack Hired its First 5 Sales People
Thinking about hiring. Showing your work. Scoring candidates. Lying at work. Finding our first 5.
How do you hire for sales people when you’ve never hired for sales people? And (ahem) you’ve really only officially just started being a sales person yourself?
After we got to $10-million in ARR and I hit the wall, those were two of the question I faced as I set out to hire people and scale up our team. Oh, and I started my third job at Slack at the same time, managing sales people.
To answer those questions, I tried to think about what characteristics I used when I succeeded in the job. Then I tried to convert those characteristics into some kind of evaluation process. I’ve broken that evaluation process into 5 steps below.
In my evaluation process I assumed I would make mistakes. If I’d learned anything it’s that the first time you do something you are both learning how to do the thing and doing the thing. In this case, I was figuring it out with a live audience with next to zero experience to draw from!
So I needed a plan that allowed me chances to recover from mistakes, and I needed a process that I had some familiarity with. So I strove to create an evaluation process that mimicked the actual work I was doing as closely as possible. Why not? I knew it pretty well and it seemed to work. The guiding principle I stuck to in my hiring was asking candidates to show they could do the tasks the job would require them to do. Show their work.
Around the same time I had just read Thinking Fast and Slow by Daniel Kahneman. In the book, Kahneman tells a story about the selection process for Israeli officers. He evaluated the selection process and found no connection between its recommendations and its ability to accurately identify officer candidates. It wasn’t any better than a random lottery. So he created a new process that used faceted evaluation to score candidates. Here’s how he describes that new process.
Step 1. Select 6 traits that are useful in the position. Make sure that all 6 traits are independent of each other.
Step 2. Make some factual questions based on each trait.
Step 3. Score each question on a scale of 1–5
Step 4. To avoid the Halo effect score one at a time before you move to the next question.
Step 5. Do not skip around between questions.
Step 6. Add up the scores of all the questions. Select the candidate with the highest score even if you think otherwise.
This seemed clever and scalable to me, and a better template to follow than the nonexistent template I otherwise had to follow.
Then I imagined someone smarter than me and more experienced and what they might think of the process. Faceted evaluation had worked for the Israeli military. I imagined that someone smarter than me might agree with it, so I copied it as closely as I could.
Job Descriptions
But before I could get to evaluating candidates, I had to get some candidates. I had to write the dreaded job description.
I say dreaded because so many comparable job descriptions I read in researching the market prior to writing our own were so terrible. Bland, boring, undifferentiated — there had to be a better option.
I thought that our job descriptions should be easy to read and easy to like. It couldn’t include any boilerplate corporate language unless absolutely necessary (footers and such). It had to read like a very concise story the prospective candidate could place themselves into.
This aspiration proved harder to do because I was starting from scratch. But if you’re considering using our experience at Slack as a guide, I want to say it’s worth the effort to write excellent job descriptions because of their asymmetry — hundreds or thousands of people will see the job posting, make a judgement and prospectively respond to it. Each following job description your organization creates will likely use the original as a template. It’s the kind of work that, when done well, I thought of as money in the bank that could pay me back.
(I’d love to at this point to share the exact job description. But I’m afraid it’s lost to time. The Wayback Machine does have the slack.com/jobs page archived from around this time, Oct, 2014, if you’re interested.)
Once I thought I had the job description together, I forced myself to read it aloud in front of someone I liked (thank you to the endless patience of my wife). I had to force myself to actually read every word, and it was painful. But it was also fruitful. Where I stumbled or added a word or it just didn’t feel right, I marked it and changed it. The forcing function of reading it aloud humanized and simplified the job description.
Then I posted the job and sent it to people I cared about. I sent it to people who I might have wanted to apply and people who might know the people who I wanted to apply. It felt like a mini marketing campaign and, like any campaign, required a bit of planning and persistence.
More specifically, it required me to directly ask for people to put themselves forward as candidates — “If you are curious, please hit the button and apply. If you know anyone who might be curious, please share it with them and encourage them to hit the button and apply. I’ll be on the other side of the button and I’ll read every application.”
At the time, I didn’t really screen for sales experience so much as for the quality of the application, the sense of a story of the applicant and the trajectory I could imagine their career on. We focused on that trajectory, moreso than their experience.
Step 1: Writing Assignment
When I found an interesting application, I gave folks a written assignment to do within a few business days (~7 or so, if I remember correctly). Enough that they could find solid time to do it, if they have a busy life and family and hobbies outside work. Not too much time that they forgot about it or deferred it or lost interest or context. Time kills all deals.
The written assignment? I didn’t bother imaging anything. I just used a super specific scenario, drawn from real work I had done, that had 3 exercises. Each exercise then had 2 parts:
Part 1 of each exercise was for the candidate to show their work. I gave them a customer inquiry (see them below), stripped of most of its identifying bits (names and company) and asked them to respond with constraints. The candidate had to make their response fewer than 400 words. They had to be clear and respond to the whole inquiry. They had to try to elicit a specific action in the customer.
Part 2 of each exercise was for the candidate to explain their response. Why did they include certain aspects? What were they seeking to achieve? Here they had more leeway on length and form. Part 2 was almost always more interesting to read than Part 1 and gave the candidate a really fair chance to unpack and describe their work product.
Since there’s no substitute for doing the thing, here are the 3 exercises drawn from real customer inquiries.
Exercise 1
You receive this inquiry:
From: XXX@maiainstitute.org – Today 07:57
Hello,
I am very interested in using Slack for my business, but I was wondering if it was possible to speak with someone regarding how Slack would integrate with our current needs and software infrastructure. We currently use SharePoint, Dropbox, Office, Gmail, and are looking to add SumAll and Yammer.
Please let me know if there is someone I can speak to that might be able to guide me through the transition? Thanking you beforehand,
<signature>
Exercise 2
You receive this inquiry:
From: YYY@ge.com – Yesterday 14:58
Hi,
Do you offer volume licensing? If so, after how many people does volume licensing kick in?
We’ve been using Slack for 2 months in our engineering team and want to roll it out to our whole IT division – 120 people. We could do an annual license.
<signature>
Exercise 3
You receive this inquiry:
From: ZZZ@pressreader.com – Yesterday 11:55
Hello,
We are interested in implementing slack across our organization, but, we need to create a business case for our operations manager. Can you please provide us with the benefits of Slack that show how it can help streamline our business and reduce complications? Thank you in advance!
<signature>
The exercises each had some clues in them for further information gathering — the domain name would lead to the organization, if the candidate did a little research. The time received on the inquiry could lead to an order to respond in. We expected the candidate to be able to pick up on those clues and interpret them and show us their work. But overall, the exercises were pretty self contained.
The exercises also proved to be a bit intimidating to folks. It seemed like a lot of work to them. It required sustained thought and concentration. If you weren’t a bit curious about the process and excited by writing, it likely wasn’t something you wanted to do.
But if you saw it as a game to see how well you could solve these tiny mysteries, then it was for you. And that filtering process was incredibly important because for the most part this was the work that a successful candidate would be doing on a day-to-day basis. I invited many folks to do the writing exercises and not all of them did them. This is good! They self selected out of the exercises, which meant they self selected out of a job they didn’t want or weren’t willing to do.
The last thing I’ll say in guidance is that I think it’s important to be ruthless about the executive function elements of the written assignment. It can’t be longer than the word limit. It can’t be late. It can’t have spelling mistakes. Any names have to be spelled correctly.
I also judged the work based on visual and usability elements. It couldn’t be a single blob of a paragraph with no thought to the eyes or brain of the reader. It had to have some flow and some consideration demonstrated for reading on screen.
Step 2: Scoring the Work
Once I had received the written assignment from a candidate I would read through both Part 1 and Part 2 twice, then score the written assignment on a standardized matrix.
I shortcutted the Kahneman 6 ratings and only used 5 facets, each scored out of 5, with a 5 as excellent and a 1 as terrible.
The 5 facets we used for our evaluation matrix, with their companion questions to be considered in the score by the evaluator were, in no particular order:
Communication — How well does this person communicate in writing? How well in conversation? How well do they read / listen? How much do they tailor their communication to the context (audience, form, timing, setting)?
Integrity — How well does this person uphold trustworthiness? Are they trustworthy as a representative? How do they act in situations where they are uncomfortable? Do they check out as a high integrity person?
Poise — How well does this person present? Do they come across as well prepared? Do they handle challenges well? Can they understand framing and can they reframe?
Savvy — How well does this person understand deal making? Do they understand influence and status well? Can they think strategically, ahead of the situation? Can they build relationships beyond simple transactional interactions?
Grit — How much persistence does this person demonstrate? Do they work to persevere in challenging situations? Can they tolerate failure and discomfort to achieve an objective?
I added up the 5 scores in the 5 facets to get a total score for the work out of 25. Then I created some thresholds for the evaluation scores so I didn’t have to think too hard, and a corresponding default next action.
Scores of <15 didn’t get any serious consideration unless there was something extraordinary about them. These candidates were default declined.
Scores of 16-20 were maybes. Was there something in there worth considering? Did they do anything compelling to merit moving them ahead?
Scores of 21+ were default yes responses unless we could find a reason to decline them. These candidates moved on to the next stage of the evaluation.
When I first started using the evaluation matrix I didn’t have the score thresholds above. Only after using it for a few repetitions did I figure out I could create another process shortcut with the thresholds and default next action.
(This evaluation / decision making model worked really well and we used it in many other situations later in my Slack career where we were trying to make one-way decisions in conditions of pretty high uncertainty. The 5 dimensions changed to reflect different decision criteria but the overall process of scoring facets and then using total score thresholds proved very valuable.)
Step 3: Presenting Slack
For candidates that moved on in the process, the next step after the writing exercise was a Zoom sales presentation.
In it, we asked them to present our product to us. Tell us about Slack with an introductory demo. We gave them a templated slide deck they could customize. We gave them a scenario (customer details and situation, role of people on the call) and guidelines (duration of call, what they needed to cover).
We wanted Step 3 in the process to feel like a continuation of the written exercise and to mimic pretty closely how our sales process worked. Once more, all the details were drawn from actual customer inquiries.
For Step 3 we used the same evaluation matrix and 5 facets as Step 2 and updated our scores. The same thresholds for scores applied too.
To receive the Zoom presentation from the candidate we always had 2 evaluators on the call acting as the prospective customer. After the call, each evaluator did their own independent scores. Then the evaluators reviewed their respective scores together and agreed on a collective score. We revised the scores in the candidate tracking sheet based on the agreement. If the evaluators didn't agree on the scores, and one had the candidate below a threshold of 21, we didn't move the candidate ahead.
One of the consistently interesting parts of getting candidates to present Slack to us was that great candidates actually taught us something about how we could present our own product. They did product demos that in some cases were embarrassingly as good as our own. They answered questions in ways consistent with our culture. In short, really strong candidates were preternaturally good at selling our product back to us, and they worked hard to prepare.
Step 4: In-Person Interview
For those candidates that we moved ahead in the process from the presentation, the next step was an in-person interview.
At first, this meant meeting up at our local coffee shop (JJ Bean once more!) and then going on a walk around the seawall because we basically had no room in our office. The coffee shop was more private than the office too, especially if you could score a table on the patio. The seawall was more predictable, even with Vancouver’s high probability of precipitation.
I remember quite clearly wrapping up an in-person interview with Mike Clapson as it unexpectedly started to rain. I could see the raindrops accumulating on his shoulders as we talked, dots on his grey sweater. Neither of us wanted to cut the conversation short so we stood and kept talking as the rain dots sprouted larger and more numerous. Eventually we wrapped up but I remember for a long time we both stood in the rain pretending we weren’t getting wet to continue the conversation. That was a good interview.
For the in-person interview, again, executive function elements were important. Were they on time? Were they prepared? Did they find the place without problem? Did they pay attention to the details?
I understand that people may scoff at this emphasis on the basics but I found it paid off. It meant the other person valued similar things to what we valued and did the small things in a way that we wanted them done, without needing to be told.
In terms of in-person interview tactics, I always wanted to let the candidate ask their question or questions first. I wanted to hear what was top of mind for them. What was the most important thing they needed to know at this stage? How did they think about this process? How were they evaluating Slack? These were sales candidates after all. They needed to be able to run a meeting with a decision maker.
Sometimes their questions would take up half the allocated time of the interview, and I thought this was fine. I knew I’d get a chance to ask my questions and I wanted the candidate to get comfortable and get to know us. This was largely culture fit and behavioural attributes in action.
Once we got into the second half of the interview, once their blood sugar was a bit lower and nerves less jangly, I wanted to shift the conversation to some harder questions I had for them.
"Tell me about a time you had to lie at work."
I found was a very revealing question. I wanted to know how they thought about the truth. Were they comfortable dealing in shades of grey? Did they understand the shades and how stories could be burnished to a version of events? Did they understand how to be transparent in degrees? I also wanted to get a sense for how they dealt with a complex and touchy subject. Did they have the bravado to say they’d never lied at work? Were they comfortable and trusting enough to be honest?
The reason I liked to ask about lying and trustworthiness originated in an article I read about Warren Buffett.
“You’re looking for three things, generally, in a person,” says Buffett. “Intelligence, energy, and integrity. And if they don’t have the last one, don’t even bother with the first two. I tell them, ‘Everyone here has the intelligence and energy—you wouldn’t be here otherwise. But the integrity is up to you. You weren’t born with it, you can’t learn it in school.”
“Tell me about something you know a lot about and I know next to nothing.”
I enjoyed asking this question for lots of reasons. It forced a candidate to evaluate me and potentially qualify the question with an example to see if they had a good topic to cover, “Do you know anything about putting together bunk beds?” It put them in the driver’s seat to communicate something, deciding on the essential information and its order. Did I need some backstory? Did I need some mental scaffolding to understand what they wanted to tell me? It let them reveal something more about themselves.
“How did you prepare for this conversation?”
This could be a bit of a softball question or could reveal the work the candidate put into the process that didn’t necessarily always show up. I had people tell me about the research they did on me. Some people told me about breathing exercises and power poses in the mirror to harness their own strength. Some people confessed they did no preparation because they were too busy or running between meetings. At least that was honest.
"What's the hardest thing you've ever done?”
This question almost always stopped candidates and made them think. It also led to some remarkable stories of resilience and grit. This was the good stuff I was seeking. I heard about someone giving the eulogy at their mother’s funeral. I heard about a ballet student whose teacher demanded perfection and how much physical endurance and determination it took for her to achieve those standards.
Often times I think candidates actually surprised themselves with the stories they told. Maybe I just wore them down a bit. But the candidates I remember best are those that consistently let themselves be vulnerable. They dropped their professional persona to reveal a bigger view of their identity. They had the courage to go off script, be honest and trust the process.
Often times I also had to be patient with these harder questions and let silence linger. “You mean the hardest thing I’ve ever done in my life?” Yes. What’s the first thing that comes to mind?
Folks were surprised by the questions and understandably often needed some time to process them and consider their response. It wasn’t a softball like, “Tell me about something you’re not good at?” I always had found that kind of question gets a canned answer with little insight to offer.
Once more after the interview, I went back to the 5-facets ratings matrix and revised the scores as needed. The same ratings thresholds applied but now I had much more nuance and detail to include. Like with customer calls, I tried to take lots of notes about the interview as soon as I could to round out the perceptions and be able to revisit later.
Step 5: References, Offer, Acceptance
For successful candidates who we wanted to join the team, we always did reference calls. People say they do reference calls but do they really? I think they’re essential and we always did them. It always proved to be time well invested.
If there were no red flags from those calls, we made the candidate an offer and they accepted. I don’t think anyone didn’t accept. We got lucky.
And it’s clear to me that we also lucked out because the folks we hired were excellent. They made Slack an even better place to work. They formed the core elements of the Sales team culture. They were excellent at their jobs and they were excellent people to have on the team and they set the stage for the hundreds of Sales hires to follow.
They were also all very different from each other. Some had lots of formal Sales experience and some had diverse marketing experience. Collectively, their ability to apply the right blend of sales and marketing to help our customers made everything work as we figured out how to combine the two and succeed.
First 5 Hires
Kim Graves joined first. Kim was young, in her mid 20s, and had only held one full-time job since graduating from university. Starting at Slack she and I shared a rectangular desk, facing each other but staggered to share the space. She was on the northwest corner. I was on the southeast. I never worried about giving Kim any task to tackle. Nothing. She learned incredibly quickly and had terrific communication skills. She also consistently just figured things out for herself with bravery and grace and humour and conscientiousness. Over the years she grew and excelled and went on to lead hundreds of people on Slack’s small business sales teams across North America.
Mike Clapson joined next, once he dried off from our in-person interview. He actually had software sales experience, combined with an engineering degree, which made for a potent combination of business acumen, process orientation and technical fluency. Mike wanted to keep score in everything, and to know how he measured up. And he wanted to excel. He injected a ton of sales know-how to our team. Mike would go on to lead our Asia Pacific (APAC) business, moving to Melbourne for a few years to open the Slack office and hire the first teams. Then he moved on to HQ in San Francisco to lead our enterprise sales teams.
Kyle Shackley joined and he also had actual enterprise software sales experience. He was the first of the Account Managers to have a family and kids at home, and that introduced a new sensibility and awareness we’d been lacking. If you didn’t watch closely, Kyle could come off as easy going. In the interview process I made the mistake to think he was lackadaisical even. But if you looked closer you could see the attention to detail, his mind consistently working on problems, and his tenacity. Kyle worked on many of our largest customers for many years and led a 100,000+ person rollout of Slack at IBM, our largest customer for a long time.
Dave Macnee joined and I can’t remember if he had any sales experience to speak of or not. He must have. If not, he was a natural. What I do remember about Dave’s candidacy is that he did have a lot of experience as a camp councillor. And how different was selling Slack to companies from organizing kids for archery or sailing? Dave also did the writing exercise and the Slack presentation as well as anyone we had ever seen. He taught us new ways to demo our product and consistently injected humour and joyfulness into work. His breadth of interests and experience meant he was well loved by his customers and he went on to work on Slack’s customer success and thought leadership teams called Future Forum.
Christina Meng joined and she definitely had structured sales experience in a larger organization, creating leads, handling opportunities, qualifying and progressing deals. She had an amazing energy and charisma and drive to learn. She also sang and performed and in the years that followed she always won the Best Deal Story contest at our global sales kickoff events. It was like she was visiting us office plebs from the showbiz planet. She also didn’t miss a stitch of detail. She would go on to get relocated to HQ in SF to run our global scale customer success efforts.
Seeing all of these folks’ starting points at Slack and then how far they built their careers reflects a few things. Did I mention they were excellent? They were trustworthy? And savvy? And tenacious? That they worked really hard and flexibly to succeed? Yes, all of those things.
Slack grew thanks to the work they did, and that growth created more new opportunities for them to succeed again, with the trust they’d earned and the relationships they’d built.
Hiring these folks and seeing where they drove their careers to, all while remaining excellent people, remains one my fondest and most meaningful memories of what we accomplished at Slack. Thank you to each of them for trusting in Slack and our weird and unproven process!
Up next — Meeting the New New Boss: Getting another new boss. Mapping how we sold Slack. Finding how strengths can become weaknesses.