Slack's growth from 500k users to 5mm
Updated: Jul 7, 2019
Merci Victoria Grace heads up Slack's growth and met Stewart Butterfield at Game Founders Anonymous in 2015. Stewart brought on Merci to head new user experience. Growth grew out of that.
Tactical Empathy and emergent complexity
Slack's approach to growth is based on tactical empathy; borrowed from the book "Never split the difference; negotiating as if your life depended on it" by Chris Voss (hostage negotiator for the FBI).
Find your 'company voice'. Slack's is courtesy and respect.
Slack thinks about how to engage new users and does it holistically. What is a seamless experience from the beginning, say Slackbot, to mobile and then back to desktop. How could Slackbot and coach marks be part of the onboarding flow. Think of your product as providing long term value to users. Evaluate if the "one weird trick" works for you (it doesn't work for Slack).
Slack thinks of it like Mario Brothers. You first learn to walk, then jump, and then jump over a chasm. By this time, you're a good 3-4 mins into the game. You're being onboarded, but in the gamer's mind, they're already playing the game.
Pretend you're a human being
"We're 1% there. Slack is a piece of sh*t but you'll are doing a wonderful job."- Stewart Butterfield.
The growth team is a diverse 25-30 people (50% women, 50% people of color). While a different team does super top of funnel activities like ads and SEO, the growth team is a team of creators who buddy up and ask one another questions: "if we didn't know what <CTA> was, would this make any goddamn sense to me. Do I have enough information at this point to make this decision".
Engineer, designer, PM go through onboarding. Go far in the pretend-you're-a-human-being-path. What's the persona of the human, what's their level of intent. For Slack, to assume this person has never heard of Slack would be incorrect because that's not their problem right now.
Regardless of role, everyone at Slack spends 2hrs/day on Zendesk.
How to break ties: Debate furiously. PMs are relied on to be the ultimate voice of quality.
The worst part of the product and the white tower of usability
Pretend you're this human: You had a fight with your spouse, just drove in bad traffic, and your boss is screaming at you. Does your product's opening screens make that person smile. Don't assume they're also on your cinematic high res high compute Mac; your white tower of usability. Your human probably also has a dying battery, cracked screen and in direct sunlight.
Find the worst part of the product. When you fix that, automatically, some other part becomes the worst part.
Run small experiments. A mistake Slack made was to launch a new onborading flow + new website + run ads. You learn nothing because it's a poorly designed experiment and you haven't isolated variables.
Insert behavioral drivers (like social nudging to add other people, FOMO). Constantly put yourself in the shoes of someone using it for the first time. Relearn the world from someone else's perspective.
The iteration experience typically is: problem statement -> PRD -> pre-mortem -> build -> release internally -> iterate -> release -> post mortem (what went well, what didn't).
Pre-mortems are like FMEAs from manufacturing (failure modes effects analysis). What are the failure points. Eg: you might be letting people think your assistant is smarter than it really is. Set expectations, what are the edges of this thing. Prevent failure even before it happens but when you fail, fail in cute and funny ways.
People are sheep and want to be led (?)
You need a balance of data and intuition. Even though the data can tell you what's wrong it won't tell you what to build. Build with your gut, and build a company that takes risk.