We recently had the honor of hosting Merci Grace, Slack’s very first head of product, at the SaaStr Co-Selling Space for a deep dive session with SaaStr COO Mallun Yen to share her thoughts on what it takes to build a truly great, viral product like Slack.
Check out the full video and transcript from the event below, and don’t miss out on hearing more insights from Slack when Kevin Egan, Head of North America Sales, and Dannie Herzberg, Head of Mid-Market Sales, take the stage at SaaStr Annual 2019! Join +12,000 Founders, Investors and Leaders from Feb. 5-7, 2019 at the San Jose Convention Center.
Transcript:
Mallun Yen: I’m Mallun. I’m one of the cofounders of SaaStr. Thanks for coming to the SaaStr CoSelling Space. I am so thrilled that Merci is here to talk to us today. The background is that, one of the questions that I’ve been getting a lot recently from founders is how do you scale a product frankly once the CEO is out of capacity. After referring about five different companies and founders to Merci, who all raved about her, I decided to gather 100 of our nearest, dearest friends together so that we can talk about it all together and hear from Merci.
Mallun: Merci was employee number 50 at Slack. She was the first head of product after Stewart Butterfield. What you may not know is that, like all great product people, she majored in fiction writing at USC and was destined for … We’ve got another one there. Was destined for a career in Hollywood with a studio when she had been designing a game, because she was a Dungeons and Dragons aficionado in her youth, and somebody offered her $500,000 to say, hey, do you want to start a company. It’s always that easy, isn’t it?
Mallun: So she launched her first company and her first game, which was The Nethernet-
Merci: Yeah.
Mallun: Apparently it just missed out on being the MMO of the year but lost to World of Warcraft. Does anyone else here need to look up what MMO was? Am I the only one? Do you know what it is
Mallun: Okay, everyone else knows. Good. Massively …
Audience: Multiplayer.
Mallun: I’m sorry. Yeah. Massively Multiplayer Online. That’s what it is. So then Merci went on and focused on engagement and paid conversion at Couchsurfing and GigWalk before joining Slack in 2015. While at Slack, she ran their growth team as director of product and took the company from 500,000 daily active users to over 5,000,000. That is quite some growth. Congratulations.
Merci: Thanks.
Mallun: So Merci, you actually met Stewart through the gaming community, how does being a game designer translate into a product lead and how does that influence you and clearly, perhaps, you and Stewart were two birds of a feather coming from the same background.
Merci: Yeah, our shared gaming background, I think, well that’s how I knew Stewart. Then it really helped us think about product in the same way. The games industry has been doing something called play testing for a really long time. Has anyone heard that term before? It’s kind of usability testing, except what you’re really looking for is the emotional response to the game because fun is an emotion, having enjoyment of something, having a reaction that the game designer wanted you to have.
Merci: Great consumer and enterprise products actually elicit the same thing. People don’t make purchasing decisions because they have really logicked it out in their mind, even though we like to pretend that that’s the case. Virtually all the decisions that we make are really emotional in nature.
Merci: A game design background really helped both Steward and I think about how is someone going to feel when they’re using this product.
Mallun: In fact, one of the things that I’ve heard you talk about is feeling or pretending like you’re a human being, right. Tell us a little bit about that. Does that come from a game design background you think?
Merci: Yeah, so that’s a Stewart Butterfieldism. I don’t know how many of you are PMs, but one of the most harrowing times of a product manager’s life is when you’re doing a review with your CEO. Imagine that that CEO is one of the world’s best product thinkers. You come into this product review with Stewart. At Slack we often would be taking him through either a prototype or a deck. You know that you really mucked up when he turned to you, and this happened to me several times, and he said, ” Close your eyes.” And you’re like, oh great. Now it’s coming. This is great. So you close your eyes, kind of wincing, and he would say, “Pretend you’re a human being. Now open your eyes.”
Merci: He was asking you to look with fresh eyes at the thing that you had just shown him always because he felt like you had gotten into the persona use case dead land of product design and away from thinking of who’s this real person who’s going to use this. Is this providing value for someone in their actual life, or is this just, I think that I want people to take this action and not this is something that people really need to do.
Mallun: Of course, when you closed your eyes and pretended you were a human being, you actually did not pretend you were Merci.
Merci: Right. Yeah, I had a particular person. This guy Juan, who I would pretend to be because different race, different gender. It’s good to put yourself into someone else’s body, imagining that I was a mid-level person at some different company and I’ve had a fight with my wife the night before and then traffic was really bad and I’m rushing in, sitting down my boss who is a dick. He walks by me and he’s turning to tell me where’s the Harrison file. So I open up Slack, and I’m like, where is the Harrison file? How am I supposed to find it in this mess that I’m looking at right now?
Merci: That’s deeply influenced by my passion for role-playing games, I guess. Let’s just have it out. I would build these characters in my mind. Then I created a product process out of this with my team as well where, when we were probably 30 people on the growth team, one of the really difficult things about always working on onboarding is that you need to constantly put yourself in the mind of someone who doesn’t know how to use your tool yet. As you can imagine, Slack Slack is insane and people use the shit out of it. It’s like every power user of Slack actually works at Slack at this point. So we knew a ton about the product and it was really hard to put yourself in that beginner’s mindset.
Merci: So these character sheets would be very helpful to people. We’d pair a designer and an engineer together. They would come up with a company that they were working at and a circumstance in real life which we’d heard from users, mostly doing user research, that this is how people came to use Slack. Then they would go through the new user experience that they had built as these new people. We got incredible insights from doing that. It’s sort of like role-playing user research. The fact that it was people who were cross functional and we had a really diverse team I think also really helped.
Mallun: In fact, when you first joined as employee 50 you were brought on primarily to do user experience to start, right, before you ended up broadening and taking over product. So take us back to you’re joining … 500,000 users is nothing to sneeze at. Tell us a little bit about what you inherited, what you stepped into, what your goals were and what some of the challenges were and maybe even some things that you thought were challenges that weren’t the challenges and things that shouldn’t have been, but were.
Merci: Yeah. One of the assumptions that I came into Slack with was that I had been using it as head of product at this company I was at before. I was like, oh I know all the features that in Slack and I was very wrong about that. I spent the first couple of weeks of my time there … I was hired without there being a team ready for me to inherit because obviously everything is getting slapped together. I was hired as the new user experience PM. That was a role that Stewart and I decided on really because of my games background. I had built really great onboarding experiences. There’s a whole thing that he really believes in, which is that games are basically onboarding experiences forever until the game is over because you’re always learning something new. We very much had that similar mindset around games, so he thought I would be a good person to come in and totally revamp onboarding.
Merci: Most of you probably used the onboarding that existed in 2014, given that founders tend to be early adopters. It was broken in a lot of ways. One of the ways I found out it was broken is that I spent the first few weeks just listening to users, both talking to people but really answering a lot of email in ZenDesk. So going through and using that as a way to learn how to use the product. I would get screenshots from people who had been in the beta group and people who had been using Slack for four or five months and they would, in that screenshot, still have the little throbbers, which were the indicators of somewhere to click, that were a big part of the new user experience that was out at that time. So it was like, oh shit. No one is seeing these. That’s the interesting tension between product and design is subtlety is beautiful but the point at which subtlety becomes invisibility is something to keep your eye out for.
Merci: So these had verged into invisibility and we were getting these power user screenshots with those in. So I put together a design sprint with the team at Google Ventures. We built prototypes and tested our existing onboarding against a couple of other ideas, one that Stewart was really into that was beautiful but I think just too complex, and one that was very simple that ended up being our new onboarding, which then we released and optimized all throughout mostly 2016, which you can see the effect of in our DAU charts.
Mallun: How long did that process take with Google Ventures?
Merci: The actual thing itself is like a four-day sprint.
Mallun: Okay. Then you take all the work and figure out what to do with it.
Merci: Yeah.
Mallun: When you came on board, there wasn’t a team that you inherited. So you did a lot of hiring.
Merci: Yeah.
Mallun: How much hiring did you do and any tips for how to hire great PMs?
Merci: Oh man. I think for hiring great PMs, always have them do work. I think the other thing is they really need to … PM is such a weird, maladapted role in so many ways, where they’re filling in the gaps and they become a communication and prioritization mechanism for your entire company. So who that person is is incredibly important. Cultural fit really needs to be there. I’d hire about half the PMs that worked at Slack at the time that I left, so I don’t know, 14 PMs or something like that. User researchers, including the person to manage the team, UI writing, and then I helped out building up the data teams.
Merci: One of the things about PMs is that a lot of people feel that they’re unqualified to interview PMs. Chances are, at a 20-person company you have one or maybe two PMs. Giving them a real problem to work on that you actually already have and asking them to solve it, asking them to go broad, show a couple of solutions and then pick one and defend it. Then say, okay, what would you do to build this and then what would you do after launch, can reveal a lot of interesting things, including whatever tools or data they’re accustomed to having, especially if they’re coming from a larger company that your small startup just won’t have.
Merci: So people who … This is one of my controversial opinions. As Mallun was interviewing me, I have, apparently, a lot of these. I’ve seen just hard fails from people who have that Silicon Valley passport that you’re looking for. Oh, three years at Google and then four years at Facebook. Fantastic. Not for your 20-person startup. That person, especially people from Facebook that I interviewed were often very reluctant to make a decision without data, especially in enterprise when you have low low volume parts of the funnel that are incredibly important, like conversion to paid. And especially for startups. You’re just not going to have … There’s no such thing as a five percent test. There isn’t a one percent test. There’s all of your precious potential customers and nothing else, right. You should be very conscious of protecting their experience and not just flippantly testing things.
Merci: People who have worked at large companies are often not aware of the amount of tooling that has been built over the last decade for billions of people and really time tested. Then when they come in and they’re asking for all of this data, it’s hard for that person, not impossible, but hard for that person to transition to, okay, let’s just spin up Amplitude or let’s just up Mixpanel and get some initial data and then I’ll run some usability or play sessions or do some customer interviews on my own. Bridging that gap can be really difficult, which is why I think just startup experience is actually one of the key things that you should look for, even if it’s not a startup that you necessarily recognize.
Mallun: You mentioned that you give little problems or tests to people. Maybe not tests. Let’s dig out a little bit. When you do that, do you make them do it on the fly there or do you let them take it home with you?
Merci: Let them take it home, yeah. I don’t why we do this, but someone in recruiting told me that I should do this. I’m like, I believe you because this is your job. They said, “You should time box it. Only give them three days to do it.” Then people would write in and ask for more time, which we always gave them. We always set the expectation that, oh this should only take a few hours, but if someone is really motivated to be at your company, it’s okay if they spend 10 hours on it because that’s actually a great signal that this is someone who really gives a shit and is really gonna … Is a craftsperson who really cares and is gonna put a lot of time into it.
Mallun: Speaking of user testing, you don’t wanna necessarily use your users and give them a bad experience. So some people use outside user testing. I know you have some thoughts about that as well.
Merci: Yeah. I think one of the things that I strongly believe is that, first of all people are often dismissive of qualitative data. They’re like, no I want quantitative data. That’s just qualitative data. But the truth is that the human body is an incredible information processing machine for the data of other people. We have evolved for a very long time to understand what a little shift in tone means or a little flash of color or a small twitch of your eye. User researchers are this incredible resource for you to put as a listening device in front of people who are just spewing information all the time.
Merci: So I think one of the things that people often want to do is to say, oh well, but my sales guys are bringing back information to HQ. But they’re not really listening. They’re selling. And they’re listening for the name of the competitor, for the problem that matches the use case because they have that great line about the problem that matches the use case ’cause they have that great line about this feature, which is awesome. That is their job, but it’s not their job to just be listening for things that potentially you don’t even want to hear. So, it’s incredibly important to work with user research early on. You don’t even necessarily have to hire someone full time to do it. And then we were talking about things like user testing and not to hit on a specific company, but I found that resourcing people to participate in these tests is very important. It’s important that it’s not just like this every person, especially early on. It should be your early core customer. Especially for enterprise companies, it’s gonna be someone who’s pretty specific.
Merci: It’s very worth your time to go trolling on Reddit and in Facebook groups and on LinkedIn and in places where people who really care about their careers are hanging out. And you would be surprised how people open up. Everyone just wants to be heard and they’ll tell you everything that they’ve ever thought about their job if you just give them the space and are silent and listen to them.
Mallun: I’m giving you notice here, so in about two questions, I’m gonna open up for your questions so get them ready, raise your hand and we will try and get to as many as we can. Okay, so I have a couple more questions. What product is not out there that you wish were out there?
Merci: Yeah, so for product managers I really with that there were something that made this resourcing for user research much easier because that’s one of the key blockers that I see. And then I’ve been doing angel investing, particularly around blue collar, pink collar work tools because I think that something like Slack is going to emerge for pink and blue collar workers. And then I think that the massive opportunity for that company is going to be in flexibly deploying these workforces into other places. Imagine a future where I’m a sous chef and I’ve worked at California Pizza Kitchen, I can probably take on a shift at a cute little place in the Mission District because I have a good work reputation, I know what I’m doing, I’ve worked at Starbucks, I can also work at Pete’s. For most of the working population, there isn’t this sort of work reputation that follows them around to say oh, this is a reasonable person. She’s already passed a background check. And businesses of all kinds need flexible work, especially as flexible work becomes just the norm and you’re competing against things like Uber or Lyft where if someone just puts in more time, they can earn 10000 dollars a month if they need or want to.
Mallun: And are there particular areas that you’re focusing on with angel investing and/or advising companies?
Merci: Yeah, mostly, so things that get into the hands of end users at the end of the day is one of the things that I tend to specialize in both because of my understanding of product metrics but also how important the product experience is, and then anything that’s a self-service adoption as well.
Mallun: So, two questions. Two more questions. One is you founded a little group a couple years ago or maybe three years ago now called Women in Product and I think you started it by just posting a LinkedIn post. Is that what happened? Or tell me a little bit about that. But it’s grown to over 5000 women now. So what motivated you and how did you start and what’s the focus and how might some people get involved?
Merci: Yeah, I had been thinking about doing something like that for a while. So before I joined Slack, I had been working as a product manager and a founder for almost 10 years at early age startups because I’m a startup kid. And that’s also where the out-sized returns are, right? If you get in before the B or the C at a company, then you get to be an early employee, which is a life-changing experience. So that was what I was out consciously looking for and I realized that in that time, before joining Slack, I had worked with one other female product manager. And I was often the only girl in my company. And I can hang. I literally designed games, which is a very kind of boy thing to do. I like D&D and I like science fiction. I should note that the production company that I worked at was Bad Robot, which made Star Wars and Star Trek, so I’ve only worked at really cool companies.
Merci: But I often found that I was negotiating this really fine line with the guys that I worked with where it’s like friendly but not too friendly, assertive, but not bitchy, and it’s a kind of claustrophobic place to find yourself in. And now, only really in the last five years, it’s become okay to be a woman, okay to be out as a woman. So when I was raising money, I often talked to women who were 10, 15, 20 years older than me who wanted nothing to do with being a woman and they would explicitly say to me, listen, I’m not going to join the ghetto of women’s issues. And now things are different, thank god, and I realize that there wasn’t any sort of place for me to find other female product managers.
Merci: And I also had kind of an open question for myself because I had worked with so few women in product, was like how many of us are there? Is this really the kind of out layer that I think that it is? So I invited, I started the Slack group. I got a Slack group for free at Slack, which was a nice employee benefit, and I invited a few people and then actually this guy on Twitter, men on Twitter can be helpful, suggested to me that I write a blog post and put up like an application and everything and I was like okay, that’s a great idea. Thank you. I will do that. So I did that. Shout out to that guy. And then people just started hearing about it and really flooding in, so now there’s 5000 people all over the world, some really fire channels.
Mallun: You guys have conferences, you have gatherings.
Merci: Yeah, there’s a different group, we started a non-profit around the same idea and started a conference, which I’ve never been to but seems very cool.
Mallun: Great. So, I’m gonna save my final question for a little bit later. So we’re gonna go ahead and open it up for questions for about 15 to 20 minutes. So do we have any questions? I see one over there. Actually, do you mind speaking into the microphone? Thank you.
Audience Member: Thank you for your insightful talk. If you’re trying to develop a product that hopefully by the end has a lot of features but you have really limited developer resources, how do you go about deciding which features to develop first and allocating your resources so that you’re not building a product that may or may not be useful to the end user?
Mallun: Oh, one thing. Can you introduce yourself and maybe what you do in your company?
Audience Member: Oh, sure. My name’s Rebecca. I’m the founder at the startup Flodesk and we’re working on a way to make email marketing super simple for the end user and focus on design and templates rather than on super geeky logic and flows that confuse a lot of people. So we’re just trying to simplify a process that’s previously been very difficult.
Merci: Great, well it sounds like you know who your early core customer is, which is I think step zero for anything. Having a very clear idea of who’s this human being that you’re building for and what is their actual life like is really important because it’ll keep coming back to that as you are successful or fail to make other features stick.
Merci: And then I think the other thing for you to consider is sort of before you even think about something like product metrics, the first thing to think about is when someone is using your product, what is that like for them? So if you imagine a woman who is using your product and she’s going through it, like what are the key things that she does in it? And not what are the key things that someday it would be cool, but what is just the absolute nuts and bolts that she would have to be able to do and then write that out on a list and then as you are speaking with your early core customers, even in the prototype stage, and this could be … People will respond to anything. It’s sometimes shocking. But even drawing out a little wire frame on a paper or whatever and saying, oh what if this did this? And then what would you do next? Then asking a bunch of open-ended questions can be very revealing to you ’cause what you’re looking for is what is this core loop that people are gonna come in and do over and over and over again, and it’s probably just two or three features.
Merci: You’re going to need more to sell into enterprises and things like that, but ultimately for Slack for instance, you come in, you can talk to other people, and you can add files. There are so many more site features than that, but that’s the core loop of Slack, right? Cheers.
Audience Member: Hi there. This is a great talk. Thank you. I’m Tim Murray. I’m from the sales side of the software house. You had mentioned earlier working with enterprise sales organizations, so could you just talk through best practices that you’ve established and how you’ve built this strong relationship between product and sales? Because I’ve seen good, I’ve seen bad, and it sounds like you’ve developed really good ways to work amongst these two groups and I’d just love to hear your thoughts.
Merci: Yeah, sales always had feedback, which is great because they were often in the types of companies that we weren’t in. So Slack is a self-service product and there’s also the enterprise product, enterprise grid product. Those are two different users and people who work at Slack have very little experience usually, even being the type of person who works at something like IBM or Comcast, very similar job, but the network of facts and the way that people socialize with each other and things that they need are hugely different.
Merci: The other thing that we did was we added in some segmentation and onboarding in order to give our sales team better and better data because then they would often to go into the C-Suite of this larger company and be able to say, oh hey, we’d like to sell you enterprise grid. By the way did you know that you have 100 pay teams at your company already? Everything gets easier when there’s that insane amount of product market fit.
Audience Member: Oh thanks. My name is Fred. I’m a founder of a company called Squad. We do feedback for, feedback insight research for fast fashion and apparel brands. So your comment about the onboarding thing is really fascinating to me ’cause that’s what I’m trying to figure out. But like you said, you don’t have a lot of users, you don’t have a lot of testing capabilities. Do you have some Cliff Notes in terms of the process for thinking about onboarding?
Merci: I mean isolating the few things that people really need to know about your product is really important. So, oftentimes the onboarding flows that I’ve seen fail are ones where people are like, let me show you everything. And they like oh, this thing, and then you’re on like step eight. And it’s easy to look at someone else doing that and be like nope. We’ve all been there. You really need to have just three things, ideally. So people can remember kind of up to like five or six things. They can but it’s harder, so just make it easy on their brain.
Merci: And then I think one of the other Cliff Notes that is something I picked up in games that goes really underappreciated is when you’re looking at the screen, mimic what a person is actually doing. So we always have these big displays on our high res computers and our fast connections and our stuff is full screen. No normal person is using your product full screen. They’re certainly not scanning around for anything interesting happening. We are very much directed by movement and by our reptilian brain, so another thing is if you have someone in this lower part of the UI and you’ve given them a hint about oh, this is how you open this menu down here, they will not see something that happens outside of roughly like a coffee cup sized area around it. And that’s often the case where you’re like, oh, but I need to show them this part of the UI that’s over there. And it’s like one, they should make you consider whether if those things are really so interlinked that they should be that far away from each other and then two, you just need to actually move the person over there. Even if it’s with some sort of light animation or something like that. But people get really zeroed in on actually pretty small areas of the screen and this is true even for mobile apps.
Audience Member: Can I just follow up and ask one more question? So one of the things I’m always thinking about is should we say it or should we just copy the most popular app that does that one thing in front of you, like let’s say upload a photo?
Merci: In terms of like the UI pattern?
Audience Member: Yeah. In other words, should we say upload a photo or should we just make it look as much like Instagram as possible? Because then people are just automatically visually triggered. Do you ever …?
Merci: Yeah. I mean, people definitely try to invent things kind of for the sake of inventing things, which I strongly disagree with. It’s like, let’s not be too clever here. Make it look like upload if that’s what it is. It’s also good though … Oftentimes we think that a glyph is really understandable when it’s not. A great thing to do at the beginning when you’re still sort of in black and white prototyping stages is to get very specific about the language and then that’s the thing to really hone because people don’t read, but people read actually button text and then headlines. So, read that stuff out loud to yourself, show it to everyone that you know, make sure that it’s very, very crisp and then the rest of the visual design will follow that, but get tight on language first. Yeah.
Audience Member: Thank you. On a related note, what sort of- Oh, sorry. Bernard Sled. I’m here with Voxeet in business development. A related question on the competitive analysis. What sort of processes do you put in place? Microsoft Teams just released a new version, you’ve got an open source project and so this is a moving, even though Slack is the most visible, prominent and best product, it’s a moving dynamic environment. What sort of processes do you put in place to track what’s happening and what do you do with that?
Merci: Definitely ignoring everything else that’s happening is the best thing to do. Competition can be very distracting and it’s easy to look at someone else and sort of assume that they’re more successful than you are. Or that the flow that you went through is working for them, so back to the throbber thing that I took out because it obviously wasn’t working during the few months it took us to develop the new onboarding experience, I saw five or six products that were mimicking our onboarding. It’s not just like this one feature you mimic and then it’ll be fine, or ever your one feature. Products are a whole experience that include sales. They definitely include your support team, the include the language on your website, and that’s how real people experience them.
Merci: So I think I wouldn’t pay too much attention to the competition. Someone was recently congratulated me on, “Oh, you guys officially beat HipChat.” And like, we beat HipChat a really long time ago by never caring. And when I joined Slack, I, like a lot of early users, had this irrational love for the product and I really wanted to be a part of this team that I thought must be so special because they built this incredible thing.
Merci: And then, someone was saying, “Why are you joining them?” But look at Flowdock and look at HipChat and there are not any features that are different. And in fact, both of those products had more features and they had things like Verplius that we would consistently hear from people, but it’s not just one thing and it certainly very distracting to spend a long time looking at the competition.
Audience Member: Hey, my name is [Seraj 00:31:45]. I run product at a software company called Stella Service and we do employee success software for call centers. And I’m curious to hear, sort of, as you have buyers who are making decisions about using your products with their team versus users who are actually then using it. How you think about prioritizing and building or not building features that you’re getting asked for? Maybe specifically from the enterprise side, or from larger businesses that are sort of saying like, “Well, we need this thing in order to buy.” And then going through users versus your buyers.
Merci: Yeah, if you are selling to both enterprise and SNB, then I think just like anything else, nail the hell out of SNB first before you try to scale up in enterprise. Because that is actually where the truism of you know, “Oh, just follow the end user and don’t pay too much attention to what the C suite is saying that like they have to have,” is really true.
Merci: People at those giant companies will have already adopted your product, probably, right? Because people who are high performers are always looking for ways to optimize their team, especially something like that. So, you will get to the point, when you’re selling into the enterprise, that you do need certain levels of compliance, you need all of that kind of stuff. And that’s a real and moveable, but really knowable list of things. But ultimately if it’s not something that’s getting ravenously used by a lot of people, then it doesn’t matter if you built that stuff. And it can definitely get you ahead of your skis.
Audience Member: Hey. So I’m Sharon. I’m a growth marketer at Vara Money. We’re a mobile banking start up.
Audience Member: So, as you know, you scaled a number of users and the product at Slack. How did you work together with marketing, and it could be product marketing or growth marketing, to do that together? And you know, what were some of the things that were product specific versus something that you’ve collaborated on?
Merci: Yeah so. We, by the time … I guess by early 2017, that’s when we first really, and kind of late 2016, that’s when we first really had a substantial marketing org going. They had done, the existing marketing team, had done a bunch of brand marketing and then the product marketing team helped the product team launch specific features.
Merci: And then on the growth team, we added a performance marketing team and then we also had a sort of product marketing team that was focused on email campaigns and they took over slack.com, as well.
Merci: And then how we really worked together was we were all part of the growth team together, along with the growth marketers. And one of the most important things that we did was the way that we designed our goals was around activated teams. So those are teams that have reached a certain of users and a certain number of messages within a certain period of time. And we had found sort of end of 2015/early 2016, that if we increased the percentage of teams coming in every day who were hitting that metric, they just stayed. So cohort after cohort, they would retain better, they would convert to paid better, and so we knew that, “Okay, this our seven friends in seven days,” or whatever the urban legend about Facebook’s metric was. And they were gold on that same metric, so they weren’t incented at any time to push in a bunch of people who are not the profile of user and then have us sort of deal with decreased numbers later. And the same thing too for all of the PM’s on our team as well.
Merci: But everyone had the same goal because it’s one team and that’s one of the most important things that you can do. You’re … I mean this a little bit of the game design thing, but think about the behavior that you’re incenting, the good behavior and the bad behavior and you’ll actually be able to kind of predict how people will respond within a system like that.
Audience Member: Hi, my name is Kay. I work at a company called Build Zoom on the product team.
Audience Member: My question is, so at my company, we use Slack on the desktop. I also have it on my phone. What happens when it comes to feature parity? When you think about a feature that really works on desktop, is there a different mobile team? Do they have to figure out how to work it into the mobile app? How are your teams structured in that sense?
Merci: Yeah, so Slack, as you probably have noticed, or maybe not, has really gone for parity on the mobile app to an extreme. It’s like one of the things that I disagreed with. I felt it should be more of an alert inbox and not something that was fully fleshed out. But for the people who like to use Slack almost exclusively on their phones, so people who would otherwise be an email, like executives and VC’s and consultants, they’re really using it on their phone.
Merci: I feel like there’s a lot of downsides to that. It’s much harder to sort of manage everything, but anything that touches the database, anything that touches the core loop, you do need to have parity on.
Audience Member: Thank you.
Audience Member: Hi. Martha with FlowDesk here as well. So for wording specifically, whether it’s games or companies, what are some of your top favorite sources of inspiration?
Merci: Playing games is great. Let me think … Monument Valley is particularly good. It’s super simple. Mario games are great. Nintendo has just nailed this time and again. A lot of the Zelda games are also really good. And that is primarily what I push people toward. In part because they do things that you can’t directly copy and so you’re less likely to go for, “Oh, let’s do this little explainer about the keyboard and just the way that Bitmoji does it because it seems like it works for them. Because again, you do not know what works for them and just because something’s in the product doesn’t mean that it’s working.
Audience Member: As an early employee at multiple startups, how do you decide when it’s time to leave?
Merci: When I get bored, I guess. That’s when I … I mean, if the culture becomes toxic or I get bored, that’s been historically what I’ve done. And yeah, usually when I have something else going on too
