How We Almost Messsed Up Trello | Michael Pryor, Trello & Paul Kenny, Ocean Learning| BoS USA 2016

Michael Pryor, CEO, Trello and Paul Kenny, Ocean Learning.

Michael Pryor is CEO of Trello, the tool that helps entrepreneurs organise their businesses and lives. He’s also the co-founder and President of Fog Creek Software, sits on the board of Stack Exchange.

In this interview, with Paul Kenny, Michael talks about some of the things he got wrong building Trello and discusses the reasons things went wrong as well as what was done to get it right.

Video & Transcript below

Video

http://businessofsoftware.wistia.com/medias/p1zu607lkm?embedType=async&videoFoam=true&videoWidth=640

 

Transcript

Paul Kenny: Good morning, everyone! For those of you who haven’t met before, my name is Paul Kenny. I have had the pleasure of presenting at Business of Software I think every year bar one since 2008. I consider it a privilege to be invited, I never take it for granted. I’m just kind of curious from the audience who our veterans are. So, who’s got more than 6 Business of Software’s under their belt? Just shout. I can’t see you cause of the lights. Who is here for their first one? Alright, fantastic! Great!

So, this year I get a chance to do something a bit different and I’m delighted. I get to do an interview with someone I met way back then in 2008. Joel Spolsky invited me to FogCreek to talk some of their staff and I met Michael Pryor. And I’m so excited to do this interview with Michael today, but there’s a huge challenge cause there’s so much that you could talk to Michael about, you could talk to him about the early days of FogCreek. I think I met you about version 5-6 something like that but all we could talk to him about many of the other products, co-pilot, hyper drive and the new product. You could even talk to him about the challenges of keeping a salt water aquarium in the middle of the office and keeping a ton of fish alive, really difficult to keep going. He’s agreed to come and talk to us today about his experiences as CEO of Trello. So, without further ado, could you give a round of applause to Michael? Thank you very much!

Ok so what we’re gonna do is we will chat for about 30 minutes or so and then we’re gonna open up the floor for a longer question session than we might with a speaker. So, as we go through please jot down the questions that you want to ask and if you think I haven’t pushed him hard enough on any of the detail, feel free to go back.

Welcome, Michael! I guess it would be useful – actually how many people use Trello? Ok, you’re ok, you’re amongst friends. So, you don’t have to explain what it is, but perhaps you could give us a brief history of Trello in terms of the users and customers.

Michael Pryor: Sure! You said you’re amongst friends, it’s funny cause I came to the 1st or 2nd conference, came to a bunch in my early days. It’s always been an amazing place to connect with other people and learn about what’s actually going on inside software and running a business that has a software business and I want to give back and be really honest and open and tell everyone about how we got to where we’re at and the mistakes we made to get there.

We’re now – we’re at 16 million people sign up to use Trello, we’ve got over 1 million using it every day, a couple months ago past 10 million in annual recurring revenue. If you go way back to when we first launched it, Joel got on stage and said hey, we built this thing, we hope that 100 million people will use and maybe 1% will pay us $100 a year and we will have a $100 million business. So, it was ridiculous to say let’s build a product for 100 million people, but 16 million checked it out – if we keep going –

Paul Kenny: Are you gonna revise those goals?

Michael Pryor: They are supposed to be stretch goals. We still have a long way to go. I tell people when they join our company I’m always like our mission is to Mars, we’re passing the Moon right now but it’s a long way, it’s like 4 days to get to the moon and 4 months to get to Mars, we’re not sitting on the launch pad, just taking off. There’s a lot of risk that’s been removed but there’s so many things and low hanging fruit we need to go after.

Paul Kenny: Do you mind if I take you back to the launch pad? Because I think it will be – it’s an interesting company. You have this habit of keep shipping, delivering and planning. Great tools for all sorts of market and sizes. I think most I’ve mentioned earlier, what’s your process? Where do these come from?

Michael Pryor: If you go back to the late 90s, Joel and I were working at a software start up in NYC called Juno which was a free email provider at the time. You had to dial up and get your emails and the .com bubble was kind of happening at the time and we thought let’s go start our own company which was ridiculous. I’d only had one job for 2 years and we were young. So we did and we started a company. And the idea was if you were in NYC at that time and you were a software developer, you could only work for a bank or advertising agency, there were no start-ups, that was all west coast, Facebook didn’t exist. You were back office a bit and were like hey let’s build a company around the developers, developers first. And so we started Creek not with a mission to build a particular product but rather to build a place that we wanted to work at. And it was interesting when Gail was putting up that chart with the long slog in the beginning where it’s just flat. It was like day in and day out we’d go and in the beginning we would make $100 this day and then $60 the next day. It was like this long, slow process that then when you look back, you go like the company is growing but it took so long to get there. But that was the idea, build a company that good developers want to work at and just experiment and try different things. And over the years we’ve built a lot of things and one of the things was a content management system back in 2000 that was probably a good product at that time because the right product, people needed content management software, but we decided because people were having trouble installing all these scripts on their servers, we were like let’s build a desktop app that does it. So, the wrong incarnation of the right idea. So, that was one of the first mistakes we ever made, was city desk. And low and behold we had something that we had built in the closet and we were like hey, let’s spruce this up a little bit and start selling it. And that was FogBuzz which was our developer project management tool – our bug tracker, and that just slowly gained traction and traction and it ended up generating a lot of profits for us to experiment with other things.

Paul Kenny: So that was your core?

Michael Pryor: Yeah, that was the driving engine of revenue that has allowed us to experiment with other things like Stack Overflow and Trello and –

Paul Kenny: So let’s dig into that cause I want to get to the heart of the process cause I guess it’s more than a bit of magic and luck to it. What’s the point in which people go wouldn’t it be really cool if we had a product that did? Or is there a design thinking process that goes into it? How do you – where does it come from?

Michael Pryor: So, I will talk about Trello specifically and how we got there. So this was later, about 2010, so we had built a bunch of products, learned a lot of lessons, gotten beaten in the market a bit and were sort of coming to this point in time where you’re like what’s next? And we went to – we had these things called creek weeks where we’d spend a week, let’s experiment and do different things. And one of the ideas at the time was…Joel had this idea – as managers you’re like what is everyone working on? I see them come in and are typing code but I have no idea – I don’t understand if we’re doing things at a strategic level or they are solving the problems that are painful to them as an individual? He had this silly idea at the time that was what if everyone had a to-do list but it was only 5 items long? And you say two items you’re working on, one you’re working on now and one while your code is compiling and then two things that are up next to fill the slots and one thing you’re never gonna work on. Everyone can see it and don’t even ask me about this cause I won’t do it.

And he was getting – historically we were not too great at naming and we will come back to that. But we were gonna name it 5 things or 5 camels [laughing], it was animal themes. But there was a nugget of something in there that we were kind of like look, we build an app that lets you track things and wrap your head around development projects. So, what is the problem that we’re having as managers with that detail– we’re trying to raise it up and be like I don’t wanna constrain people – just be like what are you working on right now? And that was the idea of 5 things. That kind of merged with, as we went around and saw all these start-ups and incubators you walk into one of these offices with like 20 start-ups working and on the wall would just be sticky notes everywhere. We go in our own software development office and they would have sticky notes on their whiteboard so everyone was doing it. And so the trick there was then we said they’re getting a lot of value out of this and we decided that the world can get a lot of value out of this, that this metaphor that people were using and the level of perspective that 5 things gave you was important to people to wrap their heads around something they were working on and that everyone could understand the sticky notes on a whiteboard thing. And if we were gonna build that, there were already 1000 tools out there, we weren’t building a tool for developers, we decided let’s bring this to all teams, doesn’t matter if you’re in sales or marketing. In fact, developers are not our market which is tricky for us because pretty much everyone that knew us were developers.

Paul Kenny: I was gonna ask because this is a real departure, because all your products to date have really been for the fan base, people who you knew and had automatic communication channels to them. How would you go about articulating what a product like Trello is? That’s so much more –

Michael Pryor: I think the idea is we were competing in the developer tool space for so long and doing well. Clearly Elastic at that point was doing much better than us and we were trying to think in the future, what do we want to do? Where do we want to be going? And building that horizontal tool had never been something that we did before. So, it was hard because who are you targeting or marketing or selling to? It’s everyone? It’s kind of great, cause your target market is huge but it’s awful because you can’t target anybody individually. We knew we were gonna get these feature requests but early on, a lot of the things people asked us for, were coming from this developer/project mindset, tasks, milestone – they would say how can I do dependencies in Trello? How I see my gant chart or milestones? And it’s like all that language that you’re using is not what we’re doing with Trello. So, we had to listen and sort of go below the surface of what they were asking for, not the feature but look at the workflow and find the pain point they were having. Tell me about why you need that.

Paul Kenny: And how do you do that? Was it that your devs talking to the users, networking?

Michael Pryor: Yeah, support, sales and PMs. And in the beginning when Trello was 15 people it was the devs because we were PMs and support and we were doing everything. But it’s just customer interviews, that sort of things. We actually had a Trello board up in the early days that was a 2-way communication channel for what was happening like the roadmap basically. And the way it worked early on – this is gonna be another mistake we made – is you could just go and put up a request or you could vote for an existing feature request. And I think very early on, at an early stage where you don’t know which direction your product is going that can be useful because it’s a way to collect information and hear from your customers. But as you start to hone in and see where you’re going it can also create friction cause there were a ton of people that ask for dependencies and it was not something we were building. I wasn’t building a project management application, but we were trying to define a new category of applications and I get that a lot of people use Trello for project management but that was a tiny piece of what we wanted people to use it for. So, that board actually evolved over time to be less of it – vote for what you want and you will build it and essentially that was the message we were giving them early on which worked for a while to here’s what we’re working on and what’s coming up. And so it took us a while to get there and caused a bit of pain in the interim because there were people that said this feature has 1200 votes on it. Why haven’t you done it? Well that’s not Trello.

Paul Kenny: Do you tell them directly that we’re not doing it or just let the interest wane once you decided to do this different channel of-

Michael Pryor: Yeah, we changed the way the board worked because we needed to have this conversation with people, it wasn’t gonna be just vote for the feature you want cause wasn’t helping us understand…cause we had to build features at a very horizontal layer, if I’m gonna build something I need the sales and the marketing persons to get value out of it. So you may ask for something from your sales perspective for example you look at a Trello list and how many deals in a list? Each car is a deal and I want to know what my pipeline looks like. What’s the total value of my pipeline? Some other developer comes along and he’s like I put all my tasks in the list and they all have story points and I want to know my sprint, how long it will take and that kind of things. Both things are great, you could have feature that adds up and will tell you the total on the top. That’s a very simple way of understanding that there’s a connection between and it’s two requests but if you just listen to the person that’s asking it, all you will get we have to have stories on Trello. Or we have pipeline values –

Paul Kenny: Ok, I can move on a bit and you touched on the name Trello and I think this is an area that people at this conference most of them have had naming nightmares of some sort or the other and I was in a meeting with an events organiser in London and halfway through this meeting, the account director leaned across the table to the manager and said have you got all that? And she said don’t worry, I’ve Trelloed it. You managed – when you become a verb then you’ve hit the nail on the head. Where did it come from and what went into the creation of it?

Michael Pryor: So look, FogBuzz is an awesome app with an awful name! I think like it’s hard, naming is really hard. Especially when you’re trying to own the .com and you have these constraints like not misspelling it because if you tell somebody the word and spell it wrong they end up to the wrong place. You don’t want it to be too long and want it to be unique. So when we had named Trello Trellis internally. That was the codename of the project. So early on it was called Trellis and the reason is we were trying to provide a structure for people to put their process on top of. It was just a building block for you to apply your process to like a Trellis is a structure for plants to grow on. So when we got to it and knew the deadline was coming up and had to have a site, we tried every manner of that name, we tried to get trellis.com, trell.is, which was less – you really want to own the .com. So we bought names and a list of names from people to go through, we tried to crowdsource our name internally – Stack Overflow was crowdsourced from the community and that worked out fine and it was pretty long – but we got to this point where we were two weeks away and had a deadline – we got to get this site up and pick it, and we weren’t getting any closer to a final name. So we got everyone together in the kitchen of the office and we started to have people vote on different names and this is the worst idea ever, like [laughing] it’s like crowdsourcing design, there are some things you can’t just have everyone vote cause you will get the least objectionable name instead of a good one and the ones we came up with were awful – I will tell you Trello was this far away from being named Planatee, which a combination of planning and manatee [laughing].

Paul Kenny: Is it true that Cardvark was also –

Michael Pryor: Cardvark was the other one! Cause we like mascots – so this is Taco the husky which is the name of Joel’s actual dog who is a husky. But before we had Taco we had Amanatee, that was the standing mascot. So people were like it’s fine I guess – we had been through this process and they just wanted to have it over with. And I said to Joel, I was like come on, we can’t repeat the FogBuzz thing, we can’t do this! And I had come up with a Trellis code name so I was watching it go from Trellis to Planatee. So he was like you got 2 hours, come back with something different. What do you want me to do? We have a deadline. So I went back to my desk and one of the early employees who was working at Trello then and I just sat there and we were in this instant domain search website where you type things and it tells you if it’s available, you type trellis and all this different stuff and we’re just a bootstrap start-up and it’s not like we got a lot of money so what domains are available for a couple of grand. I did this for 30 minutes and sitting with my head in my hands and I’m like ok that’s it. I guess – I typed trellis one more time and an ad pops up for trello.com. and it just – it was like short, you can’t misspell it, it’s weird and sounds like jello, hello. So I ran through the office and I was what do you think about Trello? And people said the same thing, it’s fine. It sounds like jello, hello…but no one hated it, whereas Planeteee there were people saying this is the worst mistake you will ever make. So we were like it was I think 3000 or 5000 or something but when you buy it through the brokerage service, you buy it and have to wait cause they go back to the seller. I was like buy and then waiting like please and we got it so that’s where we came with the name. And the idea now it’s interesting because I think if I could go back in time the one thing missing from the name, it’s short and easy spellable and easy to say in other languages. The one thing it doesn’t do is it doesn’t give you any implication of what it is. It doesn’t imply what it’s for or what it does so that’s missing. But because we’ve gotten so much scale and people were using it we got ahead of that and when you see everyone that clones that boardless card metaphor, Microsoft has a clone, it’s a straight up clone of Trello but other aps added that visual interface to their apps and when they copy that, when the press writes about it, they always write about a Trello competitor or Trello like. So I think we accomplished that goal, there’s a lot of people that haven’t heard about it, but we covered a lot of ground there.

Paul Kenny: That’s part of the core that you now protect. Ok, so you started off by giving it away and I guess the first people to use it were your fan base, the developers and there must have been a lot of organic growth just as they invited people to their boards. And after this early spike of growth, what was it that made you decide to start charging when you did?

Michael Pryor: So if you will build a tool for a horizontal market and especially the collaboration tool like it’s – we were sitting there saying it’s weird being at BoS cause back in the day it was charge for your product. Basecamp and it was like don’t give it away for free – charge for it. Why did we give it for free? The market we were targeting was horizontal, the fact it was collaborative meant we didn’t want to put any impediments in the way. Pricing is an impediment. So we were like if we want to be Trello and go from nobody knows about this to where people never heard of FogCreek, how do we do that? It had to be free, we had to give a tonne of value to people. That was when Joel said 100 million, 1% of the people get more value out of it and pay 100, that was the model. Big number, charge a small fraction, make a bunch of money.

But we were making a decent amount of money so we were self-funding the creation of Trello and weren’t worried about monetising at this point in time. We were just focused on the product. Couple years in, we started hearing people say I’m not gonna use that because they will just shut it down. They aren’t charging for it and it’s gonna go away. This was at the time that google reader shut down, there were a couple of free aps that people bet on and were shut down and it turned out that not charging people was actually friction to them using the product. So we stopped for a second and laid down a dirt path so people can understand where we’re going with the monetisation. Later, we will pave it and make a 4 lane highway but right now it’s just gonna be a dirt path, let’s not spend too much time on it, just give them a way to pay us. So we went around a bit on that and what we were gonna do for pricing models and one of the arguments that we had was we didn’t want to invent something that was totally inorganic, like charge per board or charge per card – because that was something that you have to explain to people. If you’re a storage provider and people get you pay for storage, this is a thing people understand. But in Trello no one would get they need a paper board or whatever – so we couldn’t come up with a great metric there, the other obvious one is charge per user – the problem is it isn’t a collaborative tool that inhibits the growth. So Joel said look, we’re never gonna come up with a perfect answer, we just need something for right now. Let’s charge a flat fee per company and it did this I think and we did it – dumbest thing we ever did. I said a couple about other things so I don’t know –

Paul Kenny: I was gonna quote that back to you, what were the consequences?

Michael Pryor: So the problem with this is – and I’m not saying that you should never do this, it just depends on your specific market and target and – but for us, we left it for too long, like it was good for that moment in time cause what we needed was to put down the path but we left it for too long and all our customers grew but the amount of money they were paying us were the same so they got anchored in this small number – in fact we had people that were paying us 4 cents per user per year because of the number of users they had. So when we finally decided – we’re gonna switch to a per user model – it was easier for us to do that for new customers and did that and left the old people for a very long period of time but at some point in time it was confusing for the metrics. How many paying customers you have? This number but this portion only pays us this amount – at some point we have to move them into the future. We also left the flat rate – it’s what I call the one price – customers on a specific version of the product that only had a couple features. So, when we added new features, we added them to the new product with the per user pricing model so they were on this very old product, they weren’t getting the new stuff we were working on and it created this weird dynamic so eventually over time we made the decision to say we will turn that off and that’s a process, and you’re gonna break some eggs. I don’t really have great advice on how to do it perfectly. We gave people warnings and discount codes but some people wrote me emails this long and I responded to each one say look, I’m trying to run a business and you are paying me 4 cents. It was always the people that were paying me 4 cents. I literally said I don’t want you to pay me, use Trello for free, it’s barely any different than what you pay me because at that point in time there wasn’t much in there, so let me convince you to pay for the thing that I’m selling now. But luckily they didn’t have to stop using Trello, they could keep using the free version because we were basically – they didn’t have to leave, they just had some features disappear and we weathered that storm and we’re still in the process that every month people that were renewing aren’t renewing.

Paul Kenny: You have a sense on what proportion of people migrated from that $200 flat fee?

Michael Pryor: I don’t know percentage wise but it’s more than 50%, I don’t know exactly.

Paul Kenny: Cool! One of the things that I wanted to ask you about because I like many people in this room get my regular email and Twitter updates from Taco to tell me the latest features and the things that they are adding – I think it’s been a consistent theme of this conference and the 8-10 years that I’ve been coming here is people have stood up and said be careful of feature bloat! Gail said it this morning, Steve Johnson said it, Rich Mironov said it, how would you prioritise those features? Given the breadth of your audience.

Michael Pryor: Yeah it’s tricky. I will tell you what our strategy was. If you’re gonna build a tool that everyone uses and you get to that scale where you look at a company like big fortune 100 companies and there’s 1000s of users at the company using your tool, there is gonna be a lot of requests for different things. If you think about what we were trying to build with Trello it was this building block for you to define a tool to work the way you work instead of giving you a grammar and a construct and for you to put your data into. So for example, we could have built a tool that was like we have tasks and milestones and tell me what your task is and I will tell you when it’s due – and the nice thing about a tool like that, a vertical tool, and focused on managing a specific project, is that when you put the data in, it can tell you all the things – you will be late, these people are overdue. But I think a lot of the ways that people use Trello is it depends on their process and their kind of business and they mould it to work the way they work. It was much more about presenting a metaphor which is a list of lists, cards on a board, and then figuring out how to allow people to extend that. So, from the very early days we had this idea of what we call power-ups which as integrations or advanced features. The idea it was that it was a separate thing that you could turn on per board. Basically if you understood cards in a list on a board and you went to Trello you would get that, that was the base level. You didn’t have to learn anything about Trello but then over time, you could turn on these advanced features like custom fields or card voting or aging where you don’t touch a card and it start to look old. In features like that, if you think about a pivot table in Excel, it’s a very similar model where we’re gonna give you a grid, you will enter your data in there, the numbers you want and you will define this cell is a formula – here’s where the headers are, I will put another chart over here and add a sheet. I am defining a model as I go and adding the data, it’s similar to the way you build a Trello board. You don’t have to know what a pivot table is when you get in Excel but as you get more advanced, then somehow someone shows you this thing and you’re like wow, you become like an advanced Excel user and even then when you still think you’re an advanced user and Joel sits down and gives you a talk on why you suck on Excel – which is on YouTube – it will blow your mind and it’s called literally why you suck at excel. You could use Excel for 15 years and go like I never knew I could do that. The idea is to give people these building blocks through power ups and allow them to take aboard and turn it into an app that solves some specific business process and that’s what we’re doing and attacking the problem of feature growth, it’s build these horizontal features, give people what they need to build the tool that they need to solve their job.

Paul Kenny: Ok and who do you take inspiration from? Other products they use for collaborations – are there other people you take inspiration from? In pricing, packaging –

Michael Pryor: Yeah, we obviously work super close with Slack and I think we use it internally and they use Trello – it’s like the way they price their product is we’ve taken a lot of – looked at what they’ve done and copied it. We’ve taken inspiration from it –

Paul Kenny: That’s the term yeah. We will edit that bit out of the video.

Michael Pryor: But those two are complementary to each other. I tell the story about if you were out doing search and rescue in the woods, the things you would need is your radio to talk to people which is Slack but also your GPS, you’d want to know where you were and were going and where you’ve been and that’s Trello so both those things work closely together, about providing the direction and blueprint for what you will do. What is the current state of what we’re doing? And everyone can come and see that on the board. Whenever anyone looks at a Trello board, they see the same thing. It’s not your view and my view, or like I’m putting the database to tell me what issues do I have to do today. It’s like here’s the blueprint for the house, we can look at it and talk about building it – that’s the slack portion, but they work well together for doing something whatever that project is.

Paul Kenny: Ok being the salesman that I am, I have to ask you how you are – how you’re charging and getting bigger clients, how are you professionalising the sales?

Michael Pryor: Yeah, so being developers, Joel and I don’t have a lot of sales experience and if you look back at the early days of FogCreek we didn’t know anything about running a company either, we just learned along the way. But at the time we had to build our own system, so we had this philosophy of we’ll just figure it out, we’re smart people! We tried to do sales, total failure every time we tried to do it. And I think maybe it’s because of our mind set as developers but the one time it started to work for us is when we said let’s hire a talented VP of sales, instead of hiring the sales people first and bringing the VP, let’s hire the VP, have him do the selling, understand the product and let him build a sales team cause they understand that world and we don’t. And that was super successful for us at stack overflow and I took that same model at Trello, I went to an amazing sales woman that worked for us at Fog Creek previously and then gone somewhere else and said we don’t really have a sales team, but would you come and try to sell Trello into these enterprises? We’ve got these companies with 1000s of users, we don’t have an enterprise product, I don’t know what the pricing would be, just go and sell it and see if you could find out. She did, she went out there and sold the product initially it was like a dedicated support person, it was no code, it was like these fun things that you would get and took the price up, up, up to see what we could get for it and over time we started to build these SSO features and user administration features and as we were talking to them and finding their pain points and then we would hire the sales people to sell it. It was this process that we went through that worked really well.

Paul Kenny: And what’s the future at Trello now? What’s gonna make you scratch your head over the next 12-18 months?

Michael Pryor: You said before the woman who said did you Trello that? The idea of becoming a word that people use that they understand that this is a – if we’re the word they use for that metaphor, then awesome! Now everyone is copying that metaphor, that’s awesome too but what’ next? If now that starts to become well everyone presents their information like that, what makes Trello special? I think a lot of the work we’re doing in the recent release we made to make power ups free so all the 60 million people can use them, so it’s around the idea how do you take a Trello board and turn it into a real application for you to use? Like how is this for the sales team, how do they use Trello as their CRM? Not that we’re gonna necessarily build all those vertical apps but that we understand you can take these power ups, layer them on your board and build a process that’s unique to you and solves your own business needs.

Paul Kenny: Is this with external developers working?

Michael Pryor: Both. That platform is opened and 60 million people have access to it. So we’ve had people do it, the time tracking and sales force, zen desk, a lot of slack, bigger companies have built integrations and you will see that more the experience that those apps deliver, people will be delivering that same experience in Trello.

Paul Kenny: Brilliant! I think it’s time to open up to the questions. So if anyone has a question for the CEO of Planatee [laughing]. You’re right it doesn’t roll off the tongue. So raise your hands so we can see – there we go!

Q&A

Audience Question: How is it going? That’s the first question. How are you using usage data? So is it impacting feature requests or are you using it to improve sales? Just curious.

Michael Pryor: Right now, I think there’s a lot of low hanging fruit in the new user experience and how people get invited to boards and a lot of what we’re focused on has been looking at the journey and someone tells someone about Trello and they come into the website, how do they get into the app and learn about what Trello is versus someone invites someone in that specific board. Basically it’s new user acquisitions is one thing that’s helping direct where the product development is going. We look at a lot of stuff on mobile, we’ve been doing a lot of experiments about offline mode and we had this goal very early on that when you use your phone we want it to work offline if you’re on the subway. There’s a lot of different nuances there, do you want sync conflicts to happen when the person comes back up and someone edited it while they were down, how does that get resolved? And maybe that matters and maybe it doesn’t and people don’t care and the last person to change their cards wins. So we tried to study how often that happens, when does it happen, is it worth targeting and fixing? Because maybe people don’t care, they want to know that when they open their Trello on their aeroplane, they can see their board. So it’s very product focused at the moment. One more thing about that, we use a tool called Snowplow that’s been useful for acquiring analytics and then we put that up in Redshift in Amazon and then we used a product called Mode Analytics to write queries or reports on top of it, those three things worked really awesome together.

Audience Question: I am fascinated by this idea of power-ups which adds value to your users and customers and I am curious about how you feel about monetising it cause as a privately held company, if you care about revenue that much or just making a decent living? How does the idea of power ups which adds complexity which can result in new packaging and new monetisation opportunities play in to your revenue model?

Michael Pryor: The question was about power-ups and revenue and we’re in the early stages of what people want to build and how they want to do it. People before we even had the power up architecture, we saw a lot of them building Chrome extensions on Trello and some were selling the service for Agile reports or time tracking so we knew there was a need for people to extend their boards so building that architecture was a part of that, that platform to allow people to build on top of it. The natural extension of that is if you look at other platforms like slack or sales force or their app stores and the app exchanges, maybe working with the partners as a revenue stream. For example, you might build a really amazing cool CRM experience on top of Trello and then be able to sell that as a competitor to close.io or salesforce and then if – it’s probably not something we would focus on but if someone would do it, how would that process be seen. We don’t have immediate plans and there’s other stuff we’re working on but it’s in our vision of the future.

Audience Question: You talked a bit about onboarding and the use case of somebody coming in for the first time and maybe they heard or read about it or heard somebody in the meeting and we’re curious about that. What did you learn about the a-ha moments of someone coming in? Where do they find value? Did you learn something that surprised you? What have you done to get people to that value early on?

Michael Pryor: I think early on, it wasn’t something we ever made a priority because we just had so much organic growth coming from people inviting other people to the boards and we were focused on the product so we didn’t pay too much attention to the website and how they signed up, we have a website and it’s working, it’s great! 100,000s of users a week but as we get further along and were a more mature company, there’s a lot of low hanging fruit there to look at. We’ve taken a lot of inspiration from what quip has done. If you look at the new user experience, they had a dedicated product team to go through every step and looking at how that happens. For example, if you sign up for quip with a google apps address, so now just put your domain is managed by google apps, when you type it in and hit next they show you a button to log in with google and going on the server side and seeing your mx record is coming from google and they are like just come in through google and let you skip the other stuff. You don’t need a password, you’ve authenticated the domain you’re in so you can now connect the same people in the same domain this is all kinds of little tricks that you can do on that experience of coming in and not thinking of it, like there’s a website and signup process and then there’s people in the app. It’s not – that’s not the experience from the end users perspective, it’s this whole seamless thing and I think that iteration changes what we’re going through internally and thinking about right now.

Audience Question: Hi, my question is you said that at some point you decided to bring in a VP of sales cause neither of you had any experience with that. I’m curious how you evaluate that particular person and make sure they come in?

Michael Pryor: Well, we had to – maybe we’ve hired like 5 people in that position over the years cause we had Fog Creek, stack overflow and Trello and two of them didn’t work – I think like to find people that had direct sales experience and a lot of times you don’t know until a person gets in there. And in the case of Trello today, our VP of sales she worked at Fog Creek and we knew her and what she was capable of. And she was amazing and continues to be. So that was a bit easier, even though some of the challenges we threw at her, we didn’t know how she would perform. As far as – particularly for us being a developer focus company and our background is product people, Joel and I, it’s important we get the right kind of sales philosophy and team which is difficult in sales I think – it’s probably getting better I think a lot of the people like Mark from Hubspot and a lot of the education around sales and how it’s done in this new economy where software sales is a lot of bottom up instead of the top down with my Rolex, I roll in to your CEO and CTO and say let’s go play golf! That atmosphere doesn’t happen anymore. You build a good product, the team brings it in and that’s the sales experience. The sales function is getting a lot better and the philosophy around it so I think it’s important to – when sales people come in, we have them sell our product. It doesn’t matter if it’s an actual sales person. We’re big on the audition based interview so we give them a bit of prep and say sell our management team, sell us Trello! They come in and do a PowerPoint or whatever and a lot of times we don’t give them material so it’s interesting what they come up with. I don’t care about the facts they use, because they don’t understand how it’s sold, but the level of presentation they come up with – do they hand you a think they printed out or no presentation? Those sorts of things help you understand is this person prepared? And then at the end of the interview when we say awesome! Thanks! Do they then try to close us? What’s the follow-up? You’d be surprised how many people don’t follow up on the end. All right we’ll put you on the calendar for next week. We look for that – it’s the sales you gotta have the base level.

Audience Question: Tony Pappas and great interview so far! Curious, Gail spoke of the fact that engineering teams, there’s never extra capacity, free time or whatever, a non-busy moment. At Trello how do you go about thinking about true forward looking RD? Stuff that will be monetisable in the next 6-12 months? How do you allocate resources to that plan and get the go ahead to what’s good and pause things that aren’t?

Michael Pryor: It’s hard at a company where you have so many things to do and there’s so many opportunities to improve on what you have and ideas, there’s always gonna be a longer list of things to do, how do you carve out time? I think that’s been a cultural thing from the early days but even in Trello Joel would tell people early on you gotta shoot gamma rays at Trello, mutate it and figure out what works and doesn’t, it’s ok if we shut something down. So very early on, I don’t know if you actually know what a web RTC is but there’s this video chat you can do in browsers and Chrome supports it and it’s built into the browsers, a standard, but it’s still a little janky, it’s not like you’re using zoom or skype or something like that. It’s like kind of breaks sometimes and have to refresh your browser. Early on we were like let’s play with that and see if you can go to a Trello board and then you can hit a button and appear in a meeting. It’s like a video conference at the bottom – we built that and used it internally and the experience was cool when it worked but when it didn’t work – it wasn’t our fault, the web RTC wasn’t prime time yet and so we knew if we shipped it, that that pain would fall on us, people would be mad at Trello, they wouldn’t be you implemented this with web RTC so clearly that isn’t Trello’s fault! They would be like sending support emails of your video chat stinks. I think about that a lot when you talk about the features before and you’re like when you add them, think about are you guys in the developer tools space? Like the big elephant – and there’s a lot of people that are like Agira stinks and hate it. A lot of times it’s very configurable so you can put out this whole workflow, add custom fields and build it the way you want it and in many organisations you have an IT person or maybe a PM that they build the perfect structure and are like voila! Use this tool! And everyone is like this is convoluted, weird and gets in my way. It’s not the tool’s fault, it’s the psychosis of the person that implemented this crazy structure but no one blames the IT person, in most cases they blame the tool and say this tool is too complicated. So it’s a very fine line when you’re building out these features and giving people complexity and what they do with it, there’s some – as developers, interesting ways to deal with that. If you’re in slack and type add channel and then a message, it notifies everyone in that channel and that’s ripe for abuse, you’re like you just woke up people in Australia cause their phone buzzes they have a message and you didn’t even know. And slack added this fun little thing where if you try to do that a rooster pops up and say you will interrupt 22 people in 15 time zones. Do you really want to do this? And then you have to say yes, you can still be a jerk but at least you know you’re being a jerk. There’s a humanisation of software that’s going on and it’s like certainly with all the tools and the way we built software has gotten better and better so it’s much easier to build something today and it’s like what are you offering then? What are your viewpoints and how is it presented? The marketing and brand and voice of your tool. It’s much more important nowadays than before where if you built a database website that was awesome! You know how to code in Pearl? You can make a lot of money – those days are gone, so it’s a tangent –

Audience Question: I was really – touched my heart to hear about the reason you started the company with Joel being the place you want to work. It’s the same reason I created my company 18 months ago. As you’ve gone through the experience of sprouting multiple organisations that came out, any advice for someone who started with that mission? I’m curious about corporations, in your experience with all these organizations is that a thing that would be useful to look at?

Michael Pryor: I don’t know much about b corps – I will tell you this though. That – so we ran Fog Creek for 10 years and when we spat out Stack Overflow, that was a really a joint venture between Jeff who was a very famous blogger and Joel and the programmers but when we spun it off, there were some employees that went with it, including Joel. And now is actually a very tough time for us as a business inside, culturally, splitting off the leadership and some of the employees and it created this us versus them and even though it was the same people that you were just working with and it was hard. And when we did it with Trello we learned our lesson and tried to give people a lot of information upfront and tell them about what will happen and have many questions. It’s much smoother – but having done it twice, I definitely don’t want to do it again. It creates a vacuum, you need a leadership vacuum to run an organisation that like. There’s Joel, me and our time on Fog Creek and there were already constraints on our time, what we’re focused, what we’re spending it on, so yeah. Don’t underestimate the cultural impact and the pain of doing that.

Audience Question: You mentioned that Trello is at the moon right now and headed to Mars. How do you set the rhythm for your team and sticking along that journey to Mars?

Michael Pryor: Specifically, I sit down with every new hire, our group of new hires and tell them the story in person. It’s called story time with Michael [laughing]. As we’ve grown actually it’s interesting because you realise you have to insert formalities for things that used to happen informally. And that’s hit us because we’re mostly remote, 60% of the company is remote so you realise you have to do things that you wouldn’t need to do if you were on the same place. Every Friday we pair up 4 people together to just shoot the shit over a video conference. It’s like a 30 minute work, but it’s not for work, talk about whatever. And it’s just to try and create a connection between a sales person who didn’t talk to a dev cause they don’t work together and they’re not in the same office so they’re not going to hit the coffee machine or they wont have lunch together. So we had to invent things to create connections between people and build that culture. We do town halls every month, but that even in the early days that was a 2-way street and now it’s just a presentation cause there’s too many people so they won’t speak up at a town hall. So now I actually do skip level meetings with the teams so there’s like the individual contributors, level of management and I meet directly with pieces of the sales or marketing team and just kind of talk about higher level vision, what’s going on this week – I try to do that once a month. But as you scale, you change up what you’re doing and realise where there’s friction and try to come up with something to help it.

Paul Kenny: Ok. Michael, thank you very much!

Michael Pryor: Thanks for having me!