Launch Details and Decisions

Kimberly (00:00):
Welcome to Rework, a podcast by 37signals about the better way to work and run your business. I’m Kimberly Rhodes, joined by the co-founders of 37signals, Jason Fried and David Heinemeier Hansson. It has been a minute since we’ve been with you guys and during that time we have been busy. We’ve launched a new version of Basecamp. I thought we would spend the time today to talk a little bit about some of that process and some of the decisions that you guys made prior to launch. But before I jump into that, guys, kind of give me an overview. How do you think this went in terms of launching a new product? You’ve done it several times. How do you feel like this one went?

Jason (00:34):
This one I would say was the most critical launch we’ve probably ever done because we transitioned everybody from Basecamp 4 to Basecamp 5. So in previous versions, not always, but historically from Classic to 2, we let people stay on Classic and then some people could opt into 2 or new customers could have 2. From 2 to 3, we did the same thing. From 3 to 4, it was more of a gradual slow transition over a number of months where we just added a bunch of new features but slowly at a time essentially. And then with 5, we had hundreds of new improvements that we launched all at the same time. So this was a new thing for us and all things considered went really, really well. I mean, there was no downtime, no big technical issues. We had a few slowdowns here and there, but for the most part it was invisible to customers.

(01:17):
I mean, not invisible. The product was very visible, but the transitional was invisible. And we did it on a Monday night, on a Monday evening. I think it was around 5 PM, which is sort of the best time for us. 5 PM Pacific, I think it was, because the East Coast of the US is kind of off in a sense. I mean, people stay up late and work, but for the most part it was kind of 8 PM East Coast and then Europe was sleeping and we have customers all over the world. So it was kind of a really good time to slowly ramp this thing up, launch it all at once and then let people sort of come on the next day. Now we had a bunch of emails, huge volumes from customers. So that was sort of a pretty hectic, I’d say two, three weeks of answering questions and helping people with the transition and whatnot. But technically everything went really smoothly.

David (02:02):
And I think when we remember back to the other launches that we’ve had, especially the most recent major launch we had, which was the HEY email service all the way back in 2020, that was very different. And we have some new people at the company who’ve not lived through some of that. And then we also have some veterans who remember what the HEY launch was like where it wasn’t just a couple of weeks, but closer to six months when things were quite hectic. That was a brand new product that did not get the luxury of a slow ramp because we had a big fight with Apple that gave us a ton of visibility just as the product went live and with a snap of a finger we had 30,000 people who have signed up for that service. And that was a brand new service.

(02:48):
There was a bunch of things that we needed to figure out and then we had all that stress of the big launch. So in comparison to something like that, this was very mild in terms of the issues that we had. As Jason mentioned, we had a couple of performance issues, which is always interesting because we had spent actually quite a lot of time testing and probing that, but it wasn’t, well, this screen that everyone uses is kind of slow. It was, oh, here’s a feature, for example, a bookmarks list that hadn’t been paginated. And there just happens to be some people who have thousands of these bookmarks or thousands of bubble ups set aside and, oh, we hadn’t tested that. So that was kind of the cleanup that we have to do, the kind of cleanup that you can only sort of really do once you go live because all your test data, all our own internal use, even all the beta tests that we had on didn’t find those things.

(03:42):
That’s the magic of going live for hundreds of thousands of people that you get to see all those things. But still, given the criticality that we were essentially replacing the engine of the airplane mid-flight, there were trillion things that could have gone wrong and they didn’t. Part of the reason was we gave ourselves a little extra time. We actually had an early launch date that was three weeks earlier and the company had been scrambling and pushing for months to hit that date. And as always, when you set a date, there’s a big push right up until that date and then you’re like, “I need a little more time.” If we had set the dates three weeks later, that moment would have arrived then, but it arrived at that time and we decided, you know what, deadlines in our industry, certainly for us, are made up. We picked that date, not totally out of a hat.

(04:43):
It was somewhat informed by our intuition on how long it was going to take to wrap things up, but it was still arbitrary. So we actually did have a bit of a debate back and forth, Jason and I and Brian, about whether we should try to push through and just ship it or whether we should give ourselves another few weeks and spend the meetup together to really polish this thing all the way. And I’m really glad that we did, that we spent those extra three weeks, but that whole phase could only have happened because there was a date picked that we were rushing towards and then you can do that stuff. If you just get yourself into when it’s ready, well odds are often that date just keeps disappearing out into the ether because you’ll never just feel like you’re totally ready and everything is bolted down. There’s got to be a little bit of tension, a little bit of stress, a little bit of like, “Oh, I wish I had five more minutes.” And then just give yourself two minutes, not five.

Kimberly (05:45):
And also add that earlier deadline, like the three weeks before, that was a date no one knew. That was an internal date. It had never been publicized. I’m assuming that was intentional for that very reason.

Jason (05:57):
Yeah. We tend not to share release dates with customers until we are absolutely certain. And with the transition, this time we had to be absolutely certain at some point because we had to give customers a heads up because this wasn’t an opt-in situation. This was a, you’re going to get this new version. And so we had to pick a date at some point and we just, as David talked about, we finally picked a date after the meetup. It was actually right after the meetup. It was a Tuesday after and about a week ahead of that we started telling customers exactly which day it was going to be. Prior to that, we’d been talking about it with customers for a couple months. So we’d give them some heads up broadly. I did a big video. There were some other posts, but we kept saying like, in the next few weeks, we’ll be launching this because we didn’t know yet. And once we did know for sure when drew that line in the sand, we did tell customers the day, but it was about a week ahead of time.

David (06:46):
What was important about that date too was that once we’d picked it, we had this point of no return and that pertained to the mobile apps. We had designed new mobile apps for Basecamp 5 and they were not backwards compatible. So once those new mobile apps went live in the app stores, we could not roll back. We could not go back to the old version because the new apps didn’t work with the old version and the old apps didn’t work with the new version. So we kind of had this hard line where we just had to say, this is now. We’re doing it because we have to submit the apps to the app stores. And then once they’re approved, that doesn’t roll back. So we had to have some surety that this was going to work at that time. I think this is another reason we were quite glad to give ourselves those extra three weeks just to nail down that whole transition because on top of the mobile apps, we also changed the domain name and that sounds like a really trivial thing.

(07:45):
Well, you could just change the name. Well, we have a bunch of things that point to the old name. We have APIs, we have the mobile apps, we have a constellation of helper app, something called Launchpad for identity verification. It was actually one of the major technical, not unknowns, but risk factors, because if you screw up DNS and the new domains that the mobile apps expect to be there aren’t working correctly, the whole thing just comes apart. I mean, that’s one of the jokes in operations that it’s always DNS. Whenever anything goes wrong, a lot of the times it’s DNS. It’s these names that point to specific IP addresses that get screwed up in some way and we were doing it intentionally. We were screwing it up. We were screwing with it and we moved from 3.basecamp.com to app.basecamp.com, which was something we had wanted to do for a long time.

(08:40):
I mean, Basecamp four ran on 3.basecamp.com, which is always sort of a little bit weird. I mean, the reason we were doing the 3 was we had done two previous versions of Basecamp that were complete rewrites and when we picked 3, we were like, “Well, maybe that’s also going to happen at some point in the future.” And then it just turned out that the chassis we now have, the technical underspinnings that started in Basecamp 3 have been so resilient and so malleable that we’ve been able to not just take that set up to Basecamp 4, but take it all the way to Basecamp 5. And if you compare the version of Basecamp that launched with Basecamp 3 with Basecamp 5, they look like very different products because they are. I mean, there’s, what is that going to be? 11 years between them on the launch date, but a lot of the technical underpinnings are actually the same.

(09:34):
So we’ve been doing that transition moving forward to, do you know what? We’re probably not just about to throw it all out and rewrite it over. In fact, just this morning, Jason had noticed that our APIs were still referencing. it was called BC3-API on GitHub and the same thing with the integrations and I just snipped out that little three in the mornings. And that’s just BC-API and BC integrations. But that’s the price and in many ways celebration of having been around for so long. I mean, Basecamp launched in February 5th, 2004. I mean, 22 plus years ago is when we went live with the first version of Basecamp and here we are still evolving that concept. So, you are supposed to pick up some scars. You’re supposed to have a little bit of marks on your technical infrastructure and so on. It’s remarkable how little of it that we have, but you shouldn’t look upon that as like, “Oh man, that really sucks.” It’s like, “No man, that’s amazing. Jesus, you made it 22 years on an application that still have hundreds of thousands of people using it all the time?” That’s incredible. Pick up a couple of cigars and be proud of them and a few gray hair too maybe.

Jason (10:55):
Yeah, lots of them. By the way, if you go to basecamp.com/live, L-I-V-E, you can actually see how many people are using Basecamp right now and where they are in the world. That’s something we just launched recently, which is pretty fun.

Kimberly (11:07):
Yeah, it’s very cool. Okay. So you mentioned that we pushed the launch date a couple weeks from our original internal date. I’m curious if you wish you had even more time. Now are you like, “Ooh, we could have used two more weeks,” or are you like, “Nope, that was the right decision.”

Jason (11:22):
No, the answer is no, because you need to get the thing out there. But another version of that answer is yes, you always need more time if you are trying to get to this point where everything’s perfect, which is an impossible point to ever hit. It’s like trying to cut the distance down by half every time you’ll never get to the end. It’s impossible if you do it that way. There’s always stuff you kind of wish you had more time, but you also don’t actually want more time. You need to get the thing out there because what could happen otherwise is that you never launch, you never ship. Now we weren’t going to be in that situation, but I know a lot of people who launch new products from scratch for the very first time do find themselves in that situation. They’re like, “It’s not good enough.

(12:00):
It’s not good enough.” And that can be true, by the way, for a while. And there is a point where you shouldn’t launch too early if something’s not good enough. You got to find that right point, but deadlines are your friend and also the good thing about software is it’s malleable and you can make adjustments on the fly. And the other thing is the things you thought were going to be important that were going to maybe break or be showstoppers or whatever turn out oftentimes not to be those things and it’s something else. And you don’t find those things out until you ship. And like David was mentioning, we had about 500 customers on BC5 prior to shipping some for almost a month, some for a few days and we picked up a lot of issues. There’s a lot of things that they found for us, but then we shipped this thing to hundreds of thousands of people and they found a whole bunch of things those 500 people never found. You just don’t know until you get it out there in the wild. Of course, you don’t want to get out there in the wild and have it just conk out and die that you don’t want. But that’s a deeper fundamental flaw if you’re putting something out there in the wild and it just breaks down completely. More time probably would not have helped you because you didn’t even see that. So anyway, that’s my general take on it.

David (13:07):
I’d say that’s the thing I love about going live is it actually makes the job of prioritizing what to do next incredibly easy because your customers will not just tell you they will scream it at you. The issues that are showstoppers for some, like you’re not just going to hear vaguely about it, you’re going to hear very loudly about it. And we did and we made some key improvements to the product, things we probably sort of kind of knew, but we kind of sort of knew a lot of things and it wasn’t until we went live that the customer base told us, do you know what? Out of these, I don’t know, 50 things, 100 things you think, oh man, if I had a little more time, I’m going to spend time with it. There’s like five that the overwhelming majority are either really upset about or feel is missing or actually something is broken and this is where we have to focus our attention. Out of the hundred issues, we could have spent months polishing all those off and we would have spent a bunch of time on 95 of those issues that just didn’t move the needle. It didn’t matter instead of being informed by here are the five that really do matter. Now I actually just shared this article from mid 2000s, I think I forget his first name, Buchheit, the creator of Gmail.

Jason (14:26):
Paul, Paul.

David (14:27):
Paul, that’s right. Paul has this wonderful essay entitled A Great Product Doesn’t Have to Be Good, which sounds counterintuitive, but his point is that if you have something that is novel and appealing and as a core that’s really strong and really appealing and people are highly motivated to get that, all the other stuff like the laundry list of checkbox features of things you ought to have doesn’t have to be there. Those are all the things that are about a good product. A good product is well rounded and it has all the table stake features and so forth, but a great product can actually get away with far less of that because there’s a kernel of it that’s really strong. And I think about that essay when I think about a launch like this, that if we can just nail the key interactions, the key reasons why someone would look at Basecamp 5 and go like, “Oh my gosh, so much better.” Through the same eyes as we did, I mean, we had this discussion internally quite a few times as we were getting closer to the end where it was almost painful to go back to Basecamp 4 because the product felt so inferior to what we built with Basecamp 5.

(15:37):
That’s the kernel where you know there’s something great here. When it actually kind of stings or hurts a little to go back to your previous version, you know you have something that’s great and therefore that value needs to get out there. If we just sit on a clearly superior version of Basecamp and we’re just polishing and we’re waiting and we’re waiting, we’re forcing everyone else to use the, in comparison, lesser version. Now that version was perfectly fine for many years because that was the best ideas we had at the time and then we had better ideas and I think you really have an obligation to put your best ideas forward, to make them available for sale. If you just hoard all those good ideas and they’re running on this little beta server and you’re just tinkling for months and months on end or even years on end, you’re robbing your customers of your best ideas and that just seems disrespectful.

Kimberly (16:30):
So let me ask you this, Jason, you mentioned this was the first time that we forced, that sounds like a terrible word, but did not give people the option to opt into the new version of Basecamp. I think that is probably one of the biggest decisions with this version of Basecamp. Kind of talk us through that and why we decided that when there were people who just want to stick to the version that we don’t think is as good.

Jason (16:53):
For sure. And by the way, that was the number one email we got. I’m just being real here, right? Which is we shipped Basecamp 5, hundreds of emails come in, can we go back to the old version, right? Now these emails come in literally four minutes after we ship the new version and I understand the reaction. It’s very natural human reaction. I’m in the middle of something, everything just changed. I know you maybe told us about it, maybe I wasn’t paying attention, maybe I was, but right now I wish I had the old one because I’m familiar with it. So I think a big part of this is to recognize that no matter how much you prepare people for a major change, a major change is going to surprise them or catch them off guard or even catch them fully prepared, but familiarity and comfort is very powerful and what people are used to is what they want to keep using.

(17:37):
Now some people are really motivated to always use the new stuff and the new this, the new that. Most people actually are not. So in our heads we could be like, well, we could just let people stay on the old version and make a new version and let people opt in like we did before. But I feel like what we’ve sort of found is that while there’s some advantages to keeping old versions around, which we still maintain the original version of Basecamp, Basecamp Classic for thousands of customers, and Basecamp 2 for thousands of customers. I really personally at this point, and I think in our career, we want everyone using the best version of Basecamp that we know how to make and we think it’s way better. It’s just way better. Now there are some things that are different and people have to get used to that for sure.

(18:17):
Even internally, we had to get used to some of this stuff for sure, but we just want our customers using the best ideas that we have. And even though some of them may not want to use the best ideas, they may want to use our older ideas. I just don’t want to do that anymore. I think there’s been some advantages in leaving people, allowing people to stay on old versions, but for the most part people are then stuck with some old ideas that I think we can help them with if they moved over to the newer version. So it’s also just like yet another thing to maintain. We’re already maintaining two old versions. Do we want to maintain a third old version? Also, the technology underpinning Basecamp 4 is basically the same as Basecamp 5. In the previous versions, it was a total rewrite from the ground up.

(18:54):
This was not the case. So transition was a whole lot easier actually. And we also were very careful to make this feel, even though it’s fresh, feel familiar in the key points, like the project page, which is sort of the heart of Basecamp, it’s a little bit redesigned, but fundamentally it’s kind of the same thing. You’ve got boxes for tools and yeah, the boxes are a litle bit redesigned, but if you know 4 and you know 3, you know 5 in that respect. Yeah, notifications have been moved around. Yeah, activity looks a little bit different. Yes, some other things are different, but fundamentally it’s not that jarring once you give it just a moment. So anyway, we wanted to give everyone that moment, even if it was a hard moment. And what we see over time, of course, is that fewer and fewer people are asking for the old version and more and more people are like, “Oh, I actually see what’s good about this now. Thank you. Thanks for showing me. Thanks for helping us. This is actually much better.” And it’s been on us to help people see that. But anyway, that’s sort of part of the reason why we went with this direction.

Kimberly (19:51):
It’s also interesting that you say that because I’ve met with customers before who have a frustration with Basecamp and as I’m talking to them, I’m like, “Oh, you’re on Basecamp 2. So much has improved, so much has changed. Let me show you how all the things that you’re running into are fixed in later versions.”

Jason (20:09):
And also just to be clear, people’s perceptions are real for themselves. So if they preferred the other version, the question is when did they prefer the other version? Three minutes after the new version came out, you probably are going to prefer the other version, especially if you’ve been on it for 11 years. It’s a lot of momentum. So for us, it’s a matter of we knew this. In fact, at the meetup prior to us shipping this the following week, I was on stage and I was trying to make it very, very clear to everybody, next week’s going to be a hard week. People are going to push back strongly. This is a major change and we’re making them make this change. The natural reaction’s going to be pushback from a lot of people, but it’s going to be pushed back five minutes after we launch it and maybe five days and maybe even five weeks for some, but we have to just stay steadfast, stay focused and let people get adjusted and get used to the new system. And we think overall the majority of people, not everybody clearly, but the majority of people are going to find that this is a big upgrade for them after they kind of figure it all out.

David (21:10):
I think this is one of those areas and times of the business where having a stellar support team makes all the difference because what you’re selling when you’re trying to convince someone that your new version is better is not just a feature checklist. It’s not just like, “Here’s some concrete, very objective way Basecamp 5 is better.” No, what you actually have to wrap that in is a heavy dose of empathy, heavy dose of skilled support people who can make someone take two steps back from being all the way in the red. They might have had a deadline, they had someone they had onboarded. There’s a million situations that in their lives are already producing a bunch of stress and then here comes a little feather called Basecamp 5 and it lands on top of that camel that’s already packed to the hilt and it just goes snap and you think like, but it was just a feather. Yeah, that’s just the reality of people’s lives, that there’s a lot of other things going on, not just your goddamn piece of software and therefore if you have a great customer support team who’s been experienced in dealing with those kinds of situations, it’s going to go over a lot smoother.

(22:29):
And I think back upon the times where mostly Jason and occasionally me would answer those customer emails and the tone was not quite as refined as it is now. The amount of empathy collectively available for distribution was a small little puddle compared to the lake we have today and I’m just really pleased that we have the setup that we have now and also that this was a situation where the entire company had to jump in because there were simply so many emails that it’s always a good thing to get appreciation for the kind of work that the support team does every day. They’re dealing with like a minuscule version of this. Some customer who’s upset about something and they’re employing their best empathy to deescalate the situation, make them feel good about it, doesn’t always convince them. I mean, this is one of the things Jason and I, well, mostly Jason and then me on CC have been on a couple of email threads with pretty persistent customers who are not the kind who were won over in five minutes, more the kind that needed five weeks.

(23:38):
And you can sort of just see the evolution of that pushback. It can start very ferocious. It can start almost personal as in, why are you doing this to me? My situation is already difficult. My day is difficult, my week is difficult, my year might have been difficult. Why are you putting this on top of it? And it’s just human nature and realizing that like here’s software, here are features, here’s introductions and yes, all of these bubbles float around and then in the center is a human being who just sometimes just needs some really well chosen words to make it past and calm down from the initial gut reaction they have. Maybe sometimes they’ll get a point or two from something that moved, the piece of cheese moved around, they can’t find it, very frustrating and then they get a little clue like, “Oh, it’s over there.” And still mad for a little bit, right?

(24:34):
But they get les mad as they realize that all the things they relied on are still there, maybe they moved a little bit and then hopefully over time they start appreciating all the things that have been added that’s been good. But the key ingredients in that is A, empathy and B, time. You’re not going to convince someone who’s deeply upset five minutes after they’ve seen the change that this is actually amazing. You got to have realistic expectations for how you can flip someone’s perceptions. And I’m thinking particularly about one case that Jason has replied to now three times where I’ve been on CC, where I’ve seen that evolution of just like there were no combination of words that Jason could have written back to this person on day one that would have completely diffuse situation. But we still ended up a couple of emails into that thread over a couple of weeks with maybe not super happy about some aspects of it, but appreciation that, no, we are actually trying.

(25:34):
And by the way, not just are we trying on the software side, but you’re right now having a conversation with the CEO of this company about your feelings about the upgrade. That’s kind of rare. There’s not a lot of major software products where you get to have a long running email thread with the CEO of that company about whether you like something or you don’t like something. So it is an opportunity for us to show all the magic of a small company with a great customer support team and a entire staff, including the CEO, who jumps in and helps customers appreciate what we’ve built or not, right? We can just put our best thing forward, wrap it in the right amount of empathy and then at some point it also just got to stand for itself. You’re never going to win a hundred percent, right?

(26:23):
I don’t know what the percentages are. I think we’re in the high 90s of people who have either immediately liked what we had or eventually liked what we had. And then there’s a core group left where you just go like, “That’s the cost of doing business.” And the primary cost there to realize is that yes, that’s a cost you’re going to turn some customers off, but what we are putting out now with Basecamp 5 is not just for the customers we already have. It’s also for all the customers we don’t have, is to make sure that Basecamp is the collection of the best ideas we have and we need those. It’s a pretty competitive market. It’s only gotten more competitive. So if we want to win over a new customer tomorrow that’s not already invested into everything that we’ve been offering for the last 22 years, we got to show up with our best.

(27:08):
The other option there, and that does occasionally happen maybe even frequently, is that you get a sort of audience capture, existing customer capture, where you get so petrified of alienating existing customers that you freeze the product into amber and then over time it gets less and less appealing to new customers who look at everything that’s in the market, including the thing that just launched two weeks ago from a clean sheet of paper. If you want to be competitive for the next marginal customer, you got to weigh that and say like, do you know what? I’m going to do my very best such that existing customers like what we have and will appreciate it, but I can’t win everyone. That’s okay. It’s okay.

Jason (27:52):
Yeah. I would also add there’s this one case where this one customer, and this is sort of a good insight into like what it’s like to be on the other side of this, which is to be presented with a massive change in a piece of software that you rely on every day and all of a sudden it’s different. This one customer wrote me this incredibly long email. I would say it might have been a couple thousand words. I mean, it was a long email and everything they were mentioning was actually answered in the five minute video, Kimberly, that you did, which was on their home screen in the lower left corner. So for all customers, we had a little home screen video in the lower left corner saying, “Here’s Basecamp 5, here’s what’s new in five minutes,” like a very concise tight thing. And I remember saying like, there’s a video in the lower left corner that explains all these things and they wrote back like, “I don’t have time for a five minute video.” Meanwhile, they wrote a 2000 word email, which I know it took longer than five minutes, but the key was not to make fun of that response really.

(28:46):
It was more like, oh yeah, this is actually a deeply emotional response that makes complete sense. This person is really frustrated and I need to back off here because at this moment I cannot turn them around. This is like, I hear you, we understand, I understand where you’re coming from. There is a video. If you have a moment, this will answer a lot of your questions. Here’s a few answers specifically. And by the way, you can reach me anytime, you can vent more if you need to, you can yell at me, whatever you want, I’m here. I’m your personal contact right now. And just saying something like that, I’m not sure how I worded it exactly, but there’s a moment when you realize you cannot convince someone and it’s not your job to convince anybody of anything. It’s your job to be there to understand and go, “Yeah, it is very frustrating to have a bunch of stuff change on you.

(29:33):
I’ve been in that position. I totally get it.” And I will say that I’m not proud of some of the emails I probably sent to some tickets where I could have been more empathetic and I was probably a litle bit more instructional than I should have been. And in these moments you kind of have these slips because there’s a lot of stuff coming in and you’re trying to help a lot of people and you’re like, I put so much work into this. I think it’s so much better and I want you to know that, but that’s not really your role in those moments. So it’s It’s an interesting experience to have and not many companies get a chance to have this experience. So for those of you who are out there who are listening, who might be in the next year or two launching a new version of something you’ve already made and you have a lot of customers, just keep in mind that it’s not your job to be professorial and teach people when they’re really upset because they’re not going to receive that. You have to be a different version of yourself for that moment.

Kimberly (30:23):
Yeah. And the last thing I was going to mention on that for people who might be launching something new, we’ve talked about everyone on our team jumping into the support queue when we needed to. And it’s like that only came to fruition because we’d had that kind of policy all along that everyone knows how to do that. No one was figuring out how to log into the help queue because it’s something that we have just done as part of our jobs.

Jason (30:45):
Yeah. And the support team raised the red flag, help. We’ve got 700 tickets. We need help. And so everybody jumps in and does their best.

Kimberly (30:55):
Well, with that, we are going to wrap it up. We will hear more about some of these Basecamp 5 updates as we go along, but this has been a production of REWORK from 37signals. You can find show notes and transcripts on our website at 37signals.com/podcast. Full video episodes are on YouTube. And if you have a question for Jason or David about a better way to work and run your business, leave us a video recording. You can do that at 37signals.com/podcastquestion.