By the Numbers
22:45 Share

- 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 Freed and David Heinemeier Hansen. This week, we're talking a little bit about analytics, what we look at, what we don't. We don't have a full-time analyst here at 37signals. A lot of small businesses don't, but we did have that role at one time. So I thought we'd talk a little bit about that transition and what numbers we look at now. So Jason, why don't we start with you? This role of just a full-time analyst, what

We've had it, we don't have it. Kind of talk me through that transition. Yes, we've had two in our history. We first hired one because we,

Thought we should, like, we should know what the hell's going on. Actually, the initial idea was like, I think there was a story that Facebook figured out that if someone clicks this or uploads a photo or something, it like led to something. We're like, we should figure out what that thing is for us. What is the engagement thing? And to do that, we need to analyze probably a lot of data. And it seemed like a really good hunt and really interesting thing. So we hired a fellow named Noah, who was fantastic at this, like awesome at this. And he was with us for a number of years. And he was like,

And then he left for a variety of other reasons. And we just didn't hire anyone immediately, I think. Then we eventually hired someone else named Jane and she was with us for a while as well.

And she left and we kind of realized like at that point, we didn't really need someone again, because what we found is that we were, we had this initial insight or this, not even insight, but this curiosity about like, would there be a trigger? Would there be a thing that if we did this and we did that and we could look at all this data and then we found out that we could look at more data and look at more data and look at more data. It turns out like it didn't really affect much of what we were doing in the end. There's a lot of data you can look at, a lot of data you can mine, a lot of stuff you can analyze and,

But ultimately, if that's just the work that you're doing and you're not really using that information so much, maybe we don't need to do that. Maybe we could occasionally have someone look into something if we need someone to look into something occasionally, but it doesn't need to be a full-time job. There wasn't really enough work to make this a full-time job.

in a way that we could use it. So now we have Ron who runs finance here, who can also look into the data and do some analytics and peek at some things if we need to, like for doing some A-B testing, or we need some data on how much does it cost to store on a product or things like that. We can ask those questions and get those answers. But we just found that we didn't need to replace that initial role. I think ultimately that was a good idea for us. So that was kind of, I mean, it's a very brief history of why, but David might have some more to add.

Yeah, what I love about that role, data analytics role, is that we were convinced there had to be the holy grail. And I remember exactly the anecdote from Facebook, which was if you could get a user to upload three pictures of them and their friends and tag them, engagement was just off the roof. Stickiness was off the roof. It seemed like here was something very specific that you could drive a sign-up flow towards engagement.

And we kept being convinced that something like that had to exist for Basecamp. And Noah, as Jason mentioned, actually did this incredibly rigorous study where we had all these measuring points inside of Basecamp. We got all this data back. And the conclusion was a little deflating. It was essentially, if people use Basecamp with other people, they'll stick around. Yeah, okay, no shit, Sherlock. I mean...

It's nice to get that confirmed, but it was also the same kind of obviousness, perhaps, because we couldn't find the Holy Grail. We couldn't find the mystery path that was going to convert the whole thing around. Another insight I remember from Noah looking into all these things was that Basecamp's feature set is quite rich. We do a lot of things which allow people to use just Basecamp and not like a million other tools. But within that usage...

A majority of it is in two parts of the product. It's in messages and it's in to-dos. And that had the taste of a little bit of something, but something what? Shackles.

Should it just mean that we should focus all our energy on just those two features and perfect them and perfect them and perfect them and spend time on nothing else? Or actually, as part of the attraction of Basecamp is that when you occasionally need something else, like piping in an email, which is probably the least used feature of Basecamp is you can add these email forwards to the product. We don't have that feature turned on by default, but it's still used by thousands of people. When you need that, Basecamp has it.

How do you quantify that? What does that mean in terms of the conversion rate? And that ultimately led me to the conclusion that the problem with data is it can only tell you about things you can measure. And that's such a small portion of product management, of company management, and it can lead you astray. It can lead you into the blind alley of thinking that the only things that are important are the things you can quantify or the things you can measure, which is just a complete

Blind alley. That's not how you should run any company. There's just way too many things that's never quantifiable.

An example of that I love is that you can ask your existing customers, well, would you do this? You can measure them. Would you use this? Would you use that? None of that tells you anything about the customer coming in tomorrow, about whether they're going to buy or not, about whether they're going to react positively to your pitch or not. Some of that has to become from a gut instinct that you have some almost like a truffle nose for finding like a market fit here. And I think that's,

When I look back on our history, 20 years of it, the amount of insights we've extracted from heavy duty, statistically significant data analytics that have turned

the company in a way where it really clicked something? Very, very few things. I actually can't just even off the top of my head think of one. And I can think of a lot more where, especially Jason, would just go like, do you know what? I have a hunch that this is the right thing to do. I can't fully quantify it, but we're going to do it.

And there's something to just embracing that. There's something to just saying, do you know what? Maybe in some abstract world, it makes sense that we should check all the corners of this before we do it, before we jump. And we've done that over time. A wonderful example of this was when we changed our pricing on Basecamp.

We spent like three months doing very detailed analysis to essentially end up at the place where we started. Jason just going like, you know what? I think the price should be X. Why? I don't know. Because the gut machine tells me so. Because I've been in this industry for 20 years. Because I've talked to a million customers. Because I think there's something here that hits.

And data really never gave us those kinds of insights. I think my favorite thing about data is just like sort of the baseline that you don't let the business careen off a cliff. We had one example of that with HiRise, where a different team was running the product for a while. They totally changed the marketing page. No one ran an A-B test and no one even checked whether conversion was better. And like six months into it, we realized we totally tanked

sign up and conversion. I think we were down by like 10%. The whole ordeal cost like, I think we calculated as like two and a half million dollars. I was like, after that, we're like, all right, when we change something major, someone just got to look at the line and make sure it doesn't go off a cliff. That's about level of rigor we need.

And we do do some A-B testing now too. So someone might go, well, I heard you guys did an A-B thing and David tweeted about the 12%. Like we'll do some stuff like that occasionally, but that's not a full-time job. And by the way, there's no analysis there. There's like software that we use where we plug stuff in and the software does the analysis. And we serve three versions or two versions and it tells us basically, we don't need to like hire someone to read that chart. We can read the chart. So we'll do that out of curiosity and also like to get an insight that we didn't know or to try a few things and see what works best. And we'll just run with that. But we're not...

constantly measuring this stuff. We'll occasionally measure something that we think maybe might lead us in a better direction. Okay, so we've mentioned that we're working on some products, Fizzy. We've talked about that recently. Tell me about any kind of numbers you're looking at as things are being built. Like, are you projecting how something is going to go? Are you looking at the potential success of a product? Or you're not even thinking about numbers until after something goes live?

Yeah, we don't really think about it. We're just trying to do the best we can and build the best product we can. And then we'll see what happens. That's just how it always is. And also, if the product doesn't work, let's just say, or it works or there's a mediocre success or...

or whatever, it's still worth doing. We're not doing these all the time. It's not worth doing a bunch of things that never work. But sometimes it's worth doing it because you learn something new. You bring those things into other things you're doing. They kick off new ideas in your head. Like making something new is permission to think something new. It's kind of hard to think something new sometimes about something that already exists.

So these are exercises as well in expanding our mind, our possibilities, our interface design, our technology, infrastructure, all these things, which hopefully they will work in this new product or whatever. You really hope they will. And I certainly hope they will, but you don't know until you do it. But even if they don't, they're worth doing occasionally. Definitely. Also just as a cure for boredom, frankly, I think this isn't something people talk enough about. Business is boring. It's

It's pretty boring, actually, a lot of the times, especially if you're working on something that's working well. It can be actually quite boring because you are afraid to change it. You just don't want to lose anything. So it's nice to make new things occasionally. But yeah, we go into things hoping they're going to work. We believe they're going to work. To what degree, we don't know. It's always best, I think, in a sense to go in with no expectations that are quantifiable.

That way, it's hard to be disappointed, actually. There's only upside, in a sense. I mean, granted, you could look back and go, wow, we spent six months and this thing did not work at all. I could find the downside in that. But I kind of prefer to find the upside in that. And then if it doesn't work, we maybe make some changes. Maybe it works then, maybe it doesn't. Eventually, you just kind of move on to something else. And that's fine, too. But we go at things differently.

hoping it's going to work because there's no reason to do it otherwise. There's really, you got to put your hopes and aspirations into the thing, but you don't have to have a goal in mind beyond just doing the best you can. And I think this is the luxury of profitability.

We have margin. We have space to experiment, to not set up complicated business analyses of whether this opportunity or that opportunity is the better. We can just go like, you know what? What do you feel like? Do you feel like this? Do you feel like that? Let's do the thing we feel like more. Because ultimately, no one knows anything. No one knows which...

new idea is going to be huge. If you look at the history of most great, huge products, they didn't start with an analysis that said, oh, my God, we've identified this enormous market and it seems like a short dunk hit. A lot of it started with like, this was kind of a funny idea. This was someone who was doing something on the side. This was an experiment. This was a lot of other things. It wasn't the product of some MBA market analysis.

That's maybe something they do at Procter & Gamble when they got a slice down the whatever baby diaper market. Oh, there's an opening here for something slightly more purple or pink or I don't know, with the seashells on it. That's not how most products that really move the needle come to life. And I think embracing that and going, no one knows anything. So we might as well just

work on the things that inspire us the most, that we are most interested in, and use the luxury we have accumulated by running a profitable company for 20 years that allow us to do just that. Because if it didn't, why else were we doing all this? Were we really working together for 20 years, sort of grinding it out to build a profitable business so we could just work on a bunch of stuff we don't care about?

What kind of reward is that for a lifetime of effort? It's not a reward that I'm interested in. I want to work on fun stuff. And as Jason said, if you line it up in such a way that the possible outcomes are everything from, do you know what? Only Jason or I like this idea. And we're just going to use it between the two of us to it's a medium sized success or other people use it, but not a big thing. Or it was a slam dunk. And you're happy with all of that.

You're guaranteed that you're going to be happy with business. Now, that may not last forever. Part of that privilege comes from the fact that Basecamp in particular has been this multi-decade, huge success that tons of people use. And that perhaps is over tomorrow. Like, who the hell knows where AI is going to go? Who the hell knows where the market is going to go? Who the hell knows what's going to happen about anything?

All the more reason to live in that beautiful moment where you can decide what you want to do and you can invest your time in the ideas that you're most passionate about. And this brings me back to the real sort of crux of why I thought, you know what, we should not hire another data analytics person. Because a lot of the data analysis that we were doing fell into the category of like, huh, that's kind of interesting. I'm not going to change what I'm going to do anyway.

And if the piece of analysis that you're looking at can't change the trajectory of where you want to go, it's a curiosity. It's not a necessity. It's not even something I think you should embrace. You should just go like, you know what, if I'm going to do the thing, regardless of what the little numbers say, I should just do the thing to not worry about the numbers, not even bother with the numbers. You know, one of the things I want to add about this is that

This is not exactly fully true, but there's enough truth in it is that a lot of data analytics is about the answer or the outcome is like, be careful. And, you know, whenever I tell my kids to be careful, I want to smack myself because it's like I say it too often and it's not a good thing to hear.

Be careful, be careful. It's like, people know to be careful. They're not going to jump off the roof. You don't need to tell them that. I just don't want to keep hearing like, be careful, which is a lot of what data can tell you. Like, well, don't screw with this and don't screw with that. And while you're 2% down, if you do this, like whatever. So if we're 2% down and we do it because we want to do it, like that's okay too. But it becomes hard to justify that.

that, well, that's okay too, when you're looking at numbers. When you're not looking at the numbers and you just do it and you find out what happens, it's like, it's more interesting in my opinion. So I don't like things that are all about be careful. I just don't like them.

Couldn't agree more. Numbers make you lose your nerve. They just do. We've looked at numbers before when we did the big pricing change. I remember looking at those numbers. Oh, man, if we get this like slightly off, we're going to be down so and so much. At some point, like those kind of things have a tendency to just inform. Let's just stick with what works. And that in itself is just a kind of death. That is a kind of death that happens.

founders usually have some allergy towards. They build something from nothing to something so they can retain that. But this sense of looking at the numbers too carefully, that's how MBA thinking creeps into your mind. That's how capital preservation keeps into your mind. And by the time you've ended up there, that it's all about just preserving something, not rocking the boat, you're already dead. You just don't know it.

So let's talk a little bit about the numbers that you do look at. I feel like we really try to keep our expenses in line. David, with the cloud exit, I know those were numbers that you were looking at very closely. So not necessarily how things are going in the analysis of product sales or those sort of things, but what numbers are you really looking at carefully?

Dave, you want to take that? I mean, costs are a big part of it as well. Profit margin is another part. I love looking at cost. And you know why? It's not even just about the money, although it's also about the money. Keeping more of it to yourself. Very nice. I highly recommend it. But to me, cost has this intrinsic aesthetic quality. Any cost to me that is sort of

accidental or squanderous feels like waste, feels like imperfections in the smoothness of the grain. I want to run my hand over the surface of the company and go like, someone really paid attention to whittle this down to just right. Like if I hold the bar up, oh, I can see it. It's not bent. It's not crooked. And

I just find the same attraction as I do from removing needless words in prose, from removing needless lines of code or concepts out of programs. It's just so satisfying to take things away that aren't contributing and making the whole thing simpler.

And I find that when we whittle down our expenses, we often end up simplifying the whole thing. You can go too far and obviously shouldn't do that. You can cut your expenses all the way down so you don't have any and then you won't have any employees and you won't have any servers and you'll be out of business in five seconds. So it's not about cutting everything down below the grain. It's got to be at just that right surface level. But I really like looking at those numbers. And

It's almost pathological to me where I sometimes have to stop myself. I do it with my own expenses, too. I just went through last night actually looking at my subscriptions I had on the old Apple device because that's what the kids still use. And it's just like the satisfaction I got from canceling something I know we don't quite use enough.

Or there's something else we could do. And I'm like, why the fuck am I caring about 10 bucks? Why am I getting such a joy out of going like, ooh, here's $10 that no longer have to weigh me down. There's like almost a Marie Kondo level of joy

Just the release, the lightness in spring that comes from taking a expense report that has 25 items onto it and reducing it to 17. And then the next time down to 14, because I think it's also connected to what we just talked about to the margin.

The more margin we create by decreasing our costs, the more freedom, flexibility and carefree nature can we enjoy when we're pursuing new ideas we can't quantify or can't even rationalize. I can just go like, you know what? I just want to waste time this week learning about this one thing.

Kamal, the tool we use to get out of the cloud, came off one of those wanderings where I just went like, oh, I don't know. I've been using Docker, this virtualization system, the last 10 years. I don't really understand it. I've never actually really looked into it.

I'm just going to do that, right? Instead of thinking constantly like, I've got to drive more business, I've got to drive more business. Because there are these phases where we shrink the cost down, a lot of margin, a lot of free space, a lot of time to roam. And then I can be wasteful in that open room. And that feels good to have that alternation between whittling the business down to a beautifully aesthetic, tight space.

Sometimes overly tight. I think Jason has taught me over the years that like you can be too lean. We've talked about this in past episodes where I get such an energy out of like reducing the expenses down, down, down, down. Oh, we don't need all these people we don't need. And at some point you're like one percent body mass and you're like a little bit of fat around the sides is actually makes for makes for longevity. Right. There's not a lot of 95 year old marathon runners around.

So just having some, some cushion is good, but do you know what? Yeah. Saving money is like saving words. It's an editing process. And there's a point where you can blow the sentence up too, by, by taking out too much.

And there's a point where like sometimes a nice, you know, flowery sentence with a lot of flourishes is wonderful too. So that's like a little bit of fat in your spending too. Sometimes you're going to spend more on something just because it feels right. Like I do think of it as an editing thing. And there is a place and a point where a paragraph and a word and a sentence just feels right. I think that's where we try to get our costs. And it feels like if we're spending too much on things that we don't need, it's like there's too many sentences in this paragraph.

You know, we don't want to whittle it down so much. So it's so lean that it's like you lose all the emotion in the writing.

Just as the same thing, you don't want to be so lean that your costs are just like so dialed in that you can't make any mistakes and that you're, you know, there's no body fat in the organization. So I think that's another way to think about it. It's like, and David, I know loves editing. I love to edit as well. It's a fun process to look back, take some things out or change some things around and look, this is a better collection of something now. And I think better question costs is a nice way to look at it.

I think one example that really comes to mind with cutting too much is we do these bi-yearly meetups at 37signals where we all get together in some city and enjoy our company or each other's company for a week and eat good food and so on.

And I approached that subject like it does every subject, like how much are you spending and what is this going to do? And then we had this one experience where we tried to save a little too much. Like my message of like not being wasteful had sunk in a little too deep. And we probably saved like whatever, 10 percent, 15 percent. But then the whole thing wasn't worth doing.

And you really have to be careful with that. You can whittle your cost away to such a degree that like, oh, I'm just spending the absolute minimum that. Yeah. But at that point, it's not even worth it. You shouldn't even be doing it. I'd rather spend zero dollars on a shitty meetup than like have saved 20 grand out of a budget of, I don't know, three, four hundred thousand. So that was a nice reminder that, you know what?

Think about cost, but then also think about like, is it worth doing at all if we scrape the bread so thin that there's no butter left? Like, we've got to have a little something. Yeah, since we're throwing more metaphors out, it's like you maybe don't want an A plus in cost cutting. Maybe you want like an A minus or B plus. Like you want to be smart about it. You want to be good, but don't be the best in class because you're just going to like take all the juice out of it.

All right, let's stop there. Yeah, with the analogies, we're going to wrap it up. Rework is a production of 37 Signals. You can find show notes and transcripts on our website at 37signals.com slash podcast. Full video episodes are on YouTube. And if you have a question for Jason or David, leave us a video. You can upload or record at 37signals.com slash podcast question.