Episode Transcript
Speaker 0 00:00 <inaudible> welcome to the rogue startups podcast. We're two startup founders are sharing lessons learned and pitfalls to avoid in their online businesses. And now here's Dave and Craig.
Speaker 1 00:24 Hey everyone, it's Craig. Just want to mention before we dive into the episode today that the MicroComp independent SAS survey is out. And if you haven't taken a look at this, this is something that Rob walling put together too to get a better sense of what non VC funded SAS companies look like here in 2019. Uh, I took the survey a couple days ago and takes about 10 minutes to really thorough and I think it's really cool that it gives us some benchmarks to which we can compare ourselves and our businesses against other kind of similar peers. Cause I think this is a kind of a rare business and world that we live in where you know, there's, there's a lot of data around venture backed companies and, and the growth rates they should expect and the performance that, that we should all kind of strive for. But for us who are not venture backed, um, you know, maybe you've taken some money but are not on the kind of go bigger go home, uh, path that, that, you know, we don't really know kind of what normal should be for us maybe.
Speaker 1 01:17 Um, and so Rob skull I think with this is to, to get a sense for for what is normal and what we should expect and what we should strive for and, and to know when we're kind of falling short of the Mark and things like that. So, um, I I highly encourage you if you run a SAS business with any kind of revenue, uh, to fill it out and participate because kind of this community sharing of knowledge and benchmarking will really help all of us who are going forward. So we'll link to the MicroComp independent SAS survey in the show notes, uh, head of her if you haven't already head over and take the survey and uh, I think Rob and the MicroComp crew will be sharing the results of this here with everyone pretty soon. Thanks Craig. How are you today? I am doing well. I'm doing well. I a I survived. We were talking about it last maybe the last time you and I talked, I was on the whole 30, kind of like paleo keto detox for, I didn't quite make it 30 days, but for about three weeks and just came off it and I'm having a beer and enjoying normal life once again. A nice, it was a no, it was a nice experience. Yeah. Good, good. How are things with you?
Speaker 2 02:27 Things are good. Uh, things. I was just out at my freelance client last week and so I'm returning to a bunch of things this week for the family. Um, and I will actually be heading out to the great American beer Fest tonight, uh, much to your dismay. That's awesome. I'm very much looking forward to it. So, you know, things are going well. We have, uh, some planned vacation with the family coming up in a week and a half and I'm looking forward to that. Hopefully it won't snow on of us while we're gone. So we'll see. You guys are going into the mountains around where you live or we're going to the Southern part of the state. So we're going through the great sand dunes, Durango, Mesa, Verday, four corners, that kind of area. I think kids have not seen any of that stuff. So, and actually my wife has not seen a lot of it either.
Speaker 2 03:22 So I'm the only one who's ever been. And you know, I've talked about these things for a long time. It seems like a perfect time to just go ahead and throw down and make it happen. Uh, it's a relatively close and easy trip to make and lots of fun stuff to do. The kids are all at a great age to be able to do everything and anything there. So it should be fun. It should be fun. That's awesome. That's awesome. I wish we did more camping. Our kids are seven and eight and definitely can do some things, but definitely like get bored and frustrated and can't really carry anything. So it's just like, all right, we can do car camping and that's about it for now. So I'm sure there are people out there listening that are like hardcore backpackers or campers that are saying, Oh, you're a wuss.
Speaker 2 04:05 Get get with the program. But yeah, we're, we're soft. I guess. It's all right. I wouldn't have done it either. We were strictly car camping until this last year and I just started taking them out tent camping the previous years. So yeah. Good. Nice. I feel better about myself. It's all good, Craig. It's all good. Are you still kicking ass? I hope, uh, yeah. I mean, we add, uh, the best month ever in recapture. Uh, the total numbers were like 30% growth. That's awesome, man. Yeah. Um, you know, we're three days into the next month and you know, not seeing anything yet, but it all happens like in bursts. You just never know. Like one day you hit expansion revenue, you're like, Holy shit, there's $300. I just didn't have <inaudible>. So yeah. I, uh, that's awesome. I definitely don't have that. I'm very dye. We have a little bit of like plan upgrades, but nothing like really built in.
Speaker 2 05:06 Sucks. Yeah. I mean, this was the smartest fucking thing I've ever done. Um, yeah, which, yeah, that's awesome man. I mean, whatever. Yeah. It's like, I'm sure you're kicking yourself a little bit, but it's better now than you know, never. D I was 29.2% growth last month. Um, and the previous month was 12.7 and then the month before that was 19.8. So I mean last month was really amazing. We also had some pretty heavy churn, but, um, the upgrades were really awesome. Like we only, we added 1100 new, 2200 were upgrades and 683 was churn. So it was also the highest churn month I've had. Yep. So that's pretty high turn, huh? Yeah. But you know, it's because there are some things in here, like I just think that profit ProfitWell could be tracking this better. Like sometimes I'll see people upgrade and then downgrade on the same day and they're still in the middle of their fricking trial.
Speaker 2 06:12 That to me is not churn. It shouldn't be counted as new revenue and it should not be counted as churn because they're still on their trial and they realize the revenue. Yeah. Yeah. It's like, is that inflating my churn numbers? It probably is. Cause I do get those, you know, did I get 683 of them? No. And you know, some of them were like, I lost some older accounts. Uh, I've got one guy that just, I can't get ahold of anymore. Another guy that was a zombie account and they finally, um, hit their expiration date on the credit card. Yeah. I mean, so some of it wasn't like the worst churn in the world. It was kind of expected. But uh, yeah, still struggling to try to get some fucking feedback from people when they leave. You know, I probably get one in 15 people emailing me back and I'm doing like a very aggressive follow up somewhere between six and nine emails. Well, pers, me personally on top of what it already does automatically. Wow. That's a lot. Yeah. Yeah. So that sucks.
Speaker 1 07:13 Just introduce self-service cancellation, like with the new, like a, with the no credit card trial stuff since we had like opt in to payment, we figured you should be able to opt out too. And their churn was up like 1%, um, from like four to 5% last month. And like that sucks. So I'm, I'm keeping a really close eye on it because if our churn starts going up, we'll, we'll yank that shit out.
Speaker 2 07:40 I don't know about that. I don't know that you should necessarily take that out. I'm a firm believer in if somebody can sign up manually without any, you know, if they do self-service sign up, they should be able to do self-service cancel. I think it's just a support thing. Like if you're, if you're signing up for a service and you had to call somebody to cancel or email the cancel, like that ratchets up my anger. Mm. Yeah. And, and I don't think that I'm alone in that. I think that pisses a lot of people off. Like they want to be able to say, Hey, I want to leave. That isn't to say you shouldn't, you know, do other stuff like, Hey, put up a dialogue that says, why are you leaving or follow up the emails and say, yeah, yeah, if you did that you like, you could actually capture more information.
Speaker 2 08:28 I'm screwed in a different way because when the uninstalled, the Shopify app that triggers web hooks that go and you know, basically pull them out of my system and register them as uninstalls, but now I don't, I can't put anything up in front of them. They've taken my app out. I don't get the opportunity to do that because the Shopify install hooks and uninstalls hooks cannot be stopped. Yeah, I'm sure that's, I can't interrupt him. Like that's just part of the platform. Yep. Yep. So yeah, I can't, I can't put something up and the ones that Shopify puts up in my place, they suck. And you know, I get people that have installed the app for literally six minutes and they're like, app is not performing well. I'm like, you didn't fucking send that email. How do you know it's not performing well? Yeah. There's no way you could have evaluated that. They're just randomly picking shit. So it's like the data is crap. It's utter crap.
Speaker 1 09:27 Yeah, yeah, yeah. That sucks. Yeah. So we don't have any kind of data collection stuff. We're gonna, we're going to put that in here soon. Um, but we have other shit, we have bigger fish to fry right now, so. Yeah. Yeah. So I mean, I guess like the early, early data for us with the no credit card trial stuff is really solid. We had growth last month about the same as the month before, which is nice. Um, so about half of last month's growth was on the no credit card trial, people converting, which is cool. Um, we had, you know, the, the beginning half of the month was people who had started trials with credit card up until, you know, the end of August. So I feel good about like, as a starting point, you know, um, and, and from now we have like optimizations in the app and then onboarding and stuff like that and in support, um, to, to help, you know, more people be successful.
Speaker 1 10:18 So that's really cool. The one thing that like I definitely can report on after a whole month is like our help scout ticket volume is exactly the same as it was the previous two months. So that's good. All the, you know, if you go no credit card trial, then your support will go through the roof. Like we've done a lot to make the app better, uh, in terms of like easy to use and in app messaging and on, you know, onboarding help and documentation right in the app and stuff like that. So I think that's mitigated a lot of it. Um, but it was nice to not see, not see that we have seen like different types of people using it, um, to maybe like more novice people are using it now, uh, than before. Both like novice to WordPress and people not using WordPress and just using cast despite self, uh, and really don't know anything about podcasting.
Speaker 1 11:10 So that's, that was not anticipated. And that's been interesting from a support perspective. Um, but, but it's cool and we're managing it fine. So I think that's the, the thing I know for sure. And I thing I don't know for sure is what our real conversion rate is. I guess I could calculate it from like people who signed up after the first and then converted to customers. But you know, ProfitWell just gives it for the month. But last months conversion stuff was kinda half the old style and half the new. So here in October we'll see what the conversion rate ends up being. I hope I'm shooting for 15%. That's the number I've told our team that we're shooting for. Um, if we can do that, that would be great. We were 40% before with credit card. So, um, yeah, just open the volume makes up for the difference.
Speaker 1 11:56 So, um, I guess the other big news, uh, for me is I am talking at microcuff Europe and all that. Yeah. What's your, what's your talk on? So my talk, uh, not to give away too much of like the secret secret sauce, but my, the headline of my talk is about SAS metrics. So SAS metrics you need to track, uh, and the ones you don't. And kind of how we track metrics both for me and communicate that with the team. So that'll be cool. We, I think we do some interesting and kind of unique things there. Um, but then the, the, so what of the talk really is going to be icy in and my app and that, that we see, I think a lot in the tiny seed cohort is what your, what your SAS metrics are telling you about where you should spend your resources.
Speaker 1 12:45 Um, and so like the interesting question of like for growth, you need more product or do you need more sales and marketing? And so your, your sass metrics really tell you that, right? And it's, you just kind of take a step back and not look at the numbers but look at like what the numbers are telling you. Um, it can tell you pretty quickly like you have a problem and people are leaving, uh, and that's like a product problem or that's all pretty good, but you just need more people. Um, and so that's, you know, a different kind of problem, but at least, you know, so that's kind of the, so one of the talk like the, the first half will be kind of about such metrics and how we report them and how we share them with the team and how I think that's really awesome and has helped kind of our comradery. Um, and then kinda the, so what for most of us, we have a marketing problem. Even people like me who are not developers, we want to just build and build a build and, uh, if you have some degree of product market fit, then building is not the answer. Even though it's like emotionally the easy thing to do cause you just go hide away or lean on your developers instead of kind of owning that initiative. Yeah. All right. So trick question for you is churn a developer problem marketing problem or support problem?
Speaker 1 14:02 I gotcha. I gotcha. I gotcha. Uh, it, it's, it's all probably right. Um, I would say it depends. It very much depends. Yeah. The reasons for why they're churning could be you're missing a key feature or the hate your onboarding, or they don't support them how to get started with their product, or you're just completely missing the ball on meeting their needs or hitting their pain. It could be any of those, right? Yeah. So when, when you see a high churn metric, then you have to figure out, you gotta dig deeper into that, right? Yeah. And like I'll, I'll tell you like what with churn, and it's a little bit outside of what we're talking about, but like at podcast motor, we see churn when we sell to the wrong people, right? So we sell to people that aren't really ready to start a podcast and they sign up for a, and they go for a month and they don't produce any episodes.
Speaker 1 14:54 And then, you know, it's a recurring billing thing. And then we bill them again and they say, what the fuck Craig was, I am not paying you another $700 you hadn't done anything yet. And we say, Hey, you know, when you signed up, it was, I hope it was very clear that this, the recurring billing subscription thing. Um, and I think that's a little harder to do with, with some of the like, you know, Casto sir or with recapture cause people are kind of, I know more searching for that solution. I don't know. But, um, that, that is definitely a problem. And I think we see that like higher up in the funnel a lot where like we have a lot of people on our email list that will never become customers because they never will start a podcast, um, for both businesses. And that's just like we are, we're giving out, you know, we're producing a lot of information, we're capturing emails, we're nurturing people and they're just never converting even to the trial process. So that's, you know, that's part of it.
Speaker 2 15:52 I'm kinda jonesing to reconnect, especially after we just sold a big snow tiny comp tickets for this year. So.
Speaker 1 15:59 Nice. Do you guys sell out again in like five minutes or what
Speaker 2 16:01 we sold out? Yeah. Yeah. And again, it's like a mastermind retreat now. So it was pretty much like, okay, is everybody coming? Yeah. All right. Yeah. So sold most of them that way. And then for those that weren't coming, it was a matter of going to the alumni list of people that were, you know, obvious fits from the past. That would be great additions to that mastermind that have wanted to come before. So, you know, sorry if you're listening and you wanted to come to big snow, tiny comp, I'm really sorry, w w would love to open it up to just, you know, a hundred people and have everybody show up and break it up. But we just, we can't support that logistic at this point, unfortunately. So. Yup. Yup.
Speaker 1 16:41 Uh, how many people do you have?
Speaker 2 16:44 12 including me. Okay. Okay. Nice. That's a good number. That's really cool. Yeah, it's worked really well. And this is actually our fifth year this year. Which time flies, huh? Can't believe it's been that long since we started it up. Yeah. And uh, Vail resorts was kind of messing with me, I guess this year. The one of the condos, we used to have two condos in the same building. One of the condos, which was like our, you know, I'll call it the workhorse of the sleeping bunch here because we had like six of us, seven of us in one condo <inaudible> five of us in the other condo. Yup. So the seven condo, they, the owners pulled it back this year. They said, no, we're going to be coming out to stay. And I'm like, damn, I mean for five years you haven't been doing this and now all of a sudden you hit it.
Speaker 2 17:32 We go in the middle of the week. It seems like the perfect time to go and the least amount of conflict. But they had conflict. And so now we're at the adjacent building for one of the condos. So it's, you know, it's not going to have all of the smooth logistic notice of before, but it's a similar setup and it's supposed to be a nicer condo. So we'll see. So I'm excited about that. Went and got, you know, ski stuff from my kids and now we're just waiting for the snowed of all on the mountains here so that we can actually go up.
Speaker 1 18:01 I heard, uh, my wife was telling me that Montana got like four feet of snow last week or something and we're recording like the 1st of October now.
Speaker 2 18:09 Yeah. Yeah. So somewhere up near glacier national park. I want to say it was like white fish, Montana. That's crazy. Yeah. They got 49 inches or something like that. That's crazy in one storm, which is just insane. Right? Yeah. Dangerous, right. Like that much snow in one time is Oh yeah. Oh yeah, yeah, yeah, yeah. That totally, I mean that shuts down roads and it stresses that infrastructure and um, yeah, and it's not gonna last long either cause it's still not cold enough. Like if this had been after a month of cold, they would stick around longer. But at this point it's still been pretty warm. So my guess is it's most of it's melted off. Yeah. So you know, great on them for getting that much snow. But unfortunately, yeah, that's, that's a lot to get in one bit. And then next week we're actually going on fall break with my kids and we're going camping in Southern Colorado. Speaking of fear of snow. Yeah. So
Speaker 1 19:01 pretty early though, right? I mean like we never get snow before Thanksgiving really?
Speaker 2 19:07 Anytime after the middle of September you're, you're on the hook for it pretty much. Although in the last five years, we've not really seen a lot of snow in September, if at all. And you know, it's usually been like late October before we get anything down here. Yeah. The, the weather report looks okay so far. It looks like it's going to be, you know, warmish for the next 10 days or so. You know, I realized that could change at any given time, but here we are on October 3rd and the, uh, the 10 day forecast is showing, uh, the Thursday before some snow and then after that it goes back up to 70. So nice. Okay. Yeah, I dunno. We'll see. Hopefully it's all gonna work out just fine. Yeah, yeah, yeah.
Speaker 1 19:53 So we were talking about, uh, like recapture growth a little bit earlier. Um, with, with the, the growth you're seeing, like is there anything that you're like feeling differently about or like planning or doing differently with the, the app and the business now that like things are going well or just like more of the same?
Speaker 2 20:11 No, it has actually changed a little bit. So it's interesting. Um, I'll tell this story cause it's kind of funny. I don't think I mentioned this the other night. I was actually out on a date with my wife. Uh, we hadn't gone out in a little while. So we went out for a nice dinner and we got an hour into dinner and we were just kind of finishing up our entrees and then all of a sudden recapture started alerting like crazy. Oh shit. And I'm like, huh, this looks bad. And then I couldn't get a hold of my developer Mike. It was pretty like, it was like six o'clock in the evening, which is like one or two in the morning, your time, something like that. And cause you're a, it's like seven, eight o'clock there, right, right now. Nine 20 here, nine 20 all right, so your eight hours ahead.
Speaker 2 20:58 Uh, anyway, so he's in the same time zone. You are so eight hours ahead. It would have been like two in the morning. Of course. You know, I'm texting them on WhatsApp and he's not picking up. So I'm like honey, I think we're going to have to stop date night. And she's like, Oh okay. And she was very understanding about it but I mean I was freaking out cause I'm like I have no idea what's going on. I got home, I like start digging into the admin stuff. I still don't know what's going on. And it's not like I don't know how to run the app or anything like that, but there was just something that was crazy that was saturating our entire web clusters. So we have like three or four servers in the front end and it was just saturating it and I'm like, I have no idea what the hell is going on.
Speaker 2 21:41 I'm looking at CloudFlare, I'm looking in the logs. It basically brings the whole thing to its knees after about 30 40 minutes of running. So I'm like, I'm restarting it and I'm watching it for 40 minutes and then it goes back down and I'm restarting it and I'm like, what the hell? So finally it, it ends up being like five o'clock in the morning by the time that I am getting to this point. And Mike wakes up and here's the messages. And he's like, Oh dude, what's going on? So we start digging into it and then it turns out that there was a store being attacked that was one of the stores on the capture and they had like an, I'm not joking on this number, $273 billion in abandoned cart in 24 hours. That was the number that showed up in the reporting the next day. So they were getting DDOSed and then that trickled down to you?
Speaker 2 22:33 Yeah, yeah, totally ended up like spamming us from the back end. And we were like, Holy crap. And this turned out to be a customer that wasn't a paying customer. If it had been a paying customers, that would have really sucked because here was the way we had to solve this. We had to basically kill their account. We had to delete it. Yeah. So we deleted their account, we unhooked him from Shopify from our side, and then that took the pain off recapture. But it highlighted some weaknesses in our infrastructure that I definitely wanted to address after that. So all of a sudden there has been high, high priority for me to make sure that our infrastructure is super solid. Um, especially coming up to black Friday, cyber Monday. So now that we've seen all this growth, it's no longer, you know, we've, I've been monitoring the platform and we've been adding stuff as needed, changing out some servers and things like that.
Speaker 2 23:23 But now I feel like there's some other guards that have to be put in place to make sure that we don't end up taking down the service for everybody else because of one bad actor in the system. So we've had to brainstorm, containerize those people somehow or so. I know that's not the right term, but yeah, no, I mean it's the right idea. So there was definitely a lot of thinking of what were the weaknesses and then, you know, how can we take better advantage of the processes on the machine and get more out of them? Cause we were, we were struggling, it was looking at the CPU and it was showing like 25% on a four core or four CPU machine. And I'm like, why is it only 25%? Well it turns out that there's some clustering things that we didn't have enabled so it couldn't handle the increased load.
Speaker 2 24:10 So it was like, Oh, there's something that we didn't think about. And then, you know, we added that to the list. And there's other things about rate limits for emails when something goes sideways and having, um, certain points of single points of failure in the system got highlighted that were like, Oh, I didn't realize that was a single point of failure. Yeah. We gotta we got to double down on that. So we have at least two of those standing by. So stuff like that. So there's been several infrastructure things that have been added in there. And then we've also been trying to push out one feature here before, um, black Friday, cyber Monday. And that is unique coupon codes cause that's a big one that everybody really wants. They don't want to just offer a blanket discount. They want to say this person gets this code period.
Speaker 2 24:52 That's it. And then they can be able to track it to say this person used the code, didn't use the code. Um, and we'll be able to do all of that now here come next week, which is great. It's probably been our most requested feature at this point and I've actually lost some customers over it. So it's definitely something I knew I had to get in there. Um, so yeah, I would say that those things have really changed my thinking. The other thing that the growth over the last few months has really highlighted is I've been watching my cohorts, my 60 30, 60, 90 day cohorts and I'm starting to see between 60 and 90 days. It looks like there's some drop-off more than there was before. And I'm still trying to figure out why. Um, you know, as I was alluding to earlier when you and I first started talking, trying to get people to follow up with me is a real challenge. I send them, I send them a fuck ton of email and you know, I, I've even tried like F email from multiple addresses and multiple sources. So I've got stuff from Intercom, I've got stuff from my recapture IO address, I've got stuff from my consulting address and I send a mixture of all of that and I'm, you know, probably going on the order of about nine emails at this point that follow up with them and I'm still getting maybe one in 15 to say something.
Speaker 1 26:16 Yeah, I mean I think the, the easy solution there definitely comes with its baggage is like offering a $50 Amazon gift card for people to hop on a 15 minute call with you. But, but, but then that I think very much can kind of biased the input that they give you. So
Speaker 2 26:32 it might, but you know, I'm, I'm at a point where I'm desperate enough, I might just start doing it. Yeah. On the other hand, I have seen the ones that I've gotten through and the ones that have actually said something to me. It's interesting they kind of, so far it seems like they've fallen into a couple of buckets. One bucket is, you know, we're testing you with other products and so we did a trial with you. Now we're doing a trial with somebody else. Then we're doing a trial with a third one. So you're like six weeks away from an answer after that, assuming that you're the first one that, that they look at. Yeah, and I have tried following up with them multiple times and they don't necessarily get back with me, which usually means they pick somebody else. I would love to know why they picked somebody else.
Speaker 2 27:18 It could just be that they wanted the sophistication of one of my competitors and that's not something I'm at feature parity with to even think about offering right now. So it wouldn't really matter to try to compete on that. But some of them, you know, like I just want a competitor a thing the other day where they were like looking at me and this other service and they were kind of hemming and hauling because the other service had a couple of more features that they liked. But they liked a couple of things that we had that the other one didn't. And they really liked our customer support and that's what sealed it for them. They were like, Oh your support is way better than them. And so they're like, I'll sign up now. And I'm like, yes, that was a great feeling and that, you know, I was on a live chat with him so I knew that like right there, Intercom just paid for itself that month.
Speaker 2 28:02 Yeah. Cause I got that feedback and I can't always jump on and catch that with people. You know, sometimes I'm like, you know, there's an hour delay between when they send it. Like if they sent it at five in the morning and I get up at six and I see it and I'm like, well there's no way I was going to get that. And I'm not going to ruin my sleep schedule to try to be on chat all the time. And I can't afford to hire somebody to be there all the time to just sit there watching for something that happens very rarely. So I just, I try the best I can with that. But yeah, I mean, I would love better feedback at this point, but, Oh, so the other bucket that I, so they, they are doing the multiple trial thing or they're switching to a competitor.
Speaker 2 28:40 Like I lost a one person that does this clothing store in Denmark and she was like, you know, I really like your service and we've never had a problem with it or whatever. I'm just moving up to this other thing that has more stuff that you guys don't know. I'm like, cool. Yeah, yeah. Okay. I mean, and I asked her, you know, I said, Hey, since you, you know, liked our service or whatever, would you mind giving us a, a review in Shopify app store? And she did get five stars. It was great. Yeah. Um, so that was, you know, at least it didn't turn into a, you know, I hated you. I hated your service. I hated everything about it. You know, that's, that's what you always feel like, or at least I do. I take everything super personally. So it would always feels to me like, Oh, they're quitting the service. They must hate me and want to kick my dog and they think I'm ugly and straight. Like it just, it's a totally unreasonable, but that's just how it feels sometimes. Yeah. Yeah. Um, yeah, I mean the infrastructure stuff kind of going back to that is, is really interesting. Like we, we have, um,
Speaker 1 29:42 we had kind of scaling issues a while back, got a lot of work done there. Um, and that is a really, really, really good investment. I think. Like, I sleep well knowing that like, you know, we have auto scaling and load balancing and all this shit in place that like costs tens of thousands of dollars of, you know, SIS, admin, dev ops work. But I, I mean, yeah, I think we have a really, really, really good setup. We do blue, green, like no downtime, deployments, all this going to stuff like, yeah, I mean it's, uh, I, I think once you get to, again, like some degree of product market fit and you have customers depending on you, you have to invest in that stuff. Um, because like the alternative is really bad, like reputation damaging, uh, customers leaving, leaving bad reviews like you're talking about. Um, let's see. I think you're really smart to, to invest into that stuff right now cause that's, you might not see the positive impact, but you're mitigating the negative potential.
Speaker 2 30:46 Yeah, and I would throw out a caveat, like if you don't, the guys that built recapture, I should say, the guy that built recapture, he, you know, he had an idea of what it was that he wanted to build and he had seen previous services and he obviously had some software engineering experience. And when you put it all together and built this thing with scalability in mind of having separate workers in web things and um, one for the beacon to be able to do the email tracking and stuff like that. It was, it was very, very smart for the size of the thing that he built and the number of customers. When I first acquired it, it was absolutely overkill. But you know, it's one of those things that a, he really knew if this thing took off, he would have to go that way anyway.
Speaker 2 31:32 And B, it seemed like he just naturally gravitated to putting that together that way, which is great. And then I don't think it was a huge amount of overhead for him to get there. But if you're like still going at this and you're not really sure, is this going to work? Is this gonna resonate? Do I have enough customers to make it? You know, I still think that doing that scalability piece is going to be the wrong thing at the wrong time for, I mean going into a abandoned cart emails, this was an established market and he knew there was a, an established need for this. And if he could get the traction he wanted, then it was probably going to take off. So, I mean, it was a gamble. I'll tell ya. It could've, it could've gone the other way with him and he could have had no customers and he could've, you know, shut the whole thing down and never sold it or whatever. And it might not have paid off at all. I'm glad that he did because now I don't have to turn around and invest in that. And I probably would have, I absolutely would have by by now I would be scrambling to scale and it would suck. But right now instead, all I'm doing is just adding servers when things go sideways. It's like, Oh, we need more capacity. All right, let's upgrade one of our servers. And instead of having a two core, let's do a four core or something like that, you know, um,
Speaker 1 32:47 you guys don't do like a load balancing and auto scaling and stuff like that on AWS.
Speaker 2 32:52 I don't use a auto scaling. We do have ELB. So the elastic load balancer is what we put in front of our web stuff. And then that distributes among four different nodes. And then on the back end we have multiple workers. And those workers, um, are basically draining from a common queue. So the web thing is sitting there filling up the queue so we can have all these nodes that are filling the crap out of the queue and then these workers individually go and start draining the queue and doing the jobs, processing information, send emails, whatever. Yeah. Yeah. So we don't, I mean there's no need to load balance that they just do the work and if they can't handle the work, we add another one in there to do more work. Or we had a bigger one in there to do the work. Like I think last year for black Friday, cyber Monday we upgraded them from T one smalls to T a or T two smalls to T two mediums to do more work and that's all I did.
Speaker 2 33:51 And then they handled that fine and they'd been doing that fine most of the year. I think at one point we added a third worker and then it turned out when everything went sideways during date night, one of the workers, the reason that went sideways is partly because there was a, they attack. And then at the same time we had one of our workers completely fail, like the hardware failed. That's a mother fucker, huh? Yeah. And wow. Geez. Yeah. And I was glad that I had two other ones in there, so that failure didn't kill us. But it also highlighted that we had a single point of failure because it was the one that was also generating the other emails off the queue. And then it was like, Oh, wait a minute. That was bad. So that did cause some havoc. We recovered from a very quickly by switching to one of the other nodes, but it was definitely something, um, that, you know, put infrastructure into our focus there to say, Hey, you know, before November 1st, here's some things we got to get done. Yeah. Yeah.
Speaker 1 34:53 That's cool man. I, um, we have a really, really, really good SIS admin consultant that we work at without a California. If you want an, uh, an intro or anybody listening wants an intro, let me know that they're, they're not cheap, but they're really, really, really good. So if you guys are wanting to do this kind of CIS admin automation, dev ops stuff, uh, I, it's really worth it. The investment, I think if you're at any kind of scale,
Speaker 2 35:18 these, uh, this is for AWS. Yup. Yeah. Okay. Yeah. Yep. Yeah. I mean at some point I think I'm going to have to hire a professional dev ops company or team or you know, dedicated person to do the support on the other side because, you know, Mike and I, we have some pretty good coverage between the two of us, but there are definitely times when like, you know, there was, it was late evening here and everything was going sideways and Mike wasn't awake yet. And you know, waking him up is hard to do and it's also unfair for me to have to keep doing that all the time. Yep. So, you know, if recapture gets to a point where we're making more money, that's definitely something I'm going to have to consider as some 24, seven monitoring. That's going to be a little more live than what we have at the moment.
Speaker 1 36:03 Yeah, yeah, yeah. Do you have something like status cake or status page or something like that they installed to, to monitor the servers and uptime monitoring.
Speaker 2 36:13 We do a, we actually use status cake on all the nodes on, I'm not super happy with it and you know, it, it tells us some things and it's okay. But I haven't really liked the, uh, the alerting ability. Like it's not quite, it's not quite at the granularity I want and it's not, I want to be able to do a level of sophistication that I can't quite get to at that service, but the services that I found that do do that are like $400 a month. And I'm like, yeah, yeah, that's not okay now I can't really go there yet. Um, but you know, status cake was enough. Like it was the one that was telling me, Hey, you know, shit's blown up over here down. No. Yeah, that's the worst text ever to get down. Well, the worst one was it said everything was down. It said your, your worker's down, your web is down, the health check is down. Like, like, okay, that, that, that was the one that sealed the deal. I'm like, yeah, we gotta go, so I need another drink. Are we going to go well, I don't care.
Speaker 1 37:18 Oh, no. Yeah, we uh, the, the <inaudible> yeah. It turns out a lot of the scaling issues we ran into were around analytics computing. This was a year ago now. Um, and that was really interesting to, I mean, we're crunching huge amounts of data for some of our customers and that was the, the bottleneck is some of our customers are just like fucking refreshing the analytics page every five minutes. And it would go and pull this from the database and do all this stuff in the app and then try to display it on the front end. And the app would just be like that. No, too much. So yeah, we got around it finally, but that was, that was a learning lesson. I mean, we were, we were kind of containerized now and that's on its own server and all this kind of stuff and we're all happy because we learned. But yeah, it's cool to be able to learn these lessons, you know, without, without too many scars,
Speaker 2 38:11 not too many. Yeah. Yeah.
Speaker 1 38:14 I think that about wraps us up for this week. If you have any questions or comments for Dave or I shoot us a message podcast@roguestartups.com and as always, the the ask is if you're enjoying the podcast or what you heard today to share it with someone who you think would enjoy it as well. Thanks so much and we'll see you next time.
Speaker 2 38:34 Thanks for listening to episode of
Speaker 0 38:36 rogue startups. If you haven't already, head over to iTunes and leave a rating and review for the show for show notes from each episode and a few extra resources to help you along your journey. Head over to rogue startups.com to learn more <inaudible>.
