Episode 775 | A.I. Coding Tools, User Experience, Racking Your Own Servers, and More Listener Questions (with Derrick Reimer)

Rob Walling:
This podcast is brought to you by Mercury. I’ve been banking with Mercury for years and whenever I set up a new account, I’m reminded why traditional banking feels stuck in the past when our previous bank faced solvency issues, we needed to spin up new accounts quickly that could handle millions in funds across multiple businesses. Mercury had us up and running almost immediately. I manage half a dozen different Mercury accounts across a wide range of companies from my personal single member, LLC to MicroComp, our seven figure global events and education platform to TinySeed our venture fund and accelerator. Mercury easily handles them all. The interface is elegantly simple for daily banking, paying invoices, and sending and receiving international wires, yet powerful enough to handle the multi-step approval processes we needed to put in place. When funding founders with large transfers, anytime founders ask me who they should set up their accounts with, I send ’em to mercury.com. 

Check the show notes for more details. And note that Mercury is a financial technology company, not a bank. It’s another episode of Startups For the Rest Of Us, I’m your host, Rob Walling, and in this episode, Derek Reimer and I sit down and we answer listener questions. We answer a question about the impact of new AI coding tools, talk about user experience and how much time you should or shouldn’t invest in the early days. Talk about racking your own servers versus cloud hosting the trade-offs between those two things, email deliverability and reliability risks around that. And we handle another topic or two if you can possibly believe it. We packed it in to this amazing episode, and if you stick around until the hidden track at the end, you’ll get to hear Derek Reimer get ambushed by my trivia questions. Before we dive into the episode, I want to let you know about some exciting updates to the SaaS launchpad course. 

It’s my nine and a half hour video course that walks you through the process of going from no idea to your first paying customer. We’ve just added a new module that dives deep into what founders should think about before adding AI to their SaaS. It’s actually a private conversation I had with Arvid call. He’s been integrating AI into his own SaaS POD scan fm. We talked through the landmines and tricky spots that founders are hitting right now so you can avoid some costly mistakes. And here’s the thing I’m really excited about. If you buy and you complete the course material in the next 30 days, you’ll be entered for a chance to win a 30 minute one-on-one chat with me. We can talk through your SaaS idea, any roadblocks you’re running into or just brainstorm next steps together. If you want to test drive the content first, grab the free sample module at SaaS launchpad.co. It’s a 28 minute video all about the DNA of a great SaaS idea. As a bonus, our team is giving away a free copy of the full course to one person who watches that video by June 1st. And for loyal podcast listeners, use promo code launch at checkout to get $150 off this course. Get all the details and access the free sample at SaaS dot co. And with that, let’s dive into listener questions. 

Derek Rimer, welcome back to Startups For the Rest Of Us. Always a pleasure. It’s great to have you back. So we are going to dive into listener questions today. I have hand picked several that I think are well-designed for you to answer. Before we do that though, you’ve been working on some pretty awesome stuff with Savvy Cal, you want to tell folks, give ’em a hint, give ’em a teaser, and maybe a little call to action if they want to reach out to you.  

Derrick Reimer:
Yeah, I think this actually might be the first time I’m publicly talking about this, but yeah, we’ve been working on some interesting new product lines around appointment scheduling. So Savvy Cal has historically been sort of a Calendly competitor. That’s mostly for meeting scheduling. You need to take a sales call or just book something with a colleague or something like that. So meeting scheduling has its own dynamics and appointment scheduling. It’s something that has come up basically the entire time we’ve been in business people inquiring about using Savvy Cal for scheduling appointments for more service-based businesses. And we’re just now starting to work on kind of a second product line that repurposes a lot of the goodness that we’ve developed over time with Savvy Cal meetings, but for the appointment scheduling use case. So right now we’re kind of partnering with one agency who’s using this for a cool new project, but I’d love to have more conversations with folks who are sort of in the space of building sort of custom flows for service-based businesses, whether it’s telemedicine or whatever, and need kind of lower level scheduling infrastructure. Very interested in having those types of conversations.  

Rob Walling:
Awesome. So how can folks reach you?  

Derrick Reimer:
Yeah, you can hit me up at derek@savvycal.com over email and I’d love to hear from you.  

Rob Walling:
And that’s D-E-R-R-I-C k@savvycal.com. All right, let’s dive into listener questions. Our first one is from Zach, and his question is about the impact of new coding tools. He says, Hey Rob, I just finished your book and I’m feeling inspired to get started on my next startup Rocketship emoji, which I love. One thing I was curious on is your take on the new wave of AI coding platforms, cursor lovable, et cetera, and the impact of these specifically on the micro startup. Does your 300 hour rule still hold up if you can build an MVP in just a few hours? And so I believe he’s referring to and start small, stay small. I used to think, I thought I said four to 600 hours, but whatever. I wrote that book 15 years ago, so maybe I said 300. It’s approximately directionally correct, right, of like, Hey, I have the 2 2200 framework even these days, which is two hours to if you’re trying to validate an idea to do some research and to go to Google and to go to look at keyword volume and other HFS and other things. 

The 20 hours is, hey, maybe I’m going to put up a landing page, have customer conversations, and then if I make it past both of those points with an idea, maybe I do spend the 200 or 300 or whatever that number is to build that MVP and get it in the hands of folks. But AI coding tools do in fact change that timeline. I think it’s significantly shortened. Now, I don’t know if a few hours feels a little, I don’t know, it feels a little exaggerated. I dunno how maintainable that code will be and beyond that, but let’s kick it off. Are you using AI coding tools? Have you found them to live up to the promise that people are talking about of, Hey, I’m a non-technical and I can build everything in three minutes? What’s your experience and thoughts here?  

Derrick Reimer:
Yeah, so I have become a pretty deep user of AI coding assistance tools. I kind of came late to the game. There’s been GitHub copilot, which was launched, I don’t know, two or three years ago I think. And I just never really caught onto that mode back when it initially came out, I think because the models weren’t as good. And so the kind of autocorrect as you type was just a bunch of bad suggestions. And these days it has gotten considerably better, especially if you’re in certain languages that the LLMs really know well. Like JavaScript, Python, it knows Elixir decently well, and that’s what I mainly use. TypeScript. I regularly now use Windsurf, which is similar to Cursor. These are kind of the big two AI assisted code editors. And then I think you mentioned also V zero and lovable, and I think those are more specific products from companies that allow you to kind of type in, here’s the app I want you to build for me, and it just spikes out an MVP based on what you tell it. 

And yeah, I think it’s really interesting how these tools do considerably shortcut a V one. Now the question is how long are you expecting to use this V one is this purely for validating something? Think of it like a working mockup wireframe thing that actually persists things to a database and you can show flows off. It can get you to that place really quickly. I am skeptical about the maintainability of some of these code bases that were generated from scratch fully by ai. I haven’t found the code that it generates to be good enough for, I’m trying to imagine someone who doesn’t know code at all using this to try to maintain an application for the long haul. And I don’t think we’re there quite yet. We might get there. I mean, this might be a horribly outdated take even in six months, not sure, but at the present moment, it makes a lot of mistakes that if you’re not technical, you may not catch it and that would not serve you well in the long run. But for sure, for the purpose of proving something out and being able to demonstrate something, I think it has radically changed the game.  

Rob Walling:
I would agree. And is an, it depends to all of this, and the it depends is are you a developer and are you a senior developer or junior or not all? And I do think that AI coding tools, I think of it as almost like a Mex suit, right? You think of Alien or did they have many aliens do anyways, A Mex suit makes you super powerful if you know how to use it. But imagine getting in it for the first time and letting it run on autopilot question mark. It’s like, right. And the analogy breaks down because you don’t have to maintain a ME suit five years from now and have infrastructure and spaghetti code or weird bugs that nobody can really find because there’s been machine generated code for 40, 50 years, I mean for a very long time. And back in the day, I actually was a contractor at L-A-D-W-P, Los Angeles Department of Water and Power, and I was a developer working on T net stuff there, and they had a system, I think it was in cobol, and they had generated code in the seventies to do a bunch of stuff, and it was machine generated and no one wanted to, I didn’t know cobol, but they had COBOL contractors who are like, everyone’s scared to touch it. 

It is a complete disaster and people are scared. It was like a payroll system or maybe it ran the grid, the electrical grid in LA for all we know, 

And that’s just the state of affairs sometimes with government systems. But I remember being like, wow, they generated code and no one can touch it. That sucks. What’s the point then? I mean, it works until.dot. So I’m not saying AI code is that bad, but I think that’s the big concern is you and I were having a conversation a couple of weeks ago when you came over for DD and I believe he said, you given instructions and then you go through that code to really make sure that it is not screwing around and it makes mistakes, right?  

Derrick Reimer:
Yeah, I would say I’ve still found this to be true even in my few months of using it aggressively. Now that it is best at writing test code and test code is just, it’s usually kind of an afterthought when you’re trying to move quickly and especially earlier stage. You want to make sure stuff is kind of baseline sanity checked, and you have some support for preventing regressions and stuff, but we don’t love spending a ton of time writing tests if you’re just trying to deliver value quickly. And I have found it is very good at analyzing your code, figuring out all the different permutations that should be tested, and then writing that test code for you. And ideally it writes it in such a way where you can pretty quickly, easily read through it and see, all right, this is the setup, these are the assertions that looks correct, and then it’ll run it. And if it fails, then it’s pretty good at least helping you start to think about how to solve the bug. But it’s not always perfect at that. But even just having those test cases where it’s actually rigorous and does every single kind of important permutation, that has sped up my process a lot. It made me feel like I have a stronger foundation of tests  

Rob Walling:
Faster and better, more thorough than you would care to do yourself. 

And I think really getting back to Zach’s question, because we’ve talked about AI and how it generates code and the dangers of that, but really he’s asking, so what if it makes me twice as fast, 3, 4, 5 times as fast? Can I get it done in a day? What used to take me a week or two days? It used to take me a week and in the future, will it be even faster? How does that change all this methodology? It’s really interesting to think about because one of the reasons that I talk about that 2 2200 framework, the reason that framework exists is because I don’t want you to spend 200 or 300 or 400 hours building something that no one wants. So that’s why there are those steps before it. But he’s asking question I think, what if it doesn’t take me 200 hours? What if it takes me 40 hours to do what it used to take 200 hours? 

Should I just build more stuff more quickly without doing all the validation? Here’s the thing for me, as I said, the two is doing research online looking competitors and SEO traffic and interest in demand and blah, blah, blah. To me, I still want that. If I’m starting, I’m going to spend a few hours plugging away. If I have a list of 10 ideas that I’m thinking about, could I feasibly go and build all 10 of those ideas in a weekend? Probably not now, but let’s say you could and I could just, I don’t want to do that without knowing there’s some demand. I don’t want a product if there’s no search traffic or if there’s no, I can’t find anybody who wants this. It’s not solving a problem. So I still want to do the two part. And then the 20 part is either it’s a landing page or customer interviews conversations or both, frankly, which is usually what I recommend. 

I still want to do that too. Now maybe could I shortcut? Can I build mockups really fast with AI or can I build kind of a clunky click through paper prototype thing with AI that’s not maintainable and build an MVP in order to do the 20 maybe? I think the danger, I mean it could cut both ways. If sometimes the danger is you’re getting ahead of yourself on giving too much specificity to someone of here are 17 screens with text boxes and buttons to click through, and it’s like, is that what they need to know if it’s worth it? Or are you basically saying, Hey, I’m going to solve this problem. I’m going to build a system that manages your entire business and allows people to log in and log out and click and audit and this and that, and is that what they need to hear? Especially if they’re non-technical. And sometimes I think they maybe do need to see a few screens of like, ah, I get it. But other times I think presenting them with completed software could actually be a detriment.  

Derrick Reimer:
Yeah, I think that that’s a really good point that the process of building a product for a particular market segment entails that you are actually speaking to that customer set and figuring out what needs to go into that product because it’s hubris to assume that you have all those answers from the get-go. So building something based on a bunch of assumptions that you haven’t actually confirmed by working with customers means you’re probably going to miss the mark if you come with this fully built out product. I mean, we talk about this kind of new appointment scheduling stuff that I’m working on right now, and a big part of this is getting real-time feedback from the actual end user, the clinic who’s going to be using this product on a day-to-day basis and incorporating that in as we build stuff. I can just make a bunch of guesses and try to envision what they need. And that’s certainly a big part of it is just trying to get into their shoes and think about what they would need, but that’ll only get you so far. You need to actually talk to the real customer to figure out how do you really nail it for all of their workflows and use cases and that kind of stuff. AI can certainly help you workshop some of that, but I think ultimately you’re still selling to humans and you need to accommodate human needs. So getting the actual human data is pretty key there  

Rob Walling:
Still selling to humans for now, Derek, for now, until AI starts, this  

Derrick Reimer:
Also might be an outdated  

Rob Walling:
Take two years from now. People are like, guys, AI buys everything now. So anyways, yeah, I appreciate the question, Zach, and I hope Derek and my takes were helpful. Next question is an audio question from Arthur Ky.  

Arthur :
Hi, Rob Arthur here, an aspiring entrepreneur living in Denver, Colorado. I have an idea for an app that I’ve recently started developing after validating the concept with potential customers, it seems like a strong business opportunity to move forward on. As I work on building this, I’m realizing there are various ways to approach the app structure, pages and user interactions. My question for you is how do you approach the user experience when bringing an idea to life? A traditional UX designer might create user profiles, wire frames, prototypes, conduct user testing to uncover pain points and follow other structured steps. I have a college background in graphic design and now work as a software engineer. So part of me wants to go through this formal process, but another part of me feels I should focus on quickly building an MVP to prove the concept and refine the user experience later. When in the process do you believe it’s most essential to focus on user experience? How do you typically approach UX for a new idea? I think a great user experience can make or break an idea, so I’m very curious about your perspective on this. Appreciate what you do. Thanks so much.  

Rob Walling:
Alright, Derek, as one of the best UX folks that I know that exists on the internet today, what is your take on Arthur’s question?  

Derrick Reimer:
So obviously I come from a place of being biased towards wanting to solve problems from a UX first perspective, it’s something that I really value in the products that I buy and in the products that I build, consider myself a crafts person and this is what I really care about. However, my answer to this trying to be objective is, I think it really depends. Are you staking a big portion of your value proposition on better ux? When you compare yourself to the rest of the landscape, what does your buyer actually care about the most? I tried to think about some examples in my own stack. So I use linear for project management and they’ve been around for, I dunno, three, four or five years, something like that. And I think their main differentiator, I guess is that they’re trying to be Jira but with better UX experiences that people actually like to use. 

But it’s kind of solving the same underlying problem that Jira is, which and Jira kind of has a bad reputation for being a little bit of a nightmare to use. And so for them, in order to deliver on this promise, they have to be executing top-notch user experiences. Otherwise people will just use Jira because if it’s no better, no different than Jira on the UX front, then why bother savvy Cal? We’re promising to be a more delightful user experience for the scheduler and the person who’s configuring the links. So you should be able to go in and fine tune your availability faster, more efficiently than you can in the other tools. So I think for us, the customers we attract are the people who are looking for those better user experiences. So I would say in these cases it matters a lot for other products. I’m just trying to think of examples, like software for  

Rob Walling:
Construction firms,  

Derrick Reimer:
Yeah, where you, you’re logging in, you’re doing stuff, there’s screens with forms and you’re viewing reports and charts and things, but the level of UX attention that a lot of the products that a lot of us use, linear is just probably not as important. What you’re doing is you’re driving a different kind of value. I think of even more stark examples like hit tail back in the day product that we worked on was SEO keyword tool. So you could log in, you could get reports on these are the keywords you should be targeting in your content. And the setting screen in that app was not very important that it was top notch ux, like it’s form fields and buttons, and you have the basic essentials. It had to be obviously navigatable, but the core value prop there was the keywords it was giving you. So let’s say in an app like that, the quote user experience of the product mattered a whole lot less. The user experience was really like is it delivering the right keywords? And so I guess that is essentially a different kind of user experience than what we classically think of as the way that you lay out menus and form fields and things like that.  

Rob Walling:
I don’t know that I have much to add to that. That was pretty much my take was it depends, and it depends specifically on is this your advantage or one of your advantages or not? And there’s certain spaces that just don’t know the difference. And again, I think of construction firms or maybe someone, an owner of a gymnasium of a fitness gym, a fitness studio, are they going to know the difference between linear and Jira? They could tell the difference, but are they going to be like, Ooh, this is good ux. They don’t really even know what that is, right? So I think the big thing is you end user, do they care? And not only do they care, but do you want to invest the time to make this one of your advantages? I personally would, I would tend to enter spaces just like you where the users do care, but probably most software doesn’t matter. If I think of most B2B SaaS software, I shouldn’t say doesn’t matter at all. You can have catastrophic us and then UX and then everyone hates it, but it doesn’t matter nearly as much as I think  

Derrick Reimer:
I have often opined why is the most successful software out there, the crappiest software, how does this actually happen? But I think there are plenty of examples where it’s kind of that adage of it’s always safe to buy IBM kind of thing. Like Salesforce is known for being quite painful to use, but it’s Salesforce and they have staked this incredible position in the market of this is one of the two tools. Maybe HubSpot and Salesforce are the only big two that any serious business of a certain size will be on one of these platforms and the people buying them don’t care about the user experience. There’s a whole bunch of other things that cause them to make that buying decision. And so I think it would also, this is a trap that a lot of times people fall into where you look at a successful incumbent and you say Their UX is terrible, I can do better ux. 

And you ignore the fact that the actual person buying it, maybe the actual person buying it doesn’t even use the product, but they’re charged with buying it because a VP at some level and they’re the one who has to make that procurement decision and they buy it because their peers are buying it or because companies of our level of importance buy this software. And so there’s just different motivations, but I think that often causes kind of a mismatch and the entrepreneur who thinks I can build a better version of this thing when in reality the market doesn’t care.  

Rob Walling:
And I think that’s a good distinction you just made, which is the further your buyer is from the user of said software. I think the worse your UX can be because the buyer doesn’t care, the buyer’s usually going to buy based on market and brand and reputation rather than easy to use. So thanks for that question. Hope it was helpful. I wanted to jump in here for a second and invite you to Microcom Remote, which is happening live tomorrow, May 21st from 10:00 AM to 1:00 PM Eastern Time. The event consists of three presentations, talking about early stage SaaS sales. In addition, we have a founder by founder, which is like an online version of the hallway track at our in-person events talks will be recorded in case you’re listening to this event after to get access, head to MicroConf dot com slash remote. 

Next one comes from Louis Mertons and Louis asks a question about on-prem versus cloud hosting. He says, Hey Rob, I really enjoyed the book and I’ve listened to it twice now on Audible. I think SaaS Playbook, if I’m guessing. I also love Linus Tech Tips, and they recently built a server and said that these days it’s cheaper to run on-prem. And many people were moving back to on-prem rather than the cloud. I love proclamations that many people. It’s like, all right, yeah, there’s five of them. I wonder if you could talk through the pros and cons of cloud versus on-prem. I suppose it would avoid vendor. What do you think about this, Derek? I mean he’s basically saying getting a physical server in a cage somewhere and getting it in a, we used to do this 15, 20 years ago. We also used to charge one time for our software. Should we do that as well? No, I’m just kidding. You can tell my opinion on this, but what are your thoughts? Because you as an operator could totally, you could probably save monthly hosting cost if you spun up a physical database, go buy a Dell, buy a database server and then go buy your, I say this, I remember doing this 25 years ago with clients with big e-commerce clients and going and racking the servers in a place.  

Derrick Reimer:
I mean, I have seen a lot of talk off and on in our space about this, about exiting the cloud. And I think a lot of it’s driven by DHH. Maybe there are others that have done this, but I think he’s been the most vocal lately who’s been talking about doing this. And basically 37 signals looked at their cloud spend and they said, Hmm, we’re paying, I dunno, it was like 10 million a year for S3 or something. And then there had EC2 instances that were probably similar orders of magnitude. I think they were probably looking at their staffing and saying, well, we have all of these DevOps people on staff who were kind of bored and I think we could, this is the key. And so we have the talent on staff, we’re not necessarily shipping a bunch of new products, so we have extra capacity and we could probably stand to save some money and pull this stuff in-house. 

A very mature business, and they understand their traffic patterns really well. So they know Basecamp, I’m sure just kind of mostly chugs along at a very consistent rate. And if they get a bunch of new customers, it still doesn’t really move the needle so much. So they just have very strong understanding of this is how many servers we need and maybe in six months we’ll need to buy one more server, but we can anticipate that we can predict it. So yeah, I think for them, they’re at such a stage of maturity with so much in-house expertise that sure have at it do it, but I think, I guess I wouldn’t say a hundred percent of us, but for 99.9% of the rest of us, this is not a good decision. With platform as a service, you’re effectively getting all of those site reliability engineers and DevOps people at your platform of choice. 

You’re getting all of them as a functional extension of your team for metered cost. Usually it’s like an extra $10 for the next size up server. And so you’re getting this incredible amount of expertise for very, very low marginal cost. This is just close to a miracle for us people at a smaller stage being able to start out crack open a fresh application and you have close to $0 a month in cost, and then as you get more customers, you can just incrementally expand your resource usage at your platform as a service. This is just, I think, a no-brainer to stay in these systems. And yeah,  

Rob Walling:
You and I are on the same page with this. It’s the old thing where I say, don’t use Steve Jobs and Apple as an example, unless you are co-founding with a guy who invented the personal computer, Steve Wozniak, and when you’re 20 years old, you’re worth 1,000,021, you’re worth 10 million, 22, you’re worth a hundred million. It’s something approximately that I think that was Steve Jobs situation and you’ve started this incredible company, great. Then you can take Steve Jobs advice of not listening to your customers. They don’t know what they want until you give it to them. Or if you started a SaaS that is, as you said, 20 years mature has nine figures, hundreds of millions in revenue, we would guess tens of millions a year in profit, which is confirmed. Jason free confirmed at MicroConf, and you are so bored, I’ll say that you’ve rewritten the app multiple times. 

They have a Basecamp V two and a Basecamp V three. And when I say bored, I have a ton of respect for DHH and Jason Freed, and I think they’ve done a lot for SaaS and a lot for entrepreneurs. They are TinySeed investors, they’re TinySeed mentors, so they’re in our circles, but I think that they’ve been successful in spite of a lot of the advice they give when they to say, we don’t do marketing, we don’t track analytics, we don’t track opens, we don’t have any type of web analytics or conversion tracking on our website. They used to say that, I don’t know that they do anymore. And it’s like, yeah, and if you built a SaaS in 2005 and we’re one of the first ever, and you also don’t need that, but none of us are in that position. So just really take the stuff that they do with a grain of salt, these outliers. 

And that’s what we’re talking about here. In fact, if a TinySeed company, if I was interviewing a TinySeed applicant and things were going well, and they told me that they were racking their own servers, it would be a major red flag for me that I would dig into and I would say, why are you doing that? And they would better have a damn good reason. And I believe of all the TinySeed companies, 204 investments we’ve done, there was one founder who had physical servers, and the reason was is it was like three years ago and he was doing AI and needed physical GPUs. It was way too, he built his own model. It was before the chat. GPT became a thing right about two years ago. And he convinced me in a R because a R has a PhD in science and I know my way around a keyboard. And we dug into that with him. I Wait, what? And he was like, yeah, and he kind of showed us the cost and we’re like, ah, you actually, that is the right choice, but that’s it. He’s the one out of 204 that we’re like, okay, yeah, physical, alright, fine, cloud is more expensive, but it’s not as expensive as hiring your own SREs and DevOp folks.  

Derrick Reimer:
Yeah, and I think, I don’t know, even over the years, cloud’s kind of started out as a virtual private server, whatever, where it’s just like it’s an on-prem server that they will make sure to keep the power onto it, but everything else is managed by you. So you’re still doing a heavy amount of DevOps and some people choose to go that route. But even that, it would be so hard for me to justify just getting a digital ocean droplet and just trying to rotate my own server logs and doing all this stuff where you just don’t have to deal with that anymore. And you can get so much more reliability by going with a more modern platform as a service where you have a docker file that describes what your server should be and your host would ideally allow you to just say, deploy servers that follow this spec and they manage all other aspects of it. This is the way to go unless you’re planning on investing in having your own people on constant on-call rotation and doing a bunch of DevOps work. Some people are passionate about that, more power to you. But if you’re just starting out especially, you do not want to have to be responsible for that portion of your reliability.  

Rob Walling:
Your most valuable asset as a founder is your time. And anytime you spend not building value for customers, not selling, growing the business, even if you enjoy it, it is detrimental to the business. And your number two asset is money, but only because money buys you time. See number one, you know what I mean? And so let’s say you are truly bootstrapping. If you’re around with DevOps, that’s a catastrophic mistake. And let’s say you raise a million dollars and you hire a DevOps person to run your servers out of that million and you save the same amount of money or something, just you have another person on your team that really you probably don’t need, I just can’t justify it as you said, 99.9, so maybe one out of a thousand or one out of 500, there’s some number where this is probably appropriate, but otherwise don’t make this mistake. And it is a little bit of, yeah, it’s cargo cotting in a way. It’s kind of like it’s listening to advice from the wrong folks. It’s just also like don’t follow Silicon Valley founders and see how they grow their company and then as a bootstrapper think that’s how you’re going to do it because usually their advice doesn’t apply. So thanks for that question Lewis. Hope it was helpful. Our next question gives me some PTSD Derek, so I’m going to let you weigh in first. This is from Kyle.  

Kyle:
Hey Rob. Name is Kyle. I’ve listened to the show for a long time. I’ve actually had you answer one or two of my questions on the show, I think in the past just to kind of get straight to it. Basically, we are a client and project tracking tool for tattoo artists that also allows them to schedule appointments and send appointment reminders and whatnot. So my question really is about that last piece of the puzzle. Think of like a Vagaro acuity or Schedule Lista. There are lots of them out there. One of the issues that we’ve heard from artists that we’ve talked to so far that they have on those other platforms is emails getting sent to spam or not delivered or similar issues. And what I’m trying to figure out is just kind of steps or steps I could take or tools I could use to help limit that risk. The plan is to use SendGrid for all those automated and customized messages that come out of the platform to our artists’ clients. But yeah, any insight you might have on how we can help mitigate that risk just from day one would be greatly appreciated. Thanks in advance. Hopefully you get time to hit this question and keep doing the good work. You’ve been a great help to me so far, as well as countless other founders and hopeful founders. Thanks Rob. Take care.  

Rob Walling:
So Derek, having never managed email, sending infrastructure nor dealt with blacklists and deliverability, would you care to weigh in? Should we give folks background who don’t know that we started Drip, which was sending, I don’t know by the time we left. So we sold it in 2016, left in 2018, I believe, and it was sending 150 million emails a month maybe by the time we left. We started what with Mandrel and then we used Mail Gun and we used Send Grids, we used multiple sending providers. But to our point from the prior question, we never did spun up our own email sending servers. You know what I mean? That’s kind of equivalent. It would’ve been way cheaper. I mean, our SendGrid bill by the end, it was a lot of money and I don’t remember exactly, but certainly tens of thousands of dollars a year might even have been six figures. And so could you justify hiring someone in gives service? Maybe I would do it. All that said, this is about email deliverability. What are your thoughts  

Derrick Reimer:
Here? Yeah, so I mean taking the point that even running Drip, we never endeavored to do our own email sending. I mean, I think point number one is definitely use a mature provider in the space like SendGrid Postmark is another good one that’s been around a long time. I use them for Savvy Cals sending, I think Amazon has a service. So yeah, there are a handful of these out there that their sole purpose is to manage pools of IP addresses and sending reputation with all the major email service provider or ISPs. Postmark just recently had an incident where a bunch of their emails coming out of their system were getting flagged by Google, I think. And at a certain point, a status page post came out and they’re like, we are actively talking to Google to mitigate the issue. And within a few hours the issue was mitigated. 

And can you imagine if you were responsible for getting your stuff into people’s inboxes, not using a middleman that, are you going to call up Google? No. So I don’t think that’s what our asker here is asking. He obviously knows use an email sending provider, but I just wanted to underscore that point that you definitely want to use a SendGrid or something like that. Beyond that DRC has become a very important part of email authentication these days. I think all the major email providers look for a strong DMAC policy, so you can Google that or chat tot that to get more details about it. Look to your provider of choice for instructions on how to make sure you have that stuff dialed in for their system and of course your D Kim and all the DNS level authentication stuff to make sure that your own domain is in a good place. 

And then I think the other couple other big pieces here when you’re sending email, so it sounds like this is similar structure to what Savvy Cal does, where we send new appointment emails and reminders and things to people who schedule through our system. So we have a lot of emails going out to people scheduling into the system. And so you need to make sure that any place where an email address can get in, so that’s through the booking pages that you have good spam protection there because if you have people, malicious actors hitting those and putting junk email addresses through your system, that’s going to reflect back on your domain’s reputation. And that’s the biggest thing with email deliverability. Point number one is always like, well make sure you have high quality sending. And it’s like, well, easier said than done in a lot of cases, but the biggest thing you can do is make sure that you’re protecting all the places where emails can get in. 

I also these days, like to use an email validation service, I use emailable, and you can run email addresses through it and it’ll confirm to the best of its ability, whether it’s a valid email address and that keeps your bounce rate low. And just make sure that the less invalid email you attempt to send on your domain, the better for your reputation. And then also MX Toolbox, that’s kind of an oldie but goodie tool. We used it back in the day with Drip and it’s still around and you can use that to just kind of keep tabs on your email sending reputation and just make sure that you’re not ending up on blacklists.  

Rob Walling:
There it is. That’s a clinic. We should make that into a course. The five minutes of you talking there, I don’t have much to add, although I thought of a couple things. We used mandrel and their deliverability was phenomenal back in the day and they allowed marketing email at the time and then they kicked everyone off sending marketing email, and it’s only transactional. Now these days, if you were to ask me who I would use, it’d probably be mandrel. If it’s purely transactional, I would verify that and check with other people and this and that. But I remember, I mean MailChimp’s infrastructure is so good and they’ve had it for so long that Mandel is an extension of MailChimp, so definitely add that to the list of the postmarks and the sends grids. The other thing is there is a TinySeed company that kind of helps with all this. 

They’re called Sky Snag and they help with D Kim and they monitor all this stuff in monitor phishing and this and that. So if you’re listening to this and you’re like, man, I don’t really know what I’m doing, sky snag.com is probably a place to check out. And the last thought is, I agree with everything you said, and I think that’s great advice. The other thing is this is why a lot of folks, have you noticed how many are asking for phone numbers and doing SMS now for reminders, my haircut place only does SMS, and it’s because they know inboxes are full. And there’s the multiple inboxes. I don’t use these in Gmail, but where it’s the promotions tab, reminders get in there, and SMS is more direct. Now your RSMS inboxes, so to speak, are getting crowded, and so then we’re going to have to move to WhatsApp or something. 

But that’s the other thing to think about is in certain areas, especially appointment reminders and stuff, most of the ones I receive now are via text. And so that doesn’t remove the need to comply with stuff. There’s a bunch of regulations around this. We have, I mean, gosh, there’s got to be, there’s at least 10 probably TinySeed companies where SMS is their main focus and their main value prop. And then there’s probably another 20 if not more, that actually send SMS. And so there are some hurdles there, but if you use the equivalent, if you use a Twilio and you don’t need a dedicated number for each of your customers and you just have a few, I think it’s significantly less complicated than one might think. So thanks for that question, Kyle. I hope it was helpful. Our last question for today is for Mike.  

Mike :
Hi Rob. My name is Mike. I was listening to your A MA for the SaaS launchpad course, and you made a statement that for enterprise or other companies selling to the enterprise, you’d want the salesperson. And for somebody targeting lack of a better term, SMBs, you’d want the marketer on your team. Can you expand on your thinking between the differences and how does that affect what your choices are in terms of what kind of business you might want to start or how you give advice to people deciding what kind of business they want to start, and then how critical is it to have either skillset on the founding team or can for example, the sales side for the enterprise, could that be learned from something like founding sales book by Pete Kanji? Thank you for everything you do.  

Rob Walling:
I think the first question maybe is sales versus marketing. What’s the difference? Just very fundamentally, and the way I think about it is marketing is generating demand. So this is going out and getting in Google search results or running ads or doing any of the 20 B2B SaaS marketing approaches I have in the SaaS playbook doing integrations and having someone else talk about you to their audience, any of these things that gets you and your app and your value proposition in front of them, and that is inbound. And then people come to your website and they either book a demo or they sign up for a free trial or whatever. That’s marketing, it’s spreading. The word sales is a couple things, like sales is often outbound, right? It’s approaching on LinkedIn, it’s all the cold emails we get on a typical day. It’s the Twitter dms and all that stuff. 

And then it’s doing the demos face-to-face in essence, right over Zoom. It used to be in person obviously years ago, and it is trying to close a sale. So just marketing versus sales. Those are two things. They’re complimentary. Sometimes if you have a low touch, no touch funnel, you don’t have sales at all, you don’t do any outbound and you don’t do demos. And so it’s purely a marketing driven SaaS, and that is kind of the bootstrapper indie hacker dream. It’s pretty rare. And in fact, those tend to need lower price points, so then they have higher churn and they don’t grow into multimillion dollar businesses. It can happen, but it’s often bringing sales in, doing demos and closing the $2,000 a month deal versus the $200 a month deal or the $20 a month deal that really kicks that engine into growing into that seven or eight figure mark. So I want to kind of want to level set that with some definitions, but then beyond that, enterprise versus SMBs, what do you think?  

Derrick Reimer:
Yeah, there’s also kind of different levels inside of here. The definition of a small business. Some would say a small business is up to a hundred million a year in revenue or something. And it’s like that’s usually not what we mean by that. When we say small business, we’re usually talking much, much smaller, maybe a team of 10 people or something like that versus, and we often qualify enterprise as capable of buying at a higher price point. Basically, the grass is always greener on the other side. I will say this is kind of a choose your paying. If you go with the higher price point product and you’re able to do sales for it, then you typically have longer cycles. There’s more involved, more investment involved in order to make a sale. Cost to acquire a customer is often higher, but you can offset that with a higher price point. 

But I’ve talked to plenty of founders who get frustrated with this motion and needing to have all these conversations and make it through procurement and all of the kind of sucky things that come along with that piece. But on the flip side, when you’re going marketing driven, you generally have a lower price point. People are more price sensitive, churns higher, and it can be really difficult to move the needle on these lower price point businesses. Don’t ask me how I know that asking for a friend because each individual customer is only paying you a little bit of money and you have to be really good at getting a lot of distribution and showing people who care enough about the differences between you and maybe a bunch of other options on the market. Why should they pick you? And some of this just comes from being around for a while. 

It just takes a while to build up reputation where you’re showing up in conversations and making it onto the lists where people are comparing different products and making it into chat GPT so that it starts recommending you who the hell knows how to actually do that other than just be around. So there’s so many things involved with the marketing driven approach where, and obviously if you’re a master marketer, you will have better success at building up a business with this approach. But I guess all that to say, both of these approaches take a certain amount of expertise, and I think typically the more sales driven approach is something where you can kind of brute force it a little easier, I guess, than the marketing driven approach, which is like you just have to get really good at generating a ton of traffic in the right places. And that is a bit of a dark art for many of  

Rob Walling:
Us. And I like the way you put that. And it’s not just two, it’s not a dichotomy. There’s small and medium sized businesses, which I think of like, oh, is there one decision maker? Usually oftentimes is the decision maker also the user of the software. So maybe that’s even solopreneurs or prosumers. And they’re extremely price sensitive because they think of the money they’re spending as their money. So they think of it almost like consumers and then SMBs maybe a notch up. And then there’s, there’s enterprise and there’s this whole spectrum of it. But generally the tough part is if you get these big, big contracts, let’s say you’re selling $250,000 a year, which we have some TinySeed companies that’s their contract size, but they only close a deal every quarter or every six months, and it’s just brutal and agonizing and there’s no momentum and it’s super, it’s spiky in a way that’s just not that fun. 

And as you said, months in procurement, but they close these really big deals and they have this negative churn. Everybody’s expanding. That’s great, except for it’s kind of agonizing as the founder. On the flip side, you’re charging 15 bucks a month and your churn’s really high, and it’s hard to outrun that churn. You need a really massive funnel. I’ll say it’s impossible to outrun the churn of a $15 a month business if you want to become, let’s say a 10 million or 20 million business. And that’s why when we started Drip, it was $50 a month, a hundred of one 50. I think those are the price points, but we soon realized people were reaching out with really big email lists and they were like, oh, I would pay you $500 a month based on my list size or a thousand dollars a month, or I think when we left, there were people paying us two or three grand a month. 

And these are not enterprise in the enterprise sense, but for us, they’re enterprise. It’s anything over 25 KI kind of think of as a bootstrapper enterprise plan. And that was pretty interesting. And I call this a dual funnel. It’s where you have folks paying you no churn and paying you a lot of money on the top end, and you do have some folks doing sales demos and procurement maybe, but then you have this nice low touch funnel that usually has higher churn and more price sensitive customers on the bottom end. And this only works in really, I think, pretty big markets. It doesn’t work in these tight niches, but that’s a nice way to even it out to get a little bit of the best of both worlds if you can swing it. And we’ve seen like Ruben with Sewell has that type of fun and Riverside or squad cast has that type of funnel. And it’s not always possible, but it is a way to even out the agony of, oh, we close a deal every three months. This is fun. This is so great.  

Derrick Reimer:
And I can imagine just if it’s that extreme where it’s like hundreds of thousand dollars from individual customers, and I’m sure there’s a bit of cost involved with maintaining those, so it’s like you would want to make sure you have the people on staff to maintain that. But if one of those churns and it’s like, oh my gosh, we have to replace them or else crazy. Yeah. So yeah, having a bit of diversification is probably ideal where you don’t have too much customer concentration to the point where you’re like potentially have to lay someone off from your team if a customer cancels or whatever, but it’s business. Yeah,  

Rob Walling:
You got to do it. So thanks for that question, Mike. And Mike actually sent in another question, but we don’t have time to get to it today, so I will answer that in a future episode. Derek Rimer. Folks want to use the best scheduling link on the internet head savvy cal.com, and as a reminder, if you are an agency or a freelancer consultant and you are building solutions for folks who are booking appointments, this is like service businesses and other types of folks, you should reach out to Derek to find out what he’s building. I’d love to chat.  

Derrick Reimer:
Yeah,  

Rob Walling:
D-E-R-R-I-C k@savvycal.com. Thanks again for joining me.  

Derrick Reimer:
Thanks for having me.  

Rob Walling:
Thanks again to Derek for coming on the show and a reminder SaaS launchpad.co and use the code launch to get $150 off the course as well as MicroConf Remote is happening tomorrow. That’s MicroConf dot com slash remote. Thanks for listening this week and every week. This is Rob Walling signing off from episode 775. Derek, I have four fifth edition Dungeons and Dragons trivia questions for you. Order, oh boy. From easiest to hardest. The first is in Combat in fifth Edition, Dungeons and Dragons. What determines the order in which characters act  

Derrick Reimer:
You roll for  

Rob Walling:
Initiative? There it is. Alright, that’s one out of four. Alright, second question. If you attack a prone enemy with a melee attack, what do you gain  

Derrick Reimer:
Advantage.  

Rob Walling:
You do indeed gain advantage. Two out of four. All right.  

Derrick Reimer:
Capital A, advantage  

Rob Walling:
A. All right. This one, it’s getting harder.  

Derrick Reimer:
Oh boy.  

Rob Walling:
When attacking a creature you cannot see because they’re invisible, hidden, et cetera. What disadvantage do you suffer?  

Derrick Reimer:
Do they have surprise on you?  

Rob Walling:
Incorrect? What disadvantage do you suffer when you attack them? If they’re invisible?  

Derrick Reimer:
When I attack. Oh, okay. Okay. Well wait. You can make an attack on someone who’s invisible. They’re there but you can’t see them.  

Rob Walling:
Or hidden, like if thieves or rogues, I guess can hide in shadows.  

Derrick Reimer:
I see, I see. I dunno.  

Rob Walling:
You have disadvantage on your attack roll.  

Derrick Reimer:
No, it kind of makes  

Rob Walling:
Sense, right? I  

Derrick Reimer:
Should have just guessed that.  

Rob Walling:
Totally. Alright,  

Derrick Reimer:
What disadvantage do you have? Disadvantage.  

Rob Walling:
Yeah. It has it as lowercase D. It should be uppercase. And then the fourth and final. That one’s too easy. Oh my god, these are really easy. I asked cha GBT for 10, and I said, make them easy to hard, and I meant like eight, nine, and 10. And they’re just, they’re gimmies. Listen, if you roll a natural 20 on an attack row, what’s special effect occurs if you roll a NA 20 on an attack? Maybe it’s not as easy as I thought.  

Derrick Reimer:
See, this is rare. It’s only happened a few times. Yeah.  

Rob Walling:
Happens now and again. Yeah. So if you were to roll a Nat 20 on a bow attack or a sword attack, what do we do?  

Derrick Reimer:
Do you get an extra attack?  

Rob Walling:
It’s a critical hit is what it’s called. And you roll your damaged dice twice, but you only take the bonus. You know how there’s a damaged bonus? You only take that once, but you roll the damaged dice twice,  

Derrick Reimer:
Right?  

Rob Walling:
There are some dms that it’s all, A lot of this is house rule, but let’s say you’re doing a D eight plus three damage, right? Some dms will just say you do max damage automatically. Some will say you do two dice max damage. That feels like a lot to me. That would be 19 points of damage and others roll it twice. That’s what I do. Roll it twice at add three. All right. Bonus question, what condition occurs if your hit points become negative, equal to or greater than your maximum hit points from a single attack. So let’s say you had 20, your max hit points are 20 and you took 41 points of damage with a single attack, like a dragon breath. What would happen to you?  

Derrick Reimer:
Is that an instant kill?  

Rob Walling:
It’s,  

Derrick Reimer:
Yeah,  

Rob Walling:
You immediately die versus the fifth edition. Freaking death saving throws.  

Derrick Reimer:
I was going to say, you must be extremely dead in that case.  

Rob Walling:
Yeah, I actually played as 10 below. Is that what I did? See? I said if you went negative, I think I’ve been house rolling negative 10.  

Derrick Reimer:
Yeah, I think so. I  

Rob Walling:
Think that’s what I did. Which is interesting. It’s slightly more deadly, but if you have 50 hit points, let’s say, and you go to negative 10, that can happen. Meaning if your max is 50, but as you start to get down and get damaged, going to negative 10 is not unheard of. So I like there to be dead possible man.  

Derrick Reimer:
Yeah. And in our campaign, none of us have ever died yet with that house rule.  

Rob Walling:
No, no. But in the other, the early, the delver one where it was first and second level characters, I lost two. Lost two in one dragon breath. And I was like, oops.  

Derrick Reimer:
I bet you went kind of easy on us on the other campaign that we started years ago. You probably,  

Rob Walling:
I did. I didn’t really want anybody to die. I mean, I didn’t fudge dice rolls, but I always kind of made sure it’s really hard. DMing not really hard. It is difficult. DMing first and second level campaigns, so fragile.  

Derrick Reimer:
Yeah.  

Rob Walling:
Well, Derek Kremer, thanks for playing.  

Derrick Reimer:
Well, that was nerve wracking. Thanks for having,  

Rob Walling:
You did. Okay. You got four out of five, if I’m counting correctly.  

Derrick Reimer:
Yeah, let’s call it  

Rob Walling:
That. Yeah, let’s call it that. He says.