I gave ChatGPT, Gemini and Claude the same request twice each, to price a simple operational web app for a septic company, and the answers ranged from $42,000 to $128,000. Those numbers were plausible for years, and a lot of that range is what custom software used to cost before AI changed how fast it gets built. That shift is why the custom versus off-the-shelf question deserves a second look right now.

We build custom software for a living, so read this with that in mind. We also turn people away regularly when something off the shelf is the better fit, and most of this is about how to tell which situation you are in.

Read the full transcript if you would rather not watch.

Jump to a section of the video

The deal we all took

Somewhere along the way, renting became the default. You pick a plan, then you need one more feature, then support, then more storage, then another seat. Each step is small, and over enough years businesses stopped noticing it. Meanwhile you own none of it.

That is fine for plenty of things. Fifty dollars a month for a CRM is a low barrier and you can walk away. The question worth asking is whether a given system is something you would rather own as an asset of the company or rent as an expense. Most owners never ask it, because there are a hundred products to choose from and choosing one feels like the decision.

What AI changed

The cost of getting software to run has dropped hard. A developer on Reddit described being shown two quotes for the same booking app, one for $9,000 and one for $74,000, and my AI experiment produced the same kind of spread. Owners are now up at two in the morning asking an AI what their idea would cost, and getting confident numbers back.

That access is genuinely useful. It also muddies things, because a one-page description of an app hides an enormous difference in scope, and the price gap between two quotes is usually that hidden scope rather than somebody ripping you off.

The catch: software you can bet the business on

Getting software that runs is the cheap part now. Getting software you can run a business on is still the real job: security, backups, updates, and someone to call when it breaks.

We were brought into a company whose internal IT team had built a pile of tools with AI, and the security problems were serious. We have also seen businesses whose databases were reachable straight from their public website. That is the part of what you pay for with a good SaaS product that is easy to forget, and it is the part a cheap custom build tends to skip. We wrote more about that in what AI-generated code gets wrong.

What to just buy

Some software every business uses the same way, and building it is a waste of your money. Accounting, payroll, HR, email. Payroll and HR in particular carry real compliance and legal weight, and the companies that specialize in them are better at keeping up than any custom build will be. When someone asks us to build their accounting system, we tell them to keep QuickBooks or whatever they use.

The surprise for a lot of owners is that buying an ERP rarely skips the building part. You still pay consultants and implementation teams, sometimes for a long time, to make the product fit your company. You also pay for every feature in your tier, when you might only need five of them. A system built around those five has no filler in it.

Six questions before you build anything

  1. Is this the same at every company, or is it how you do it? If you keep bending your process to fit the software, that part may be worth owning. If it is email, nobody's business runs differently because of Outlook.
  2. Are you bending the business around your tool? If the software is costing you operations because nothing fits how the work flows, that is a real cost even when it never shows up on an invoice.
  3. Will renting cost more than owning? Run the math over three years, then five, then ten. Seats, price increases, data fees, consultants. Compare the whole picture to building and maintaining your own.
  4. Could you fix what you have instead? A lot of the systems we get called about semi-work and just need fixing. That is not glamorous, and it is often far cheaper than replacing something because it feels old.
  5. Who is on the hook when it breaks? Owning software is like owning your building. When the roof leaks, it is yours to fix, which is control and responsibility at the same time. Renting means you wait on the landlord.
  6. Do you want to own your data and the system it lives in? Your data is worth more than ever right now, and the systems your business runs on are part of what a buyer is paying for if you ever sell.

The third option: fix what you have

If renting is not working and building something new is more than you want to take on, there is a middle path, which is getting the current system looked at and quoted to fix. One Reddit post I liked came from someone whose father had run a hardware store for thirty years and who wanted to modernize it without breaking what already worked. That instinct is right. Sometimes the answer is fixing the two things that hurt, not replacing the whole thing. That is most of our takeover work.

When custom is overkill

Custom is often overkill for very small businesses, depending on margins. It is overkill when someone has a vision for a system that nobody in the company will actually use. And it is overkill when a couple of off-the-shelf products would do the job, so test those first. A CRM is the usual example. Most companies do not need one built, and I will say so.

The takeaway

It is not really custom versus off the shelf. It is knowing which parts of your business are standard and which parts are yours. Rent the stuff every business does the same way, and own the engine that makes your business yours.

If the engine in your case is an ERP, we went deeper in custom ERP or off the shelf for a small manufacturer, and the specific pieces we build for plants are on our manufacturing page. The AI quote experiment from the opening is written up in how much should business owners trust AI.

If you are weighing this, tell us what you are renting today that you would rather own. Reach out here or call (507) 388-4748.

Full transcript

What I said in the video, start to finish. It is unscripted, so it reads like talking. Each timestamp opens that moment on YouTube.

0:00 And so I asked ChatGPT, Gemini, Claude, all the exact same prompt twice in two different sessions, I didn't change any settings, all about what it would take to create a web app for a septic company, right? And it was just like an operational web app, a light ERP, let's say. And the prices on all of them ranged from 42 grand to 128,000. Like when I would quote someone who came to me with that,

0:29 I'd be like, oh yeah, I'd probably take five to 10K because a web app is super easy now to build with AI. There's not a ton of labor involved anymore. The conversation between custom software versus off the shelf SaaS products for businesses is something that absolutely should be explored and discussed. And for some reason, I feel like it's something that's just been put to the back burner in some respects of the conversation because of years and years and years of people being burned either by custom software or just being trained that SaaS is the way to go. My name is Tim. I'm a partner at Norsoft.

1:13 We've been building custom software since 1987. We love this. We love helping businesses. We're very honest people. And if we have someone come to us with a question, if there's already a solution that is better fit for them, then we offer that. So the deal we all took, so this is just talking about SaaS, right? You have your base plan. You need a feature. You need a support.

1:42 Maybe you want extra memory. You're constantly being chipped away in these SaaS models, right? And I think the interesting part here is that it's something that we've just become numb to or used to, and it doesn't have to be the case, right? And you don't own any of it. As a business, this core part could be an asset to the company, or it could be an expense. It's more like as business owners, we are like, oh, you know, I need a CRM, and there's 100 CRMs

2:19 to pick from, and so I just need to go decide on one. And while that is the case, maybe going out and signing up for a CRM, it is a low barrier to entry to pay 50 bucks a month for a couple people, but you don't have a lot of control over it, right? And so there's this balance in the discussion, right? Is this something that you actually want to own and have as an asset within your company, or is it something that you're okay with just renting, right? So I thought it'd be fun to tie in some Reddit clips, you know, and I think this one's funny, the title. Anyone else angry about being held captive by companies converting to

2:58 subscription services? Like, do you remember back when you used to just, like, buy a program, and you had the license for it, and it just was sat in your company for years and years, and now they're all converting to, you know, subscription-based? So that made me laugh a bit. And like the comment says, yeah, this is all by design. This person hates it too. You can have a custom system built, if it makes sense. It doesn't always make sense, and I'll be the first one to say that.

3:28 So what's changed? AI. Let's talk about it. The thing is, back in the day, 42 grand or 128 grand, depending on the firm, these were very plausible prices, right? So that's why I think this Reddit quote is funny. An owner who paid for custom software or an app, what did it cost, and what would it do for you again? Literally the same thing. This guy's saying he's a developer, and last month someone showed him two quotes for the same booking app, 9K and 74K.

3:56 So I think that's really funny, but it's also empowering business owners who are, you know, up at 2 a.m., they're ADD, they can't turn their mind off, and they're talking to their phone, talking to AI about their next idea, and it's quoting them all of these things of what it would take and how much it would cost. But AI is making this access to information more attainable, but it's also kind of diluting it. So let's jump in to what the catch is. Software that runs, you know? Getting here is a lot easier now.

4:31 AI made this part cheaper for sure. Software you can bet the business on, right? Security, backups, someone to call when it breaks, right? So let's go over this Reddit comment here. From the owner side, the 9 to 74 gap, which you're seeing usually, isn't one of them ripping you off. It's that a one-page description hides a huge scope difference, right? And blah, blah, blah. Everything I've been talking about here. So owners get burned often by buying something cheap,

5:05 but also like the real cost might be 65,000 in this case. Yeah, the software runs, right? Like AI is building stuff right now that runs, right? People are running their businesses on them, but there's issues too. There is issues. We got brought into a company where their IT department vibe coded a bunch of stuff. And there was massive security issues. I can tell you how many companies, even from their websites, I've seen their databases publicly accessible on the front end. So this is the thing.

5:43 Software can easily be built right now. Now it's about quality software being built and software that's secure and actually functions and doesn't break. And you know, all those things, which is part of what you pay for when you get SaaS, right? You're supposed to be getting a service that is maintained, updated, secure, all of these things. AI adoption is on the rise. More software is being shipped, but it's less stable or less secure or a lot of stuff. Things that I would suggest you buy, accounting stuff,

6:17 email, calendar, I mean, depending on the calendar stuff, sure. We can, we can build that easily, but payroll and HR things that are really super heavily need a lot of compliance and are legally sticky. I would suggest sticking with someone who is a SaaS in that, in that space, right? So like accounting, like a lot of people use QuickBooks, for example, right? If someone asks us, I'll just keep using the accounting one as an example. If they say, Hey, can you build this for us? We're just, no, we would, we would suggest you stick with this X,

6:51 Y, or Z product. The next thing I wanted to highlight is that off the shelf things often need to be built or modified, right? So we'll use ERPs for example. Buying one rarely skips the build part or implementation, but you're going to still pay consultants and implementation teams a significant amount to make it fit your company, or you might have your internal team have to work on it for ages. So this is something that like on the front end, before you get out there and look at ERPs for, let's say your manufacturing company.

7:31 This is absolutely why I suggest people consider custom software being built a custom ERP. We've built many of these, right? Because what ends up happening is you'll go out there and you'll sign up for a tier and you're paying for all sorts of features that you guys don't need. Like just simply put you might only need five things. And so if something's custom built, you don't have a bunch of hamburger helper in there, right? It's just the meat and you're not paying a bunch of overhead and all it's just, it can be a significantly better investment to have something custom built, but not all the time.

8:12 And it really is a case by case basis. And that's why as a business owner leader, I really hope that you weigh the options and bring custom into the conversation. Let's go over six questions right now before you build. Is this the same at every company or is it how you do it? Like, and what I mean by that is maybe, maybe there's a lot of tools out there for other people in your space and you just like, they just don't itch the scratch that you have specifically for your processes. So if it's like you feel that you are having to

8:53 conform your business to how these SaaS models in your industry are set up, then it probably is worthwhile exploring a custom option, right? Are you bending the business around your tool, right? So is the tool costing you money? Is it costing you operations just because you can't find something that fits to the flow of what you want? Like if it's fine, you know, like Gmail, everyone's used to Gmail or, or outlook, right? You're not changing how your business operates because of that, how that email platform works, right? So you just keep using that, right?

9:30 There's no need to deal with that. Will renting cost more than owning over three years? Oh, three years. Sure. Five years, 10 years. Like I would really run, run the math equation as long as you can. Seats, price, how much you're paying for data, consultants, all of those things, try to take in that full picture and then compare it to the long-term costs of having something custom built. This is one that we come into a lot at Norsoft. People already have systems that semi work and they just need fine tuning.

10:02 So we fix it and then you don't have to put in a whole new system. We just fine tune it and help it work. And it might not be the sexiest thing on the block, but it works. And why, why cost the company hundreds of thousands of dollars just because you want to replace it with something new and shiny when the existing thing just needed a couple of tweaks. Okay. This is a fun one. It's someone on the hook when it breaks. So security, backups, updates, somebody has to own it, right?

10:34 Building it can be a safe bet. Renting it. Someone else is on the line. If you were to rent the building or own the building that your business operated in, do you want that control and, or do you want to rely on somebody else? So if depending on your lease agreement, you know, maybe the roof starts to leak or gets ripped off in a storm, right? As the owner, you're the one who's going to deal with that problem, right? And if you're renting, you're at the mercy of your landlord. Lastly, do you want to own your data, the system and the system that your

11:10 business runs on? And I know I'm probably kicking a dead horse here, but if you own your data right now, this is what's super valuable right now when it comes to AI: data, data, data, right? So if you're owning the data that your company is collecting and generating on your customers or on whatever you're doing, right? Owning it is such a huge deal. And so, and, and obviously the system, but this sixth point, this really harps for me on, on the point of how much of your business do

11:45 you want to be an asset for eventual exit or an IPO? Now for a third option, just want to throw this in there. If renting or building a new platform isn't, isn't on your scope, you can fix what you have, right? It's just a matter of having it audited and quoted out. Is this, is this something that would be more economical and more logical for me just to fix what we have? This Reddit one, this is great. My father has, has run a hardware store for 30 years. I'm 23 and I'm trying to modernize it without destroying what already works.

12:20 Where would you start? And these guys, you know, sometimes you don't need to like fully modernize stuff. When is custom overkill? It depends on the size of your company. It depends on the cost, right? But it's overkill often for, I would say very small businesses, right? If you're like a five man shop and you want this massive product built, it's probably not going to be worth it depending on your margins,

12:50 but who's to say. I think custom software is overkill when people just have a vision and they want this thing built, but actually nobody in their company is actually going to adopt it. I think that's overkill. I think it's overkill when you could test out a couple of SaaS in your, in the space that you're trying to fill in your company and, and see, you know, like if it's a CRM, right? Like I think it's overkill to go build a CRM custom built, but I love to be very honest with people.

13:22 Like absolutely not. You do not need something custom built right now. Rent the stuff every business does the same way and own the engine that makes your business yours. You're working so hard to build your empire. It's like, why wouldn't you own as much of your business as you can possibly and not be relying on other people? It's not custom versus off the shelf. It's knowing which parts of your business are standard and which parts are yours.

13:48 And in effect, this is just trying to highlight how much do you want to own of your business? That's that's the core of it. Please, please, please bring it into the conversation. If not with me, somebody else, but I hope this was helpful. I hope that all of this made sense. My name's Tim. I'm from Norsoft. Thank you for watching.