[lefstathiou]

My brother acquired an aging app (from an aging founder) built on Delphi used by many dozens (or low hundred) of the world’s leading shipping, energy and commodities companies, used as a standard to calculate “laytime” and “demurge” (myriad of fees associated when a ship docks into a port). It used to cost $5k for a perpetual license tied to usb based key that had to be plugged in to activate. If you wanted to use on two machines, you had to buy two licenses with two keys.

Customers in the US and Europe hated the usb, especially during COVID. In random places of Africa, where they greatly valued the single perpetual license, it persists. From my perspective, I don’t see anything positive from being an installed application for this use case - he had to hop through so many security hoops that when he rolled out the web solution IT departments breathed a huge sigh of relief and thanked him.

Over a period of about 2 years he converted almost everyone to saas and 4x’d the annual revenue. That also generated enough fcf to hire more developers to ship more features.

Saas is generally the way to go. Installed apps are common in financial services and industrial applications. I can think of a bunch of other niche examples but I personally would never pursue this model. We put bugs into production from time to time and it is nice to be able to instantly roll out updates.

[satvikpendem]

Thanks for stating this. Some customers (who are often the vocal minority) don't like SaaS likely due to subscription fatigue but most don't realize the amount of manpower it requires to continuously update software that will atrophy without them, not to mention adding more features.

The business reality is often not understood by the users and that's why every company is moving towards SaaS, it allows the company developing the product to continue to stay in business rather than providing a product then shuttering because it couldn't sell enough.

The former is simply more sustainable than the other, much as some (like the vocal minority) might disagree with this fact.

---

That being said, there are many who sell one-time licenses, especially in the indie hacker space on Twitter, such as NomadList and BoltAI. Their model works because they make enough money from their products to retire on, as solo devs, and their products aren't necessarily ones that require constant updates (well, maybe BoltAI as new AI advances come out all the time that need to be implemented, such as RAG, parsing PDFs, storing "memories" like OpenAI, etc, but most advances come through new models, which is just an API call away).

[massysett]

Some users do not want updates.

I understand that updating software takes manpower. Same for running servers for sync or online information or similar.

But I might rather pay once for something that works on my machine as it is now. I need no servers or sync. If I need an upgrade later, I’ll buy it.

I do buy some software as a service but for other software if there’s a subscription I just don’t buy.

[irrational]

Photoshop 6 does everything I ever want Photoshop to do. I wish Adobe would continue to sell a one time purchase copy of PS6 instead of forcing everyone to SaaS. Fortunately I own a physical copy of the PS6 disks that I purchased years ago, so I don’t have a problem acquiring PS6 on any new machines by various means.

[lylejantzi3rd]

A little off topic, but, if you'll indulge me: Why do you think nobody has been able to make a successful alternative to Photoshop? Everybody I know complains about it constantly, and yet it's still the industry standard.

[vunderba]

There are plenty of successful alternatives to Photoshop (Affinity, Pixelmator, Krita, etc). The issue as always is vendor lock-in - when everyone you're doing business with is expecting to be able to collaborate using Adobe then it becomes a huge pain point.

[thayne]

Because Adobe photoshop is what people are tought how to use. It is the same thing with excel. It isn't enough to have a superior product, you also have to overcome the momentum of the prevailing software. That's not to say it is impossible, but Photoshop has a significant advantage.

[SideQuark]

Photoshop isn’t on top simply because people know how to use it. There’s simply no better tool for most users who need that level of photo editing tool.

I’ve never seen a superior product to either Photoshop or Excel. Have you? Maybe they’re on top because they really are good products.

[sneak]

Pixelmator was a wildly successful Photoshop competitor.

Their company was successful too; Apple recently purchased them.

The speculation is that Apple will now compete with Adobe’s subscriptionware.

[trimethylpurine]

Gimp works great. The worst thing about it is the lack of high end YouTube video tutorials. That's what you're really paying for with Adobe, I think.

[araes]

Having used Gimp and Photoshop fairly extensively, I'd agree that Gimp works, and it has the features that Photoshop has (for the most part). However, there's quite a bit of user interface issues with Gimp that make somewhat simplistic activities rather irritating.

Gimp kind of has the Open Source issue where it has tons of features, yet there's a large wall of complexity, zillions of little fiddly knobs to tweak on almost every process, and the interface makes you feel like you need to, because they're all exposed immediately.

Photoshop (personal opinion) is better about having an initially functional feature, with relatively "what you expect" defaults, and then layers of fiddly knobs you can tweak if you "really" want to or need to for a project.

[irrational]

I have all the keyboard shortcuts for photoshop in my muscle memory. Those same keyboard shortcuts don't work in, say, Gimp.

[Gigachad]

You need updates these days or stuff stops working fast. Everyone at every stage is quite happy to make breaking changes without long term backwards compatibility other than a transition period because it’s understood that everything can be quickly updated.

Every time I updated macOS I find that some program stopped working and I just have to update it and it works again.

As well as the fact that most software these days has an online component that has an ongoing cost to provide.

[BenjiWiebe]

Windows is pretty good with backwards compatibility. We bought some software back in the early '00s and it still runs fine on Windows 10. You do have to install the manufacturer's update after installing it off the CD, though. Even though the update says only for win2k machines. :)

[araes]

Frankly, I may have serious issues with Microsoft, yet backwards compatibility is one of the few areas where there's almost nothing to criticize, and MS is almost off the deep end on the other side. You can install stuff from the 90's and it will still have the hardware drivers. It's really kind of ridiculous.

I tried compiling modern software in Visual Studio, and the number of includes for historical support was mind boggling. "Holy s*t, I think MS just added every printer for the last 30 years to my project. There's like a 1000 includes on a 5 file project. Doesn't even print." (maybe a teeny bit of criticism)

[hulitu]

> You need updates these days or stuff stops working fast.

Maybe some engineering course will help. If you make a product that breaks in 6 months, i won't buy it from you. This really means that the amount of testing is minimal and, instead of fixing bugs, you just rewrite the "app" keeping the bugs.

[satvikpendem]

You misunderstand, it is not the application code that changes, it is the code of the environment that the app lives in that changes, macOS is one of the most famous examples of breaking APIs.

[GoblinSlayer]

That's just Apple life: users must constantly pay to keep their stuff working. But if you evade the system API entirely with SaaS, you don't need updates for broken system API. Might as well go with PWA, java or wine.

[bruce511]

The OS is part of it (Apple bad, Microsoft good) but the bigger part is integrations.

Web APIs, import and export file formats, external resources (maps etc) are all part of the equation.

[satvikpendem]

Depends on the size of the company. Perhaps you're simply not the target customer for that company, you self-select out of their customer pipeline which makes it easier for them to handle costs, as it is more expensive to maintain separate SaaS and one-time versions (essentially on-premise, which is often much more expensive and for enterprises who can afford them due to said hassle). However, some solo devs and smaller companies do exist that make only one-time purchase products, because they don't have much overhead.

[vunderba]

> and that's why every company is moving towards SaaS...

This is a bold and not necessarily true statement. It really comes down to your target market. A SaaS is a much less disputed cost when it's targeting businesses but you're much more likely to encounter resistance to a subscription when you're targeting individual consumers.

There is plenty of highly successful mainstream modern day software that offers a perpetual license for one time fee. (DAWs come to mind: Bitwig, Reaper, Logic X, Studio One, Cubase, etc.).

Personally, I think a good compromise is the annual subscription with a fallback perpetual license, a.k.a. the Jetbrains model. I've never had an issue with paying a reoccurring subscription fee, but I take great issue with the proposition that the moment I stop paying I lose all access to the software - it's too close to rent seeking.

[mojomark]

100% this. This helps consumers understand the need for SW maintenance services on their own, without having to convince them... "You want to keep the software alive on your WindowsXP machine after the 2yr support garuntee that comes with your perpetual license is up? Have at it. However, here's a nice service package that will get you back up and running when your ready to upgrade your environment to align with the rest of the world."

I suppose the risk to the SW company is that consumers never learn and just keep opting for the one-time perpetual license every 5 yrs or so, so the perpetual license needs to be priced to bridge that time gap (effectively rolling multi-year support agreements into the perpetual license cost).

I don't belong to a country club, but I've heard they work that way. A big one time payment to join, then annual dues to maintain a membership. If you leave (don't pay dues) for a period of time, then you'll need to pay the big payment again because you haven't been contributing to the maintenance/operating costs of the club to 'keep it alive', so you need to back-pay your fare share.

[GoblinSlayer]

What do you do that windows xp is not enough for you? It has network, file system, gui, and these things didn't change much lately, only hardware support changed, really. Same goes for something like CentOS 6.

[trimethylpurine]

likely due to subscription fatigue

It's due to a-hole fatigue. These are too often just VMs running an installed solution in a 3rd party cloud, run like garbage and cost way too much. There are just too many vendors in the middle to get any expectation of a good experience. And to top it off, every time I buy SaaS the vendor is bought by some private equity giant before the first payment and the product turns to shit by the second one.

That said, it depends what the software does. If it's a platform for sharing or interacting with the public (e.g. eBay), then a true web app makes a lot of sense to me.

[satvikpendem]

> These are too often just VMs running an installed solution in a 3rd party cloud, run like garbage and cost way too much

I mean, you try making such software and let me know how that goes for you. This type of vague criticism sounds a lot like the typical engineer retort of "I can build it myself in a weekend," discounting the real complexity involved.

[trimethylpurine]

I'll try to be more specific. Stuffing a win32 app into a Citrix box and selling it to the next private equity that will offshore your support while your customer's contract milks them for another 3 years doesn't make for an enticing offer for decision makers. It makes a lot of sense for the private equity purchaser who will sell the company again before those contracts run dry and the software is shuttered or replaced.

As the decision maker (also a software engineer) I will work hard to avoid SaaS because it's a sensible move. And that is especially true if other engineers believe as you do that it's difficult to make a good product.

By comparison, the same app, installed locally, doesn't suffer from any of the above problems. There is no contract, no latency, and I don't have any risk if the company is sold. I will likely just have to find another solution provider, in the last case, but at least I'm not locked into additional years of servitude supporting a poor product for my users.

In summary, SaaS itself might be great. But the subscriptions that it usually comes with tend to incentivize bad vendor behavior and a poor customer experience.

[satvikpendem]

Why would there be no contract with a local app? At the company scales you're talking about, enterprise, there absolutely will be. These aren't going to be standalone 100 dollar apps for that level of scale.

[trimethylpurine]

>At the company scales you're talking about, enterprise, there absolutely will be.

Many very large enterprise software providers (Sage, Oracle, IBM) and small OSS shops (Grafana, Zabbix, ProxMox) offer run local versions of ERPs or entire applications with no usage restrictions. The licensing pays for support and updates, not usage. In this model, the software provider has incentive to provide good support, and quality updates. They care to maintain the product because their care is what they are selling.

So, it won't absolutely be the case at this scale. Business as usual is the opposite of that at this scale. I should have to prove beyond doubt that there is no alternative when I agree to sign for SaaS. I'm agreeing to take on a lot of risk when I do that.

>Why would there be no contract with a local app?

Because that's what I want to buy, and for good reason (see above). And someone has figured that out and sold it to me.

[whatevaa]

Companies prefer SAAS, as IT costs to manage that software is non-trivial. Individuals prefer standalone software, as managing costs for that software are non- existant. But company market pays much better. So we get this weird situation of SAAS for individual licenses.

[adriand]

> used as a standard to calculate “laytime” and “demurge” (myriad of fees associated when a ship docks into a port).

Nitpick: I think you mean “demurrage”: https://www.investopedia.com/terms/d/demurrage.asp

[jakeydus]

Very demurge, very mindful

[zxvkhkxvdvbdxz]

He inherited a existing customer base in a niche field. That's not exactly easy to replicate from scratch.

[Waterluvian]

Really shows you just how valuable the application sandbox that is a web browser can be to many.

[fullstackchris]

Nitpick here: both models are still SaaS, the only difference is the first way was deployed via (desktop?) app and the second via web.

But indeed, web is typically the most flexible option unless you are leveraging something on the OS that would otherwise be cumbersom or impossible via web (not often the case)

[satvikpendem]

How is the first one SaaS? It's a perpetual (seemingly one time) license they said.

[sirjaz]

Unfortunately, a vast majority of WebApps are hot garbage and even the good ones can never be as functional or as performative as a native app. We have such powerful machines, but we relegate them to such a horrible method of using them.

[6510]

Very few components are needed to make a bare bones web browser that is more of a vm. It would need one or more advantages over normal www browsers. Applications could be memory, and computationally heavy, it could store its data locally with some guarantees, it could run conventional web application on very crappy hardware. A new platform also offers countless opportunities to do new things or do things differently. That list is endless.

[cyberax]

WebApps (even though I hate most of them) absolutely can be as fast as native apps. E.g.: Linear, Slack.

But it requires very careful engineering.

[neverartful]

In my experience they are not nearly as fast.

[cyberneticcook]

If I may ask where did he buy the app from? Some online marketplace ? I’d love to browse these kind of deals.

[lefstathiou]

Sorry for the late reply: he’s been in the maratime industry his whole career. Met the founder years ago, stay in touched, then made the ask.

[ensemblehq]

That's an awesome story. How did your brother find the opportunity to acquire the aging app?

[lefstathiou]

My brother was a consultant and is an extraordinary networker (and salesman) in the shipping space. He knew of this tool, was looking for a challenge, cultivated a relationship with the owner and then made an offer.

[charliebwrites]

How did he price out the SaaS model vs the USB perpetual license?

[lefstathiou]

Same price, only difference is you pay it annually instead of perpetually. I think it’s like 10-15k per license per year.