5 ms·

This issue is why I’m skeptical of SaaS in general. I look at software as a solution to a problem. When I have a problem and software solves it, I’m happy. The

This issue is why I’m skeptical of SaaS in general. I look at software as a solution to a problem. When I have a problem and software solves it, I’m happy. The better the software, the more thoroughly it solves my problem without creating new problems for me (by being annoying or inefficient). The best software seems to hit a sweet spot with the right set of features targeted at solving a particular problem or small class of problems with minimal fuss. Really good software becomes extremely difficult to improve on. But services can’t deal with just hitting the sweet spot and calling it a day. They need to grow, which means they need to evolve. The constant cash flow seems to burn a hole in the manager’s pocket. Eventually they cross a threshold and exit the sweet spot. Maybe they release a major version with an awful new UI. Like a TV series which has gone on too long, they jump the shark [1]. Perhaps this problem of SaaS may be analogous to network decay [2], though the latter is more gradual. [1] https://tvtropes.org/pmwiki/pmwiki.php/Main/JumpingTheShark https://tvtropes.org/pmwiki/pmwiki.php/Main/JumpingTheShark [2] https://tvtropes.org/pmwiki/pmwiki.php/Main/NetworkDecay https://tvtropes.org/pmwiki/pmwiki.php/Main/NetworkDecay

ajmurmann 6y ago
While I share the frustration with the constant need to fiddle with things, I am not sure if going a non-SaaS route is gonna solve this problem completely. Pay-once software at least has a clearer incentive to deliver improvements, so that you want to pay for the newer versions. So the need to grow might be even stronger. Ultimately old versions will also go out of support. Maybe OSS is the best bet, as long as the project is popular enough.
- judge2020 6y ago
  SaaS software on the other hand is more popular since the revenue stream is forcefully consistent no matter if you're putting out 10 features in a quarter or 1, or if the customer even wants to pay for whatever new feature the SaaS is putting out. With pay-once software, your revenue stream will fluctuate more depending on if you're putting out features everyone wants or not, which is great as a consumer (assuming your needs match with other customers) but could lead to huge financial issues (read: layoffs) if someone makes the wrong decision and develops features for 2 quarters in a row that no existing customer upgrades for.
atoav 6y ago
Exactly this. And if you have a problem, you don't want to go back to it every time the service changes fundamentally, they go out of business, get acquires by someone etc.
dvfjsdhgfv 6y ago
Note this is not specifically SaaS-related. In more traditional models, you also have new releases and upgrades, therefore you need to evolve in order to convince your existing customers the upgrade is worth the cost. Sometimes the new product is a failure (Vista, Windows 8), sometimes they see the benefit and will migrate on their own (Windows XP, 7). In any case, you are forced by the marked to endlessly add new features to stay competitive. In practice, many of these features just use up precious system resources and create problems. Even Free Software is vulnerable to this problem - Gnome is a good example. But fortunately in these cases people can take things into their own hands, and sometimes they do - and eventually we have such nice projects like Mate and Cinnamon.
TeMPOraL 6y ago
I frame my own disdain for SaaS along three axes: 1) Control - SaaS usually ingest my data and silo the results of my work, so that I don't own the output. This is often unacceptable. It also makes solving problems harder. With a local application, I can always hack around its deficiencies, or easily post-process its output. Can't do that with a webapp owned by someone else. 2) Reliability - While a SaaS will be more incentivized to support a paying subscriber, pay-once software with no auto-updates is more reliable once it solves your problem. That's because it's frozen in time. It doesn't get new features (which usually aren't needed if you have a working solution already), but it also doesn't break. In the context of your whole computing system, pinning to a version of a local piece of software lets you freeze one component of your system - which reduces complexity, and gives you one less thing to worry about. (Longer term you may need to worry about the rest of your system becoming incompatible with the version-pinned program - but VMs essentially solve that these days.) 3) Relationships - Every SaaS I subscribe to is another relationship I enter with a third party that needs to be consciously tracked. I don't want that, it's anxiety-inducing and cognitively exhausting in the long run. What I want is to pay money in exchange for solving problems. I don't want new relationships on top of that. I tend to use the following example: when I go out to buy some bread in the morning, I exchange money for bread, and that's it. I don't enter into an ongoing relationship with the store, the bakery, the flour supplier or the wheat farmer. I like my software like my bread. (To be more accurate: I actually do enter into a bunch of relationships when I buy food or trinkets in stores, but they're all governed and fully handled by various business and consumer protection laws - which means that I don't have to care about them. They only matter if something goes wrong with the goods I bought - and then all I need is to keep a proof of purchase and know who to e-mail about it, and I can even CC the government to force-handle things for me. Meanwhile, SaaS creates relationships you do have to track yourself.)
amelius 6y ago
> But services can’t deal with just hitting the sweet spot and calling it a day. The problem with the free market is that "sweet spot" is usually defined as what the majority wants. That's why I think we need open-source, so small companies can deal with the long tail.
open-source-ux 6y ago
"This issue is why I’m skeptical of SaaS in general" SaaS is eating the B2B (Busines to Business) software market. Companies small or large are much happier not having to manage servers, installation, updates and security. I don't particularly like SaaS either, but what is the alternative? It is impossible to self-install a web app on a server in a way that is easy and simple to do particularly for non-techical users. Until that happens (which may be never), SaaS will continue to grow and dominate the B2B market. Imagine if installing a web server app was as easy as installing a desktop app. It would unlock countless opportunities for developers. My impression is that developers think it's a non-issue or have simply never thought about it. (And no, before someone mentions it, Docker, Sandstorm, Cloudron, command-line installations etc are not easy-to-use options for non-technical users.)
- yandie 6y ago
  You didn't mention other things security, scaling, keeping the service up, backing up. Securing a web application is HARD.
  - jfengel 6y ago
    Very hard. Terrifying to me as a professional developer. I can't keep up with all of the possible failure paths. The more of that I can foist off on somebody who does it full time, the better. Then I can focus on my application domain rather than the basic mechanics of keeping the thing up and secure.
- fomine3 6y ago
  For installation, install apps feature on NAS product like Synology (that internally uses Docker) is easier. Managing, keep updating, HA, etc is different story.
xtiansimon 6y ago
So much this. It happened to me when 1Password iPhone App changed and no longer required you to enter your master password with every phone restart. And that same month I made a minor tweak to my password (because I was using it so much for a new computer) and got locked out. Other zero-value improvement was with Transit iPhone App. They made cosmetic GUI changes which make no sense (cover 50% of the map, swipe interaction target too big, transit line notifications screen gone) and they made zero improvements to transfer scheduling. Weather Underground iPhone App took a compact and information rich GUI and turned it into a garish nightmare with 'speedometer' temperature graphic. I've said it here again and again--I don't understand the point of ruining a good iPhone App. Why not treat it like a video game--just make a new one and support the old one with security updates? Heck. I get it. Double the code base, without the _perceived_ possibility of a revenue boost from the _new and improved_. At the very least make improvements very carefully (1Password), and if you have a hit don't redesign. Maybe --- make the old version the subscription and the new software ad-based? Maybe then you'd at least have a real measure of _new and improved_.

ajmurmann 6y ago

While I share the frustration with the constant need to fiddle with things, I am not sure if going a non-SaaS route is gonna solve this problem completely. Pay-once software at least has a clearer incentive to deliver improvements, so that you want to pay for the newer versions. So the need to grow might be even stronger. Ultimately old versions will also go out of support. Maybe OSS is the best bet, as long as the project is popular enough.

judge2020 6y ago
SaaS software on the other hand is more popular since the revenue stream is forcefully consistent no matter if you're putting out 10 features in a quarter or 1, or if the customer even wants to pay for whatever new feature the SaaS is putting out. With pay-once software, your revenue stream will fluctuate more depending on if you're putting out features everyone wants or not, which is great as a consumer (assuming your needs match with other customers) but could lead to huge financial issues (read: layoffs) if someone makes the wrong decision and develops features for 2 quarters in a row that no existing customer upgrades for.

atoav 6y ago

Exactly this. And if you have a problem, you don't want to go back to it every time the service changes fundamentally, they go out of business, get acquires by someone etc.

dvfjsdhgfv 6y ago

Note this is not specifically SaaS-related. In more traditional models, you also have new releases and upgrades, therefore you need to evolve in order to convince your existing customers the upgrade is worth the cost. Sometimes the new product is a failure (Vista, Windows 8), sometimes they see the benefit and will migrate on their own (Windows XP, 7). In any case, you are forced by the marked to endlessly add new features to stay competitive. In practice, many of these features just use up precious system resources and create problems. Even Free Software is vulnerable to this problem - Gnome is a good example. But fortunately in these cases people can take things into their own hands, and sometimes they do - and eventually we have such nice projects like Mate and Cinnamon.

TeMPOraL 6y ago

I frame my own disdain for SaaS along three axes: 1) Control - SaaS usually ingest my data and silo the results of my work, so that I don't own the output. This is often unacceptable. It also makes solving problems harder. With a local application, I can always hack around its deficiencies, or easily post-process its output. Can't do that with a webapp owned by someone else. 2) Reliability - While a SaaS will be more incentivized to support a paying subscriber, pay-once software with no auto-updates is more reliable once it solves your problem. That's because it's frozen in time. It doesn't get new features (which usually aren't needed if you have a working solution already), but it also doesn't break. In the context of your whole computing system, pinning to a version of a local piece of software lets you freeze one component of your system - which reduces complexity, and gives you one less thing to worry about. (Longer term you may need to worry about the rest of your system becoming incompatible with the version-pinned program - but VMs essentially solve that these days.) 3) Relationships - Every SaaS I subscribe to is another relationship I enter with a third party that needs to be consciously tracked. I don't want that, it's anxiety-inducing and cognitively exhausting in the long run. What I want is to pay money in exchange for solving problems. I don't want new relationships on top of that. I tend to use the following example: when I go out to buy some bread in the morning, I exchange money for bread, and that's it. I don't enter into an ongoing relationship with the store, the bakery, the flour supplier or the wheat farmer. I like my software like my bread. (To be more accurate: I actually do enter into a bunch of relationships when I buy food or trinkets in stores, but they're all governed and fully handled by various business and consumer protection laws - which means that I don't have to care about them. They only matter if something goes wrong with the goods I bought - and then all I need is to keep a proof of purchase and know who to e-mail about it, and I can even CC the government to force-handle things for me. Meanwhile, SaaS creates relationships you do have to track yourself.)

amelius 6y ago

> But services can’t deal with just hitting the sweet spot and calling it a day. The problem with the free market is that "sweet spot" is usually defined as what the majority wants. That's why I think we need open-source, so small companies can deal with the long tail.

open-source-ux 6y ago

"This issue is why I’m skeptical of SaaS in general" SaaS is eating the B2B (Busines to Business) software market. Companies small or large are much happier not having to manage servers, installation, updates and security. I don't particularly like SaaS either, but what is the alternative? It is impossible to self-install a web app on a server in a way that is easy and simple to do particularly for non-techical users. Until that happens (which may be never), SaaS will continue to grow and dominate the B2B market. Imagine if installing a web server app was as easy as installing a desktop app. It would unlock countless opportunities for developers. My impression is that developers think it's a non-issue or have simply never thought about it. (And no, before someone mentions it, Docker, Sandstorm, Cloudron, command-line installations etc are not easy-to-use options for non-technical users.)

yandie 6y ago
You didn't mention other things security, scaling, keeping the service up, backing up. Securing a web application is HARD.
- jfengel 6y ago
  Very hard. Terrifying to me as a professional developer. I can't keep up with all of the possible failure paths. The more of that I can foist off on somebody who does it full time, the better. Then I can focus on my application domain rather than the basic mechanics of keeping the thing up and secure.
fomine3 6y ago
For installation, install apps feature on NAS product like Synology (that internally uses Docker) is easier. Managing, keep updating, HA, etc is different story.

xtiansimon 6y ago

So much this. It happened to me when 1Password iPhone App changed and no longer required you to enter your master password with every phone restart. And that same month I made a minor tweak to my password (because I was using it so much for a new computer) and got locked out. Other zero-value improvement was with Transit iPhone App. They made cosmetic GUI changes which make no sense (cover 50% of the map, swipe interaction target too big, transit line notifications screen gone) and they made zero improvements to transfer scheduling. Weather Underground iPhone App took a compact and information rich GUI and turned it into a garish nightmare with 'speedometer' temperature graphic. I've said it here again and again--I don't understand the point of ruining a good iPhone App. Why not treat it like a video game--just make a new one and support the old one with security updates? Heck. I get it. Double the code base, without the _perceived_ possibility of a revenue boost from the _new and improved_. At the very least make improvements very carefully (1Password), and if you have a hit don't redesign. Maybe --- make the old version the subscription and the new software ad-based? Maybe then you'd at least have a real measure of _new and improved_.