Cloud Software Begins to Lose Altitude | Hacker News Reader

Cloud Software Begins to Lose Altitude

bloomberg.com

88 pointsT-A25 comments

I think it's more that all types of software tech companies have been overvalued, rather than a specific cloud-based company problem. If it's a pure software product that works in a web browser, like marketo and zenefits seem to be, then you could stop innovation and maintain it with minimal development for a good long time.

Anecdotally, my father has been doing tech support for a product that hasn't seen any new features in over 10 years, but customers keep paying for it. His team is only 5 support people and a part-time developer loaned from another department who fixes an occasional bug. It makes many millions of dollars every year, and my dad doesn't even make $60k, so you can do the math.

If a company wants to sell-out to PE and stop innovating, good for them. That's the nature of business, and it opens up the market for a more innovative company.

Just to clarify, Zenefits is an incredibly manual operation. The client facing side suggests things are automated, but there are brokers and automated scrapers and fax machines to make that happen.

(I was an early ZenPayroll employee)

And an autohotkey script. Can't forget the autohotkey script.

Isn't it unfair to think of Saas as a classification of software a la "word processing software" and more as a business model?

Like, nobody sets out to build a page that will host ads. One builds the product and then later settles on a way to get revenue. I've always thought of Saas like this, just another potential business model to follow.

I think the classification has less to do with the technology itself, but rather boils down to the fact that enterprise software is hard to switch internally as an employer because of red tape. If I'm a 50,000-employee company and have used BigEnterprise Inc's non-cloud HR software for 20+ years, it's really really hard to switch to a new solution for something like payroll or HR. Also if the old software keeps working, and everyone is used to it, why switch?

A lot of large companies however are so fed up with on-premise enterprise that they are willing to make the switch. That said, new cloud-based companies aren't making nearly the same profits that BigEnterprise Inc is on their old systems. SaaS has thus become a keyword for these smaller cloud-based "disruptors", even if it is technically possible for BigEnterprise Inc to move to the cloud.

> A lot of large companies however are so fed up with on-premise enterprise

Is this anecdotal or is there some number behind this? When I was developing internal tools for massive companies, cloud was never an option. Red tape is definitely a given with large companies, but some of the red tape were a necessity to meet ISO certification and other regulatory needs.

Well it's usually not an option, but if they pay enough money for it it and have an SLA then it might still qualify as enterprise grade.

Many Enterprise companies have cloud options, and they still aren't viable because leadership is risk adverse to having their data outside of their ownership. This is especially true regarding laws in various jurisdictions.

I work in Enterprise Software and speak to 100s of C level leadership. People aren't "fed up" with on-premise enterprise. They aren't refusing to switch due to it being difficult.

I mean, this is an interesting conversation we could take in a lot of directions, but I wanted to clarify that I think there's a fundamental misunderstanding here about enterprise software.

Same is true of SaaS, it's not a keyword for these smaller cloud-based software. It's a literal description of your ownership of the service.

You are right that there is less margin in cloud vs. on-premise, but that's true of BigEnterprise's software and offerings as well.

What you're saying fundamentally makes sense as a theory, but in practice I think the reasons and thoughts just have different outcomes.

And this isn't just software, this is also for hosting the hardware as a whole.

There is a huge amount of red tape to any SaaS to store PII.

Huge.

I can imagine it's even more so for things like financials.

> People aren't "fed up" with on-premise enterprise.

Do you happen to know if the 'no one gets fired for buying IBM' attitude has changed significantly in big enterprise markets ?

UrbanCode is still absolutely huge and IBM will be serving that market segment for some time.

IT departments can be a lot like cloud providers anyway. The big enterprise HR system may well not be controlled by the actual HR department. It will be a shadowy service at the end of a network cable with a "helpdesk" for support. It may be good or bad but that actual user has little control either way. Vendor lock in at its worse!

From an Enterprise perspective (and I'm in the belly of the fortune 50 beast at a division level ~10bn/yr mfg) SaaS discussions internally aren't really about data security -- they're about lock in and leverage by the vendor. For anything of a certain level of criticality SaaS is just too much vendor leverage. Enterprise software licensing is a vicious lawyer on lawyer battleground with deeply unethical behavior and brinksmanship regularly displayed on both sides. If you think I want to give any of those vendors the possession of my server and data during a dispute, you're nuts. The enterprise SaaS model seems fundamentally flawed.

On the other hand some users/departments within enterprises see SaaS as a way to escape the IT department lock-in.

The challenge is to provide more value than the alternative. Enterprise SaaS works where it makes sense, and it doesn't where it doesn't.

On its tomb will be written : The over-promising Cloud has been an under-delivering clown.

A quite vulnerable OS (by lack of ease of maintaining the state) running in a quite hackyly (thus porous) secured envelope is no more cheaper nor efficient than correctly written software running in a diminished context (jails, chroot, pledge)... coders that write correct software are expensive and cannot be compensated by more CPU, RAM, bandwidth.

If common sense does not work, and financial sense cannot convince you ... I hope some will get your senses back before their customers change their mind.

PaaS and SaaS are here to stay. Whole industries are moving systems into SaaS solutions - the overhead of running data centers, OSes plus dealing with software upgrades is just too much.

What is happening right now that the market is shaking out the winners and the losers - Marketo has failed to keep track, Adobe and SFDC MarketingCloud (Exacttarget) have eaten their lunch.

Getting acquired is not a goal, it is failure. The good SaaS ones are standing on their own. ServiceNow, SFDC, etc.

I think this article tries to justify cloud from a financial perspective, where margins have shrunk obviously . But the volumes will keep increasing for quite some time. The value from cloud services is just increasing and The Cloud movie has just started.

I believe cloud adoption will also jump at the next economic recession as companies try to cut costs.

"Think inside the box."

If anything, self-hosted cloud software will continue to grow in popularity. It's cloud-based Saas companies that are more hype than substance (read Disrupted for an unsavory inside look).

Self-hosted cloud software will continue to grow in popularity though. Shameless plug alert: I recently released Wheatbin which is Open Source and 100% free: http://wheatbin.com. Github repo is here: https://github.com/wheatbin/wheatbin

Duane history [1] is worth reading

[1] https://github.com/wheatbin/wheatbin/blob/master/doc/seed-of...

Weird that you show a BlackBerry on the product page. It makes perceiving the software as very old.

Good point - this is exactly the market Sandstorm (https://sandstorm.io) is trying to reach as well. Will be interesting to see how cloud-within-the-enterprise will turn out.

Yeah, I was just thinking this - reading around it a little bit, it seems that it's the cloud aspect that's where the hesitancy is for a lot of players (correct me if I'm wrong). If BigCorp is always going to have some on-site IT/server resource, then surely the best middle ground is renting out the software, letting BigCorp store the data on their own servers, and then charging for training, backup and integrations?

Reply on news.ycombinator.com