Recently updated on March 7th, 2023 at 12:57 pm

I Will Never EVER Use Cloud Software. Now Get Off My Lawn!
Introduction
There have been people worrying about and bashing the Cloud since the very beginnings of the Cloud. Despite that, Cloud has been with us for a very long time, and it’s done nothing but get inexorably bigger. It has gotten into more application categories than ever, and every one it gets into, it’s market share steadily goes up. The mainstream public doesn’t have a problem with it. They don’t care about the Religious Wars, they just want better software.
Now the Cloud has come to the Industrial Software world, and more specifically, to the CADCAM world. You can see the great hue and cry in some of the comments about OnShape’s launch of F360. Like our friend the death stare cat on the right, some will proclaim they will never Ever use Cloud Software.
Frankly, Cloud’s arrival is late relative to other markets for a whole host of reasons. If nothing else, it is just a lot of work to write a competitive CAD or CAM program from scratch whether it is Cloud or not. The Manufacturing World has not been incredibly progressive about lots of things for lots of reasons. For example pay ridiculous sums for more memory or Ethernet connections on CNC Controls that are basically just PC’s down under. Let’s be honest too, we like to think we’re different and that’s reason Cloud isn’t as far along here. We like to think the rules we live by are different. We like to think our needs are different. We like to think it doesn’t matter what’s happened in all of those other markets because after all, We’re Different!
Folks, it just isn’t true. I’ve been doing Cloud software for probably half my 30 year plus career. I’ve been blogging about it over on my Smoothspan Blog since 2007–it was literally the first real blog post I did there. It’s true that Manufacturing is late to the Cloud party. But in terms of anything being new? Not so much. I’ve written about objections to Cloud for other markets many times, and after compiling this list of objections to Cloud, I don’t see much of anything new here. These objections have been dealt with for everything from mission critical accounting software to what am I going to do if my precious family photos are lost in the Cloud. These same objections were raised and overcome in all the other markets. It doesn’t take very long to get past the objection stage. In five years we’ll wonder what all the fuss was about. Meanwhile, I’ll try to guide you through how to think about the objections.
Don’t get me wrong, I’m not some blind Cloud Fan Boy. I understand and love desktop software very much too. My first company was acquired by Borland and built the Quattro Pro the spreadsheet. It made over $100M in sales during its first year. Those notebook tabs you see at the bottom of Excel were invented by yours truly and my name is on that patent. Yes, I was a General in the Office Wars. I get Desktop with a capital “D” and Cloud with a capital “C” equally well. Knowing all of that, when I set out to start CNCCookbook, I knew we had to create Cloud software for CNC for all the right reasons.
In this article I want to debunk a number of myths about why the Cloud is bad for Manufacturing Software. I will also be speaking of Best Practices to look for from any prospective Cloud Vendors to help insulate you from potential Cloud problems. It isn’t as if I’m saying there are no problems with the Cloud. It has problems just like the Desktop and I’ll lay them out in black and white below. I simply believe its advantages far outweigh those problems and that if you follow the Best Practices I will recommend, you can maximize those advantages while minimizing drawbacks.
So let’s get on with the debunking:
1. Cloud is just the latest scam Greedy Software Developers have hatched to overcharge us.
I hear this a lot–Cloud is just a scam to make the overpriced software even more expensive. I will never get done paying for it!
Funny thing about that. I just switched all of my Adobe products (PhotoShop, Dreamweaver, etc.) over to their Cloud Subscription program. It was a total bargain for me. I get all of their products and all of the updates for an annual subscription cost that is right about what I’d pay to upgrade just Photoshop once. I also signed up recently for Fusion360 precisely because it seemed like such a bargain even if all I used from it was HSMWorks.
The big lie is thinking that just because you paid a one-time cost for one version of software that you aren’t renting it. That bubble bursts as soon as you have to buy a new version of the software and suddenly you’re being asked to get caught up on maintenance. It might as well all be rental when you consider that stuff. Instead of a stream of relatively small payments at regular intervals there is a much lumpier stream of bigger payments at odder intervals.
Why isn’t that all built into one cost so it’s easy for me to understand what’s going on? Come to think of it, why do so many of these Desktop vendors make me talk to some sales person or “Value Added Reseller” sales person before I can even find out what it costs? What the heck are they trying to hide and how can that ever be a good thing for me? Cloud vendors are almost always totally up front with their pricing.
Don’t kid yourself–you’re going to keep paying either way, Cloud or not. If you think you can sit on one version forever, read on. I’ll address that myth shortly.
SolidSmack recently wrote that OnShape has “Changed MCAD pricing forever.” They refer to a Cloud offering changing pricing against Desktop offerings and they mean that the Cloud is lowering prices, not raising them. This has happened in every single market I’m aware of that went to the Cloud. The Cloud is cheaper when you really dive down into the nuts and bolts of it.
The only thing that keeps pricing reasonable is competition. Guess what, all the new kids on the block who are trying to compete with overpriced incumbents are in the Cloud. It’s true in all the other markets too. This is a tested model for competition. Embrace it. The sooner the incumbents feel real competition, the sooner prices will go down for everyone.
Best Practices: Avoid vendors who make you go through salespeople just to find out what their price is. There is no way they’re doing you a favor with that. What are they afraid of, anyway? Avoid annual maintenance fees. An upgrade is an upgrade. Why do you need to sneak another cost in under another label just to make it seem cheaper? At CNCCookbook, you get everything built into the subscription price so you know exactly what it all costs and you never have to talk to a salesperson about anything.
2. I don’t rent most things in my life, and you shouldn’t be willing to rent software either.
Nearly everything I have in my life that’s digital IS lease based already. My Internet, Cable TV, and much of my software, for example. While we offer our current CNCCookbook software both ways (annual subscription out one time payment for life), I personally like the leasing model for the digital world. Shops gotta have space for machines. Did you buy the building and land? Well then you probably have a mortgage. Otherwise you leased it. The exception is the unlikely event where you are so flush you paid cash to own it all outright. Do your machines require regular service and attention? That’s not free. There are consumables. How is that different than having a monthly fee on software.
Look at how the subscription model is priced. In most of the rest of the software world, the annual subscription cost is pegged at 1/4 to 1/3 the cost of buying the permanent version of the software (or a competitor). Hence, you are ahead cash flow-wise for at least 3 to 4 years. This doesn’t even consider maintenance or upgrade fees. For most software I use regularly, I would have to pay for at least one upgrade by the end of 4 years. Software that’s priced according to this 1/4 to 1/3 model is actually quite cheap. My business is growing, so I want to shift my expenses to the future and not pay them up front. I want to be ahead on cash flow. And if it was shrinking or fallen on hard times, I would want to be cash conservative even more.
Look at it another way. When you buy something expensive, don’t you want the seller to thoroughly prove to you that you’re getting your money’s worth before they get all of their money? That’s exactly what Cloud pricing does. You’re able to use the product for 3 to 4 years before you pay as much as the up front model. The Cloud vendor has to re-earn your business with every payment. You have them over the barrel, not the other way around. You can cancel any time you want and you’re done. Try using one of the big CADCAM packages for a year or two and then trying to get any money back at all if it didn’t work out.
Best Practice: Choose a vendor whose subscription pricing amounts to an annual fee that is 1/4 to 1/3 of what it would cost to buy an equivalent competitor outright.
3. I don’t need any improvements or new versions. I just want to pay once and use that version and be left alone.
The problem is software doesn’t sit still. It has a lot more in common with cable TV than it does with my milling machine. The mill doesn’t get 3x faster, sprout the ability to use R8, CAT40, and HSK tools, or magically grow a 4th and 5th axis just sitting there. But my software does, and that’s valuable to me.
Even programs where I think I just want the same version forever without any new features, it turns out I don’t. I encounter a bug and want the update where the bug is fixed. Or Microsoft ships a new operating system and my old stuff doesn’t work as well (or at all). I want it to be compatible with some other program’s file format that just changed with their last version. I got a new machine and want a different post for my CAM, but that post wasn’t even available when that version of the CAM came out.
I recently talked to a Shop Owner who hadn’t been upgrading his CAM license. It was great for a couple of years, but now he now has to redraw any CAD his customers bring in because the CAM is old enough it can no longer import the later CAD file formats. This has been a terrible thing for him, but in the meanwhile, the cost to get back to a current version has steadily piled up. Now he can’t afford the huge lump sum cost of getting up to date. What a pity he didn’t amortize that cost in small subscription payments so he always was up to date. Your CAD and CAM are mission critical Digital Tooling for CNC.
I had the experience myself. Had a CNC utility that I needed the latest version of. This was all triggered because I couldn’t install the old version on my new Windows 7 computer, it only ran on XP. I had to pay maintenance for all the years since I had stopped paying plus the upgrade price for the latest version before I could get it. I was only 2 annual versions behind, but when all was said and done it cost me only $20 less than the original $400 purchase to get back up to date. Excuse me, but I think maintenance and new version costs are the real scam here. I don’t see the advantage of dealing with those costs versus a straightforward annual or monthly subscription.
You are going to have to deal with new versions for any software you use very often. I know, I’ve tried to sit on a version of some software in the past. If it’s software I use very often, it has been a mistake every time. If nothing else, just trying to figure out how to operate the newest version of the software when I finally did upgrade was way more painful than just keeping up would have been. It’s easier to deal with little changes along the way than the big Tectonic shifts that feel more like terrible earthquakes.
Here’s another reason to deal with new versions: fewer bugs.
Frequently, users are afraid of new versions because they might bring new bugs. There are two important realities that argue against that.
First, in a former life, I surveyed a large number of Business Software companies that sold On Premises software. That would be the software you buy, own, and install on your own computers. I visited their VP’s of Customer Service and asked them just one simple question. Out of all the open service tickets you have, how many of them would be fixed if your customers were running the most current version of your software? The answer was astounding to me. It ranged from 40-70%. Let’s take a moment to understand what that means. For 40-70% of bugs, if customers had the latest version of the software, the bugs they’re reporting would have been found and fixed before they ever had a chance to discover the bug. That’s huge, and it is a big part of the reason CNCCookbook wants everyone on the latest version of our software. It’s hard to beat having bugs fixed before you even get a chance to see them.
The second part of that involves Boil-The-Ocean software releases. One reason many are gun shy about new releases is that the vendors don’t release very often. An annual release cycle means there are a huge number of changes that have accumulated in the software over the course of a year. If they spent too much time writing new features and not enough on testing, fixing, and stabilizing, you may be handed a real hairball of a release. With Cloud, that just doesn’t happen. The rule is many smaller releases. Very little of the product is touched in any given release. It’s much easier to thoroughly test such a release and there is also a better opportunity for anything that escapes testing to be discovered and fixed before you need to access that part of the software. It’s just a better way to operate a software business.
Imagine a Job Shop that only ever inspects parts that are finished, even parts with many tight tolerances. In fact, they don’t inspect until every single part the customer ordered is made. That’s the old way of building software. The non-Cloud way. The potential for expensive scrap is much higher with the Old Way than it is when inspecting as you go. It’s exactly the same for software.
Best Practice: Look for a vendor that does updates frequently so you get the benefit. Don’t wait a year, or worse many years to upgrade. As an added bonus, consider vendors that include both mandatory and optional updates. CNCCookbook does this. We try to limit mandatory updates to once a quarter or longer. The optional updates tell you immediately when they are ready, but you don’t have to take them. We do the optional updates as often as required. And, I almost never designate the latest release as mandatory unless it is required for an infrastructure update or a really bad bug. Mostly the former. Instead, I designate releases from several months back that were tried and true with few or no reported issues. If you’ve been taking the optional updates, you may not even see the mandatory message pop up. But even if you do, you’ll know the mandatory release is solid and well tested.
And if you’re running desktop software, keep up with the current version. I know it seems expensive, but it’ll be a false economy if you get too far behind because you’ll wind up paying anyway. If your view is I’d rather pay it later than now, you’re totally speaking the Cloud language.
4. The Cloud doesn’t add any value.
We often talk about the Cloud as if it is only a matter of running the identical desktop software in the Cloud or on the desktop. That plus one has a subscription model and the other I buy outright. But that just scratches the surface of what the Cloud can do. Of course we’ve talked about the value of always having the latest version, but there’s a lot more the Cloud brings to the table as well.
Products that need to be collaborative or that share data across multiple people or devices pretty much have to be in the Cloud–that’s how the sharing will happen.
Products that need a lot of CPU horsepower (photo realistic rendering or complex mathematical simulations, anyone) benefit from having tons of cpu’s available, but your need of those cpu’s is not continuous. You’re not going to roll in a bunch of servers and then take them back to the store when you’re done with your rendering or simulation. It’s too painful and cumbersome. But, with the Cloud, it’s easy to fire up tons of cpu’s for a short time and then shut them off when you’re done, and it doesn’t cost much.
Then there’s the issue of running on a lot of devices. Like it or not, there is more to life than just the CAD Workstation. More on this in a minute, but the Cloud is the right way to do this.
Best Practice: Look into what Cloud-Specific your vendor offers. If there are none, look for another vendor because there should be many.
5. My Data isn’t safe in the Cloud from hackers, cyber criminals, and IP thieves.
This old chestnut about malicious security risk has been around for Cloud since the beginning of times, and it is totally bogus. Here’s why:
What you’re doing is making a bet that you as an individual or your shop’s IT department (you do have an IT department, don’t you?) can do a better job on security than a good Cloud Vendor. I have people essentially making this claim when they tell me they’re worried for example about the security of having a machine even be connected to the Internet or when they’re talking about state funded Cyber-Terrorism from places like China or North Korea. The complacency that you can do that job part time better than full time professionals is a mistake. Doing security really well is something that only the big players can do.
Think you’re okay just not connecting to the Internet? It’s not going to happen. The most common way of exchanging CAD data today is via email–70% of the time the data is exchanged via email, or in other words, via the Cloud. What’s that you say? You never connect your CAD machines to the Internet? Really? And nobody ever accesses them with a USB key or takes a drawing home to work on it after work because the job is late? Well, guess what? That USB key can bring the viruses in just as much as Cloud Software and you’re even less likely to be protecting against it than with the USB key.
Remember that little thing called Stuxnet that zapped Iran’s centrifuges from their nuclear program? None of those computers controlling those machines were on the Internet. They got infected because one person used a USB key.
Still think you’re going to hold of Chinese Cyber Terrorist attacks all by yourself?
Consider that Amazon’s Cloud Service is certified and audited by the US Department of Defense. I don’t know about you, but my computers and security practices have never been audited by the DoD. Wait, that’s actually not true. All Cloud data for CNCCookbook’s software lives on Amazon which is DoD certified. In fact, the only way we could’ve afforded that certification is to get it indirectly from a vendor like Amazon.
Best Practice: A good Cloud provider can tell you which of the many 3rd party audits they’ve been certified for. There are many ranging from the DoD certifications I mention to SAS 70 to HIPAA (required for sensitive medical documents) to credit card PCI compliance (so sensitive credit card data is protected from fraud). Amazon’s Cloud has all of those certifications and more, BTW.
6. My data isn’t safe in the Cloud due to everything from natural disasters to the Cloud vendor screwing up.
The answer to this one is pretty similar to protection from Hackers. Are Cloud Vendors immune from natural disasters and internal personnel screw ups? No, of course not. But they’re a better bet than you are for avoiding these problems.
Let’s take the vendor screw up issue. It has happened very rarely. It’s very unlikely to affect every customer and all the data. Most of the same issues that come up passing the security audits require that controls be in place. Moreover, these guys are very scared of such screw ups. This is their entire business that’s at stake.
What about the natural disasters?
Here the Cloud guys have a huge advantage. It costs a fortune to put redundant data centers in more than one location. Individuals seldom do it. Small businesses? Well maybe someone is supposed to take home a backup disk every night to get it off site, but did they? Is it up to date? Will it even restore if called on? Did you ever test restoring it?
One of the biggest Cloud vendors out there is a company called Salesforce.com. They were at almost $1 Billion in revenue before they could afford duplicate data centers. Guess what? CNCCookbook has multiple redundant data centers all over the world and we’re nowhere near Salesforce’s $1B in revenue (I wish!). We did it because we use Amazon Web Services and the redundancy is an automatic part of their service.
Folks, your data is much safer in the Cloud than it’s likely to be on your own machine.
Best Practice: Find out about multiple datacenters and their other data redundancy measures from your Cloud vendor and make sure they have good answers.
7. The minute my data goes into the Cloud, I have lost control of it
This will only happen if you let it. The only way you will let it happen is by not following the recommended Best Practices for data in the Cloud:
Best Practice: Insist on being able to get your data back out of the crowd easily. Make a practice of grabbing regular snapshots of your data. Insist on being able get that data in popular formats used by other software. Getting an impossible to decipher proprietary format is going to make your life tough in the event of a crash and burn. Extra credit for vendors that make this super easy to do for all of your data as often as you want.
8. I lose everything if the Cloud Vendor goes out of business
This one is particularly worrisome to many, but there are two realities that strongly mitigate this concern.
First, vendors using subscription based pricing are far less likely to go out of business than vendors who let you pay once and then never again. The statistics on the likelihood a SaaS/Cloud business will go out of business after it reaches critical mass are wonderful. As a result, Cloud businesses cost far more to acquire than businesses that sell perpetual license software. In fact, the advantages for Cloud are so great that very few investors will even put up money for a non-Cloud business these days.
Second, you’re in trouble regardless of whether it is a Cloud Vendor or a Desktop Vendor that goes out of business. If the software is mission critical to your business, you need to start looking for a replacement the day you hear the bad news. Every business you deal with that might be using the same software (and its file formats) is doing the same thing. The clock is running and eventually the music will stop. Software companies are like sharks. If they quit growing they don’t live very long.
Best Practice: Make sure you have control of your data, like we talked about in #7. If there is a problem, worst case is you hear about it the day the vendor went out of business and you have to purchase an alternate that can load your standard formats that same day. Best case is you get plenty of warning that things are not well and you can plan an orderly transition. See my comment about the value of a Cloud business. You are MUCH more likely to read they got acquired than that they went out of business.
Extra Super Best Practice: I worked with a Cloud Vendor that did go out of business one time. It needed to raise more capital immediately after the 2008 crash. Very bad timing as no one was able to raise money for anything at that time. Despite that, there are still customers of that company many years later that are still able to use the software. How can that possibly happen? When big companies sign big deals for software they’re purchasing outright, they frequently ask for a clause in the contract called “Source Code Escrow.” Essentially, if the Software Vendor goes out of business, the customer gets a copy of the Source Code so they can soldier on. This clause is actually pretty useless, because without the developers, you can’t do much with the Source Code. At the company I was with, we had a better idea for how to make something similar with the Cloud so we could guarantee our customers could carry on without us if need be. We ran everything on Amazon Web Services. So our customers were given the opportunity to receive a private running version of the service on Amazon. It was easy. We set it up for them, trained them on how to manage it, and handed them the keys to their kingdom. I haven’t seen another Cloud company promise this since, but it’s actually very doable for those that use 3rd party Cloud Data Centers like Amazon.
9. I have to be in control of my Application
Control over Version, Plugins, and Customizability are all lost with Cloud Software or so the saying goes.
The reality is that you should give up control over the Version for reasons we’ve already talked about. It may not be all Versions (recall CNCCookbook’s Mandatory and Optional upgrades), but by and large you need to be on a pretty current version.
Let’s talk about Plugins and Customizations. There is absolutely nothing about the Cloud that prevents Plugins and Customizations. Their availability is entirely a function of what the vendor envisions or wants to offer. They realize they’re not competitive unless they can develop these functions and it isn’t that hard to do.
Best Practice: Pick software that makes it easy to keep up with the latest versions and keep up. Pick software that supports add-ons/plugins or customization in some way. Look for signs of API’s or partner pages and you’ll find evidence they’re on board with this. Great Cloud API’s have created some of the most vibrant add-on ecosystems ever.
10. It’s Not Cloud Unless it Runs In a Browser
This is a bit of a Religious War that rages inside the Cloud world, from one vendor to another. A great compare and contrast would be Fusion 360 vs OnShape.
F360 has a locally installed client on your desktop. OnShape runs in your browser. Both are Cloud. Both have advantages and disadvantages:
– It’s easier to always be up to date with browser-based software. They can slipstream little changes in every time you open the app. With a desktop UI, it takes more effort. Even so, F360 manages to stay up to date within say a few days versus the few hours of a pure browser app.
– A desktop client can tolerate a worse Internet connection than a pure browser app. Perhaps it can even work for a time with no connection.
– Browser-based software will run on more devices than desktop clients. This one is more of a likelihood than a certainty. True, it is easier to support lots of devices in a browser. But it is possible to support lots with an installed client too, it simply takes more work. F360 works with PC, Mac, Linux, and Mobile devices, just like OnShape. CNCCookbook’s software uses a desktop client and runs on PC and Mac, but not mobile devices.
Best Practice: Think about the quality of your Internet connectivity and the different devices it would be nice to run on. Choose your solution accordingly.
11. I don’t need access to CAD anywhere but my CAD workstation. Other devices are irrelevant.
So many devices, so little time and money. It’s hard to argue against accessing your software on more devices in this day and age. Devices are everywhere and there are good use cases for wanting access on these devices. Personally, I think the device makers present the biggest obstacles. It really is easier to do with browser-based apps. Not only do the device makers stringently control which tools you can use to build your software, but they exact a cost of 30% of your revenue for apps on devices. At CNCCookbook, we’ve had to decide against devices not because users don’t want us there, but because the device makers have just made it too hard and too expensive to be worth it. Big companies like Autodesk can afford it and companies that built browser apps can ignore it because (so far) the device makers have less control over what appears in the browser.
Best Practice: This area is fluid. Decide how badly you need mobile support. Not having it is unlikely to be a deal breaker, but having it is a definite sweetener for most.
12. I have to be able to run Offline in case I don’t have good Internet connectivity
– We don’t have a good connection because we’re out in the country.
– We don’t want a good connection for security reasons.
These are the two arguments I hear about the need to run Offline. I don’t see either one as an absolute unchanging requirement for Offline capability. We’ve already addressed the security issue, so lets not flog that horse again. Not having a decent connection is more interesting to address. At CNCCookbook, we allow our existing apps to disconnect for up to 28 days before they require a connection again. The idea is that even a cheap $100 laptop bought from Craigslist can run the apps and be brought to a connection every now and again. But what about an app that just can’t afford to be disconnected for 28 days?
I have two options for you on that: Satellite ala DirecTV for Internet and the WiFi Hot Spots that your phone or tablet can create. I’ve used both and here are my thoughts.
DirecTV has tremendous bandwidth (down at least, up is slow and must be via land line) but lousy latency. Depending on your app, it might work fine. A desktop app is likely to work a lot better than a browser app just because the lousy latency is going to kill any browser app that has to make a round trip to the Cloud for any user interaction. I like the WiFi Hot Spot option a lot better, and I have used it many times. We have a DSL line at home, and sometimes it acts up. I have frequently fired up my iPhone’s Hot Spot and gotten better performance than on the balky DSL. In fact, performance there has been on par with the DSL when it is behaving well. If your mobile device can get a decent Internet connection, you are set.
If all that fails, well, I’m sorry to hear it. You’re not going to like this Cloud software and will have to look elsewhere. Fortunately, the vast majority of people are not in this situation.
Let’s do talk about a third reason for poor connectivity:
– My Shop won’t allow an Internet connection because they’re afraid of security and afraid we’ll use it to slack off too much.
We talked about security. As for the slacking off, yes, there are many businesses who’ve enshrined the management concept pioneered by the English Navy: The beatings will continue until morale improves. And lest you be confused, they mean Management’s morale, not yours!
Even so, there are no end of ways to limit Internet access to particular web sites, web pages, and even ports for the API’s. This is not a problem that must be. It’s a problem some businesses create for themselves that they can also choose to undo.
Best Practice: Try for satellite or WiFi Hot Spot connectivity. Have IT limit access to just the Cloud services the business needs, if the lack of connectivity is self-imposed.
13. Cloud can’t possibly be good for performance or usability
This is another fine old chestnut that’s been around since way before it was taken up for Manufacturing Software. It’s been disproven time and time again. If nothing else, every single time you participate in a teleconference via Skype, Google Hangouts, or whatever, you just experienced a case that disproves the rule. Every time you watch a YouTube video, same story. Do you get glitches and slowness from those things? Yes, absolutely. But most of the time the glitches are due to cost saving measures taken by those vendors because they’re making their money from ads. Software you pay for should have far fewer glitches. And are you telling me your desktop software never glitched when your virus checker or other utility kicked in?
Try some of these packages. Both F360 and OnShape have very decent UI response. As for performance, remember, we can harness as many cpu’s as needed with the Cloud.
14. There is no successful Manufacturing Software that runs in the Cloud
Folks, this horse has already left the barn.
Consider Carl Bass’s (CEO of Autodesk) rebuttal of OnShape’s launch wherein he quotes lots of very large numbers about the adoption of Autodesk’s Cloud Offerings:
We launched AutoCAD 360 (formerly AutoCAD WS) in January 2010. AutoCAD 360 is a web and mobile CAD product that has 20,000,000 users who have created almost 100,000,000 models. TinkerCAD, a browser-based 3D modeling tool for kids launched in 2011. It has more than 500,000 users and more than 350,000 people use it every month. More than 4,000,000 models have been created and its usage usually peaks at about 20,000 concurrent sessions in a single day. And our broad-based collaboration platform A360 was released in December and already has over 100,000 users with more than 6,000,000 new files being uploaded every week. Onshape first? Hardly!
But Onshape is probably most similar to Fusion 360, our cloud-based CAD/CAM tool. Fusion 360 launched in June 2013 and has more than 50,000 users that have created almost 500,000 models. Let’s compare it to Solidworks just to give you a feel for the size of the market. In 2000, five years after Solidworks launched, it had 50,000 subscribers on maintenance. The numbers for Fusion are pretty impressive especially when you compare them with legacy desktop applications but the rate at which people are discovering and using Fusion is staggering. What’s even more pleasing to me is that more than 90% of these users are new Autodesk customers.
Consider the case of GrabCAD, a Cloud-based repository for CAD files which was recently sold for about $100M. That’s about 1/3 of what Dassault Systems paid when it originally acquired then highly successful SolidWorks. It’s roughly the same price the combination of Cimatron CAD and the highly regarded GibbsCAM sold together for recently.
Consider that OnShape has raised $64M in Venture Capital from some top firms. Those guys don’t invest that kind of money unless they smell a huge success in the offing. They are literally betting this company can change the whole face of CAD and garner a big share of that market.
If those things are not success, I don’t know what success is. It isn’t a question of when the first Cloud Manufacturing Software will be successful. It’s a question of how many new Successful Cloud packages there will be as well as how much more successful the existing packages will be.
Cloud is good, it’s strong, and it is here to stay. Take advantage of it. At least try it out. Come on in, the water is fine.
Like what you read on CNCCookbook?
Join 100,000+ CNC'ers! Get our latest blog posts delivered straight to your email inbox once a week for free. Plus, we’ll give you access to some great CNC reference materials including:
- Our Big List of over 200 CNC Tips and Techniques
- Our Free GCode Programming Basics Course
- And more!
Just enter your name and email address below:
100% Privacy: We will never Spam you!

Bob is responsible for the development and implementation of the popular G-Wizard CNC Software. Bob is also the founder of CNCCookbook, the largest CNC-related blog on the Internet.
What about all the garage ‘Shops’ using hacked copies of SolidWorks 2013 and Mastercam V6?
What will they do when they can’t steal the software anymore?
Isn’t that what this is really all about? Software companies protecting themselves?
Yawp,
Bob.
Bob, in a word, “No.” Sure there is piracy. Sure, some firms are obsessed with it. But most are not for three reasons. First, the incidence is lower than you’d think when companies measure it. Second, many of the pirates were never going to be paying customers anyway. Three, when you have one and two working properly, many of the pirates are actually not bad at helping get the word out to more paying customers.
For Cloud firms, there is other news. It’s just so easy for them to offer something free. OnShape gives anyone who wants it free access to their CAD software. There’s no need to pirate free software.
Honestly, it is way easier to convince the boss to pay $40 a month indefinitely for a constantly updating program than it is $2400 up front that some software companies want. My biggest issue with the graphic design program we use is that I rarely close it and it won’t check for a license update without closing. It then closes the things I had open.
Andrew, so true about convincing the boss, LOL. Most of the customers who buy the CNCCookbook software on a 1 year wind up coming back to buy the 3 year or the lifetime. They used the 1 year to get used to it and determine far more than a 30 day trial allows just how useful it would be to them.
Another advantage of subscription based software: The vendor has a steady, more predictable revenue stream. This allows them to fund operations and development without going though dry spells between major releases.
Likewise, the software user’s business has a predictable monthly expense that makes budgeting easier. It always seemed that my various programs would get a new release around the same time, blowing out my budget. Now I just treat my software subscriptions like I do my utility bill – a predictable monthly expense.
Ed
Ed, great point about consistent budgeting!
I had forgotten that one, but now that you bring it up, I do remember a lot of customers telling me they appreciated that on expensive SaaS/Cloud software.
Thanks!
I still don’t buy it! In the end most moves to the cloud where in reality a price increase, the effort here is to drain the customers. That is the primary reason for the move to the cloud.
The other reality here is that after a bit of capital costs, software is largely profit. This is one reason why MicroSoft was insanely profitable for so long. The prices we pay for CAD software are simply outrageous and frankly stunt the acceptance or wide spread use of these tools.
As for performance it isn’t a render that is the problem, that can run over night if need be. The problem is interactive use, scripting and other activities that lead to high productivity.
Here is a riddle for you then, David Frantz. If the Cloud is a scam to make even more obscene profits, why is it that companies that convert from “regular” software to Cloud software are instantly less profitable? There are many many examples of this and I am not aware of a single case where it made the company more profitable. It’s been quite the opposite.
In fact, it has been quite a challenge for non-Cloud companies to move to Cloud because of this. The move doesn’t happen over night, so when investors hear it is happening, the stock for these companies frequently crashes for years until they can complete the transition.
The answer is because Cloud is not a scam to extract even more profit. The answer to why companies like Microsoft are so profitable is that they have monopolies. If you want to get better prices, you need to encourage new competition. Cloud is that new competition.
I’ve already quoted at least one source saying OnShape is going to change pricing in this market forever.
I have to disabuse you of the notion that software is largely profit after capital costs. There is the cost of marketing and R&D for newer versions. None of those go away.
Lastly, try the Cloud software. It’s performance in interactive use is fine. There are many examples of it.
This software is free to the hobbyist, CH. You’re not being priced out of the market at all. You’re being encouraged to come into the market more cheaply than ever before.
The Cloud nearly always has a multi-layered (or tiered if you prefer) pricing scheme like this.
I can remember when we used punched cards, and ‘interactive’
was pooh-poohed. Many fights back then.
However, my CNC machine is NOT connected to the internet,
and will not be connected to it. That stops any interference with
the CNC SW, and there’s no connection available out to the
workshop anyhow.
And yet, I still want to run my CAD SWin the workshop at times.
Going a little further – Cloud may be great if you have a 20 MHz
bandwidth up and down. Much of the world does not, so you end
up with huge delays while your data uploads or the server at the
other end downloads. Google office stuff runs far too slow compared
to having the SW on my desk. It’s not “won’t use it”, it’s “can’t use it”.
Cheers
Why does it seem that technology lags behind in the world of CNC? It still baffles me at work, that our 2008 VMC has a floppy drive. All of its backup config files are stored on floppy disks.
So far I’m happy with onshape, I’ve only create a single part file so far. Last night I started to add a couple different versions to the one part file I am playing with. Its been very quick to load for me so far, but I do wonder how it handles large assemblies. I also assume that onshape doesn’t rely much on your own computers performance to run, which is a huge cost savings.
I agree with you on everything except the security issue. If you think for a second that that big companies working on proprietary designs (or classified, think Boeing, Lockheed, SpaceX, etc.) are going to connect those CAD machines to the internet, you’d be wrong.
Those companies are also going to trust their IT department far more than a CAD startup cloud service (who, let’s face it are CAD people not security experts).
Furthermore a central repository of CAD data like a cloud service would have is a prime target for commercial espionage from countries like China. However, your individual shop may not come under scrutiny because you are not high profile enough, or the effort is not worth the reward.
So while for an average user (like me), cloud based CAD is great and has a lot of advantages. However, for the big boys, who do most of the funding for expensive software, cloud based CAD is a complete non-starter.
-Jim
“DirecTV has tremendous bandwidth (down at least, up is slow and must be via land line) but lousy latency.”
Satellite internet did away with landline uplink early on because few were willing to pay for two internet connections and during that time the potential customers for satellite internet could at best get a 33.6K or slower dialup connection.
Government regulations require a certified installer because a fraction of a degree off on the dish mounting can have the uplink signal hitting the wrong satellite.
Latency is very poor. At the altitude of geosynchronous satellites the data has to make a 44472 mile (71572 kilometer) round trip PLUS any distance from the satellite service’s transceiver to wherever the data server is. Just for the trip to the satellite and back is nearly 1/4 second delay. Even worse is when packet loss happens or anything else that requires multiple re-transmissions.
Most people with satellite internet have download connections slower than 1.5 megabit DSL and a 256 kilobit upload speed. Why? Because this is a very expensive service. *Theoretical* maximum download speed is 1 gigabit with a 10 megabit upload speed. Got enough $ to pay for that?
Even with a fast and expensive satellite service, you’ll still run into problems with streaming video and other high bandwidth services due to that extreme latency. A person whose computers I regularly maintain had the lowest speed Wild Blue service. He upgraded to their mid level (at a significantly higher cost) and it did not improve performance on YouTube and other video services one bit. So he’s dumping that for a Verizon HomeFusion device with a 10 gig download per month, no bandwidth throttling. He’s fortunate that where he lives up over some hills his house has line of sight to a cell phone tower. He could also get T-Mobile but not AT&T or Sprint.
Yet another gotcha with satellite internet is bandwidth usage caps. Typically there’s a late night “all you can eat” period around midnight while “daytime” use has a ticker running on how many megabytes you download. Hit that cap in a month and your bandwidth gets throttled to little better than v.92 dialup.
Lastly, there’s the weather. The same weather conditions that knock out DirecTV, Dish and other DBS satellite TV will cut off satellite internet, which uses the same Ka band as satellite TV. If where you are has snowy winters and/or rainy rest of the year, you’ll have frequent “weather outages”.
One thing that gave many people a bad ‘taste’ of cloud services was Circuit City’s Digital Video Express (DivX) no-return DVD rental service. Every DivX Disc (they couldn’t use the DVD name and logo due to being non-standard) was encrypted using a proprietary method and had a unique serial number. A disc could be rented, played for 24 hours then *not returned*. It could be thrown away or additional 24 hour blocks could be purchased. Discs could be “upgraded” to Silver status, permanently unlocking them on ONE DivX player. There was a system to upgrade discs to Gold status, playable on any DivX player without paying more, but it was never used. Failure to use the Gold upgrade, which was touted as a pre-launch selling point, got DivX off to a bad start.
DivX players stored the status of Silver discs in a memory chip. There was a limit to how many Silver upgraded discs a player could have. When a DivX disc was put into the player, it would check its local memory for the serial number. If not there it would call up the Circuit City DivX server to check the rental status.
As the system collapsed due to people failing to fall for the scheme (it wasn’t that much of an inconvenience to return rentals), Circuit City decided to (literally) pull the plug – and stick the knife in one final time. From the start the DivX servers had a system where all DivX players could be commanded to call in and be programmed to permanently unlock encryption for ALL DivX discs, but the company decided to screw everyone who’d bought into DivX and just shut it down. Millions of DivX discs permanently became forever trash and any Silver upgraded discs had to be transferred with the players they were locked to, or they’d also become worthless trash.
DivX players also played regular DVDs, but most of the models were rather low end units.
Gregg, DivX was not a Cloud Play at all, it was a copy protection scheme. Plenty of that in non-Cloud Desktop software land.
#7/8/9 So what happens to us G-Wizard users when you, hopefully not for a long time, kick the bucket?
We are stuffed. We are in full control of our data but because of the online licensing model, we can’t do anything with it.
You just posted about OnShape – looks cool, but again, no control of your data. Even if you could download it, if they go bust, you can’t do anything with it. (TinkerCAD?)
#12 I’m sure it is a twitter thing, but people have short memories these days. Anyone remember stuxnet? An extreme example, but a valid reason not to have your industrial hardware online.
Maybe its just me, but it will be a long time before I trust a cloud service for anything of value, or anything I depend on.
Frank, you’re only stuffed if you can’t access your data in a file format that other software uses. That’s why I gave the Best Practice I did on #7.
As far as Stuxnet goes, you’re totally making my case. Stuxnet was created to attack a group of computers that were not online. The Iranian facility had no Internet connection whatsoever, yet Stuxnet got to it anyway via a USB thumb drive.
If you think all you have to do is disconnect, you’re kidding yourself.
If you’re extreme paranoid about malware and security, you use a non-critical system to plug in and scan any portable storage media. Only after it passes that, preferably after the desired data files are transferred to known clean portable media, does anything get transferred to the production system.
Not paranoid enough? Allow no files created outside the facility to touch the production system.
Bob, I think his point re: your eventual mortality is that, if someone has bought a lifetime license, the effective limit is the customer’s lifetime or your lifetime, whichever limit is reached first. Same thing is an issue with the shorter-term licenses. If I buy a 3-year license today and you get run over crossing the street tomorrow, what happens to the remaining 2 years and 355 days of my license?
Now there’s plenty of other issues with cloud-based software, too. I demoed a variety of cloud-based, virtualized and / or SaaS solutions with my last company, and without exception every single one SUCKED in implementation. Local ones that required constant check-in with mom were at the mercy of Verizon’s unreliable infrastructure in our area (agricultural dust + fog = short circuits). Things that were hosted entirely in the cloud or virtualized off-site had the same issue with Verizon constantly dropping our connection, but had as additional issues tacked on the fact that our corporate IT department was paranoid and ran the entire company (60+ locations in 46 states) through one central node before you could get out to the backbone. This is not conducive to having good keyboard / mouse / video response.
It was just flat-out terrible.
I’ll be more than happy to pay for an annual maintenance subscription for software that is hosted entirely locally on my machine.
Rusty, no, it wouldn’t be limited to my lifetime at all (which will be very long!). It’s a going concern that makes excellent money, not a hobby. It’s growing gang busters. Just the web traffic alone (3 million visitors a year) is very valuable. As such, my heirs will want to continue the business. There are other programmers involved who could take up the slack and are already familiar with the code. Matter of fact, we have a third product in the works. And, the Cloud portion of the business costs us a couple hundred bucks a month total from Amazon. So no, it isn’t at all limited to my lifespan. Folks are highly incented to keep it running, it’s cheap to keep it running, and it’s lucrative. Trust me, this is America, I know my heirs, and it will be kept running in my absence, LOL!
If you live in the sticks and can’t get decent Internet, my sympathy–you’re stuck.
It’s interesting how people look at it. You didn’t have to buy the lifetime. You could’ve paid for it a year at a time and protected yourself from sudden calamity. But, you wanted the savings even though it would’ve taken years to achieve those savings. Now you’re worried enough years won’t go buy.
I look at it much differently as a business man. The average end mill around here costs from $40-70. If I save breaking two of them, I made the yearly cost of G-Wizard Calculator. If I am manufacturing something, my hourly rate is probably on par with the annual cost of G-Wizard. If I save 1 hour a year, once again, I paid for it. If I either know feeds and speeds off the top of my head so well there is no savings to be had or I’m not doing enough machining to reap those savings, I’m not really a G-Wizard customer in the first place.
If you don’t think G-Wizard can deliver those savings, that’s fair too. But if all this is about economics and how long Bob will live (!), you’re not making much sense. I’ve got customers that tell me they’ve saved anywhere from $5000 to $25000 just during the trial period with the software. My wife hears that and says I don’t charge nearly enough. Read our testimonials. They’re all real. There’s thousands of customers like them. That’s a big enough audience that has no problem at all with the Cloud aspects of our software that I don’t worry about it too much. You shouldn’t either.
I am a Gwizard customer, and it’s already saved me quite a bit of money on end mills. Yours is perhaps the least obtrusive of the subscription models of so-called cloud software.
The example I listed of poor internet service? That wasn’t out in the sticks. That was in a city of over 60,000, and immediately adjacent to Fresno, CA – population 509,000. And the company was a 15 b-b-b-BILLION dollar a year company, so there wasn’t a budget issue on our part causing the problem.
The only reason I can see for a “cloud” program that actually has 100% of it installed locally is for the vendor to be able to “shut off the tap” for customers who elect to quit paying the subscription fee, rather than allow them to continue with the last updates they have.
For a lifetime “subscription” (which amounts to an outright purchase plus whatever free updates the vendor will provide) there is zero reason to have the software “phone home”, other than to keep tabs on whether or not the customer is abiding by the number of computers the software was purchased for.
As for the “cloud” term, it’s very much abused by internet connected or hosted software companies. Go get a Novell Netware manual if you want to know what the term is used for. The “cloud” term originated as a metaphor for all the parts of a network not directly relevant to the part you are currently working on, or infrastructure between LANs – anything you don’t need intricate detailed knowledge of *right now*. You can just scribble a blob on the network map and label it “cloud”.
If you’ve shot all the troubles on the parts you’re responsible for, then you stick your head in the “cloud” and see if the problem is “higher up” and if you can fix it or find who should be fixing it.
Gregg, you silly goose. The Novell Netware use of the term “Cloud” is as antiquated as Novell Netware itself. The Cloud has evolved as has the Internet in general. We don’t all have to access it via AOL, Compuserve, or (horrors!) MCI Mail. We don’t have to use Novell Netware either.
As for there being no use other than licensing, you’re just not using enough imagination. Collaboration, backup, and keeping current with updates are three obvious uses that have nothing whatsoever to do with backups. Gaining access to new information as it becomes available without having to deal with manually installing it is another. There are many more.
I love Fusion360’s ability to grab 3D models from McMaster-Carr as an example. Fusion360 is installed on your machine, and any CAD software that is installed could provide the same function. Are we arguing about Clouds or Angels on the head of a pin?
Neither, really. The Cloud brings value that is worthwhile. There are lots of ways to interact with the Cloud, and there are things you need to be on the lookout for in your Cloud software to get the most benefit with the least downside from it. That’s the whole point of the Best Practices enumerated above.
Cloud doesn’t have to mean “Runs in a Browser”, although it can. All it really has to mean is “Delivers the benefits of being connected and of Cloud Software in general.” Not all software will deliver all the benefits, nor will it all have the same drawbacks. Learn to understand the nuance and the possibilities and you can win. But if you start from the assumption that the only thing it’s good for is copy protection, why even bother to join the discussion? New things always bring fear of one kind or another, but we get past it. We’re far enough along in how the Cloud has touched almost every aspect of software and computers that it’s time to get past it.
Rusty, sticks or no sticks, if you’ve got bad Internet, you’re justified in avoiding Cloud services. But do keep in mind that’s a personal problem and not a problem with the Cloud. It may also be a problem with your IT, regardless of how many billion dollars a year the company is. In fact, based on my experience, the more money the company has the more trouble IT may be causing its internal “customers”. Sad but true. Running everything through the one hub is an example. If the Silicon Valley companies that know tech inside out don’t do that or recommend it, why the heck does the IT department think it’s a good thing? The answer is that it is convenient for them for some reason, and they just don’t care about you. There again, that’s not the Cloud’s problem, though you may very well have to avoid the Cloud just to get around the shortcomings of you own IT.
Most companies don’t face those challenges. If they did, the Cloud wouldn’t be nearly as prevalent as it is. For software companies, it’s a done deal, here to stay, and not even in question. I’m in a position to know that well, and there is no doubt. For those that don’t believe it, that’s fine. You can still buy brand new Bridgeport manual mills too, and a lot of guys really like them. Doesn’t mean they’ll take you to the head of the class though.