Dissecting Dieter’s perpetual points

I like Dieter Schlaepfer, we’ve been arguing for years.

Dieter and I have never met in person, but online we go back to the CIS:ACAD CompuServe days of the early 1990s. Dieter’s a good guy who has done a splendid job with AutoCAD documentation content for decades. He is genuinely interested in improving the product and customer experience, and has done a great deal to do so over the years.

Dieter’s responsible for the most-commented post on this blog, AutoCAD 2013 – An Autodesk Help writer responds with 164 comments and was a heavy contributor to the 95 comments on the recent AutoCAD 2018 – why did the DWG format change? post.

If you read the comments here, you’ll know that Dieter is the only Autodesk person brave enough to put his head above the parapet and enter into discussions here in recent times, even though he’s not doing so in any official capacity. Autodesk’s PR people give me a wide berth and the Autodesk view would be completely unrepresented here if not for Dieter. He’s prepared to put his hand up and say, “But what about this?” when it’s an unpopular viewpoint and nobody else is prepared to say it. Props to Dieter for that.

Among Dieter’s many recent comments, he outlined 12 considerations in the rental v perpetual argument. Myself and others have been having fun eviscerating his tortured pub analogy, but his more serious arguments deserve a more considered response than can be comfortably provided in a comment, hence this post.

Let’s take Dieter’s considerations one by one. Bear in mind I’m approaching this from a long-term Autodesk customer point of view. You may look at things differently, and that’s fine.

1. Cost – if a rental, lease, or membership were low enough in price, almost everyone would do so (at a dollar a month, heck, I’d lease my shoes)

Fantasy argument. If Lear would hire me a private jet for $1 a month, sure, sign me up. The reality is that rental costs more, except in the short term. That’s why companies rent things out: to make money. That’s why Autodesk is doing it; it’s an attempt, however hamfisted, to make more money. On cost, rental loses.

2. Business model, terms and conditions, and their consequences

Vague. But given the terms and conditions attached to Autodesk rental (standalone users must use a terrible licensing system) and the consequences (software stops working the instant you stop paying), rental loses big-time here.

3. Quality of fulfillment – this is to your point

Not sure what Dieter means here. ???

4. Tax consequences

This varies from place to place. I can get a 100% write-off whether buying a perpetual license, maintaining it or renting it. I may want to get a bigger write-off sooner, or not. Neutral.

5. Opportunity cost – by tying up a lot of cash, what potential opportunities do you lose?

Depending on a business’s cashflow and other circumstances, this is a possible valid argument. However, only in the short term. Because rental costs more in the long term, it costs you more in potential opportunity in the long term. Overall, rental loses.

6. Financial accounting – rental, lease, or membership costs can easily be assigned to each project and billed to each customer

If I don’t have the need to do that, it’s irrelevant. But even if that’s the way you need or prefer to do things, it’s still only partially true that rental can be a benefit. Let’s say you have won a project that is planned to take 9 months and rent Autodesk software for a year: it costs you $3000 and you pay up front (because you’re not insane enough to pay Autodesk’s monthly rental prices). You finish the project in 10 months. You use that software for other smaller projects that crop up during that 10 month period, and after it’s over. Quick, how much of that $3000 do you apportion to each project? See, it’s not as simple as it appears.

It’s really not difficult to have perpetual license software costs handled in the same way as overheads and other long-term costs that can’t be directly attributed to a project. You’re not going to be able to sack your accountant thanks to software rental. Neutral.

7. Flexibility – you can easily increase or decrease the number and types of licenses for several (not just one) products

Ah, flexibility. Let’s say I’m convinced by Dieter’s other arguments and convert my perpetual license to rental under the current so-called “discount” offer. In doing so, I throw away my flexibility. I can’t ever stop paying or my software stops working. Down the track I may not need that license for a while, but even then I can’t hop off the rental train because if I do that and then hop back on, my software costs will treble (roughly – it varies).

As for the several products thing, Autodesk has been pushing customers into suites and collections where a high price is paid for a block of products. Can you drop back from a collection to a product or two for a while, then back to other products or up to a collection again? Sure, Autodesk is very flexible. Just forego your “discount” and pay an astronomical increase, no problem.

Autodesk has been progressively removing its customers’ flexibility for decades and will undoubtedly continue to do so as long as it thinks that will make more money that way.

So please don’t come the rental=flexibility argument with me. On flexibility too, Autodesk’s rental loses.

8. A truly perpetual software license requires you to maintain obsolete hardware and old operating systems, and discourages the adoption of new technologies

No it doesn’t. A non-upgradable license might do that, whether perpetual or otherwise. That doesn’t apply to perpetual licenses under maintenance. It didn’t even used to apply to perpetual licenses not under maintenance. Whose fault is it that perpetual licenses not under maintenance are no longer upgradable? Autodesk’s. False argument.

9. Perpetual licenses put most of the financial burden on new customers rather than spreading it more fairly between all users

Actually, with perpetual licenses the financial burden is much more fairly spread. What costs more, developing a product from scratch or maintaining it? Perpetual license purchasers pay a higher amount for the initial purchase, just as the developer pays a higher amount for the initial development. The developer is fairly rewarded for providing the product for the customer to use. Following that, the developer is fairly rewarded for maintaining and improving it.

But I really hope you’re not trying to convince people that Autodesk is price-forcing customers onto rental in order to be fairer to them, because I think incredulity would be the appropriate reaction. False argument.

10. Perpetual software licenses create “a long tail” of product versions, making data sharing between users more difficult

Perpetual software licenses that are maintained do no such thing. If a vendor provides good value for that maintenance payment, then people will maintain the software. Autodesk maintenance value for money has been dreadful in recent years, leading to people dropping it. Improving Autodesk’s performance in that area would reduce the length of the tail. Making maintenance value for money even worse by racking up prices will lead to people dropping it and sticking on old releases much longer. Autodesk’s rental push is lengthening the tail, not shortening it.

Incidentally, there is a new benefit for subscription customers with multi-user (network) licenses. Guess what? Five releases are now supported rather than four. Yes, Autodesk rental is literally lengthening the tail. False argument.

11. Perpetual software licenses encourage users to use less secure software and operating systems in a time when cybercrime and espionage are mushrooming

See 10 above. False argument.

12. Providers of perpetual licenses have less incentive to support long-time customers than providers of rental, leased, and membership business models do

Absolutely wrong. This is literally the exact opposite of observed reality.

You know what model really provides an incentive for vendors to improve the product? Perpetual licenses with optional paid upgrades. With the perpetual/upgrade model, if there’s no improvement, there’s no ongoing income. But that model was too much like hard work. Easier to just remove our options over the years to manipulate customers into paying more and getting less. Autodesk priced that model out of the market and then killed it off because it wanted to get people paying for just using the software rather than as a reward for improving it.

It’s proven by history. The closer Autodesk got to the all-rental model, the worse the rate of improvement became. As an improvement incentive, rental loses.

There you go, Dieter. Rental loses on five considerations and wins on none. And I’m being generous by considering points 10 and 11 as neutral.

Feel free to add your own observations on perpetual v rental. If you want to have a go at Dieter’s arguments or mine, go for it. I just ask that you play the ball, not the man.

22 Comments:

  1. Dieter Schlaepfer

    Steve,
    Thank you for your generous words.

    To clarify point 3, this was actually my nod to CadBystander’s objection, which I think is indeed a valid concern.

    As you probably noticed, I listed points 1, 2, and 3 first. This is because they are the fundamental criteria of whether a rental, lease, or subscription is worthwhile for any individual or any size business. Point 4 is probably close behind, depending on where you live.

    The other objections you made to my points 4-12 might apply to [redacted], but you’re certainly not speaking for all organizations, right?

    For example, do you know why companies in the UK (at least a number of years ago) routinely included a leased automobile as part of the employee benefits for high paid employees?

    And yes, I was part of the subsidiary of a large company before I worked at Autodesk. After they had acquired us, they sold our furniture and computers to a leasing company–and then leased it back! Sounds crazy, doesn’t it. Why would any company managed by competent finance people do such a thing?

    Kind regards,
    Dieter

    • I’m not speaking for my own organisation, but my viewpoint is certainly influenced by my experiences and situation there.

      Yes, as I acknowledged, it is quite possible for 4 to be a valid benefit of rental in some locations, as it is a deficit in others (like mine).

      Yes, years ago in the UK, many more company cars were provided to employees than in other countries because of some fairly ridiculous tax benefits. I don’t know if that still applies. Note that those cars were leased with a residual purchase option at the end, not rented.

  2. Dieter Schlaepfer

    Steve,

    >> I’m not speaking for my own organisation, but my viewpoint is certainly influenced by my experiences and situation there.

    Yes, same here.

    >> Note that those cars were leased with a residual purchase option at the end, not rented.

    True, but it’s recommended NOT to lease a car for longer than the warranty period, which averages three years, or 36,000 miles. The reason is that this is a turning point in the car’s life, when the warranty expires and maintenance costs typically rise. For this same reason, it usually makes more sense to lease a brand new car than to purchase and drive the wheels off the now 3-year old car.

    Let’s have some fun.

    Here’s a nice little sport utility vehicle available for a 3-year lease in the US:

    2017 Toyota RAV4

    • Monthly Payment: $219 US
    • Term: 36 months
    • Mileage: 12k (about normal for me)
    • Due at Signing: $1,999 US
    • Effective Monthly Cost: $275 US
    • MSRP: $26,074 US

    Prices in Australia are probably still outrageously higher, and there are well-known down-sides to leasing an automobile–driving more than the stipulated limit for one thing, and getting into an accident for another (for which it’s recommended that you carry additional insurance).

    So you get drive a vehicle that’s perpetually less than 3-years old for $3,300 US per year, not counting maintenance or repairs. You let someone else pay the $27,000 US for the perpetual license. The break-even point comes after driving a purchased car for about 8 years in this case, but importantly not counting repairs.

    I admit it’s tempting.

    Regards,
    Dieter

    • Those poor leasing companies must be losing money hand over fist. I wonder how they ever survive?

      It’s good that car leasing is available as an option for people to consider. It undoubtedly makes sense for some customers under some circumstances. Same for software that’s only going to be needed for a short time.

      If Toyota ever stops selling cars and makes leasing the only option, the analogy will be a lot closer to the reality of Autodesk’s situation.

  3. Dieter,

    I happen to lease.

    After comparing the actual overall costs of my last vehicle purchase, maintenance, and minor non-warranty repairs (debatable), which happened to average $274/month (after resale), vs. leasing – it was no contest.

    The new car retailed for exactly 50% more than the last purchase ($36/24K), within the contracted 41K milage cap I won’t need to replace either tires or brake pads but will suffer through an estimated $350 of nothing more than glorified oil changes, fat insurance costs haven’t increased, and the lease is $5/month less, plus the initial $2K up-front gouge.

    At the end of the lease I’ll simply drop it off and drive away in the next lease. Total out of pocket expense is nearly identical to the depreciation of the purchased vehicle.

    That’s a 6% buy-in, and monthly cohabitation costs of less than 1%, or just 9% annually.

    The equivalently leased AutoCAD permanent license ($4,195 +taxes) would require $250 up-front and $375/year for the equivalent three year lease.

    That same $36K automobile under the generous three year Autodesk Subscription program ($3,970 +taxes) would require $34K in advance and would average just $945/month. I would get an updated new car each year, but I’d still have to pull up on the brake pedal to stop, front seat passengers could only exit through the trunk, and the owners manual would occasionally be available on the cloud.

    Comparing oranges and platinum. – Bill

  4. Dieter Schlaepfer

    Steve,
    According to recent statistics (http://www.statisticbrain.com/auto-loan-financing-and-leasing-statistics/)

    • 36% of Americans purchased their car outright
    • 21% of Americans are leasing their car
    • 43% of Americans are financing their new car

    I believe these statistics are for private vehicles only, and not for commercial fleets.

    According to numerous reports, leasing is growing at a very strong rate. According to Edmunds.com:

    “Lease volume has doubled in the last five years, suggesting that the automotive market could be on the verge of a fundamental shift in consumer mindset about the value of owning a new vehicle — particularly when the purchase has to be financed.”

    https://www.edmunds.com/about/press/number-of-leased-vehicles-reaches-all-time-high-in-first-half-of-2016-says-edmundscom.html

    To your points, I’m generally in favor of more choices and options than fewer, and I’d like to see a free option for Makers as is available for Fusion 360. But to represent the situation in the software industry, consider the decision made in Switzerland regarding older used cars and polluting emissions. I believe that after three years (or is it five?), you have to sell your car. In the software industry, revenue aside for the moment, it would be far better for everyone to be using up-to-date software for compatibility (somewhat), collaboration, and cyber security (more important than most people realize).

    Bill,
    With auto leasing, there’s a significant value in the used vehicle. Leased software cannot be purchased as used, which explains the cost disparity.

    But I loved your satire on passengers having to exit the car through the trunk! This is far too often true, and has been brilliantly illustrated by Alan Cooper in his famous book, “The Inmates Are Running the Asylum: Why High Tech Products Drive Us Crazy and How to Restore the Sanity.” A few years ago, I had a conversation with Alan over lunch.

    Again, cost considerations are only one (albeit an important one) of about 12 elements that should affect the decisions of thoughtful, analytic people. As Steve, I’m also confident in an efficient market that makes the ultimate decisions on price and terms.

    Regards,
    Dieter

    • So about 4:1 prefer to buy their cars. Isn’t it nice they get the choice?

      > Leased software cannot be purchased as used

      Only because of Autodesk’s unreasonable and artificial restrictions on resale of licenses, except in the EU where the courts have quite rightly thrown those out.

      > which explains the cost disparity

      No. Just no.

    • Dieter,

      Not to the leasee. I don’t own it, nor can I sell it. Similar to some software products currently under seige. What the leaser does with their vehicle at the termination of our contract does not affect me. They may offer to extend the lease at the same, or a different rate. They may offer to sell the vehicle to me for far more than it’s actual value. They may toss a mint to me as I exit their facility. I’m old business.

      Failure to accommodate current or potential customers by not allowing resale of extra, unneeded, or abandoned licenses would be the unilateral decision of the software provider. The software value instantly plunges to zero upon registration.

      “Cost disparity” is kinda hard to swallow since currently there appears to be zero physical deliveries of anything with many (not all) software products. No disks, no printed documentation, nothing but bloated downloads and eMailed receipts. The cost of providing download delivery is certainly low. Manufacturing costs are nonexistent. Costs related to continued development are obviously fictional.

      One major similarity between the development and manufacturing of AutoCAD and automobile: ACAD has several thousand commands, variables, and thingies that need to work together accurately, reliably, repeatedly, and correctly so that the product will perform as designed. A modern automobile has to do the same thing with up to seventy-five thousand individual parts under nearly unlimited mostly unknown conditions. With a warranty.

      One major difference is that each and every one of the automobiles many parts must function as specified or designed. There are no unfinished components. Recalls are not service packs. Updating the driver won’t prevent fatalities if somebody decided that not heat treating front hubs would save nearly 63¢ per unit.

      It’s not the cost of the software – it’s value received and perceived. Ribbons, silly schemes, unfinished commands, enforced punishing unknown “rental” futures, and online anything does nothing to improve my output. Neither does redefining “permanent.”

      Actually, this is academic for me. After thirty-one years – the 2016 Product Design Collection will be my last Autodesk license. Rental is not a realistic option. My workstations will never be allowed online. I would never entertain any form of cloud-like participation without an eight figure performance bond for myself – and at least nine figure bonds for each client, and especially never where I’m contractually bound to never let a single file out of my direct control. I feel that the continuous degradation of Autodesk products will cause professional damage. And not just to me. Can’t guarantee safety if I can’t trust the software.

      ‘Couple of years ago I offered (for cash) to analize every single AutoCAD command, operation, variable, and option – and deliver a written and illustrated report on what doesn’t work, how I expect things should work (based only on my mechanical greed), and what I would like to see added to make the product meet my needs and expectations. Overall it would be an incomplete exercise since my design disciplines are significant, but limited. I no longer have the time or interest to offer this valuable service. Somebody, or a team, outside the company needs to be retained so development won’t remain at the 2008 release. Nothing being done by Autodesk is helping.

      As long as I’m wasting space, why did I lease a car? Out of pocket costs are similar. It’s a plug-in hybrid with a fairly comprehensive warranty. I don’t want to own this toilet beyond half the warranty period. $8K+ battery, $15K water-cooled controller, many unknowns. I expect rapid depreciation beyond four hears. Any hybrid surprises along the way will drop the value to zip. It gets about 36 of the advertised 45mpg, but my fuel costs are $80/month less than it’s predecessor.
      Plus I get distracting green cartoon dashboard vines as I feather the throttle trying to keep it running on the battery pack. I’m easily entertained… – Bill

      • It must take a lot of fuel to cart that battery and bonus motor around everywhere. My old-fashioned comfy-seat-by-the-radio big-enough-to-take-the-family-fencing-kit 140+ MPH Accord Euro 2.4 (Acura something to you) gets 32 of your strange American MPG or 38 genuine Imperialistic ones despite me driving it. 5-year warranty is now gone but I am serenely unconcerned because it remains in typically Japanese as-new condition. Unlike some old Japanese cars it doesn’t leap out of gear when I accelerate hard (remember that from 22 years ago?).

        Yes, I own it. I like owning things, and am glad to have the choice. Lease sounds very wise in your case.

        Previous Japanese hard-driven fully-owned sporties have been on-sold in near-new condition for decent sums after 10 and 7 years. This works for me. The self-soiling perpetual maintenance 1.9 turbodiesel (Holden-badged Opel Astra) didn’t. Lesson learned.

        • * It must take a lot of fuel to cart that battery and bonus motor around everywhere… 140+ MPH Accord… gets 32 of your strange American MPG or 38 genuine Imperialistic ones… Unlike some old Japanese cars it doesn’t leap out of gear when I accelerate hard (remember that from 22 years ago?) *

          Battery eats a third of the trunk/boot too. That 36mpg snivel is on gas only. It’s about twice that when the green vines thrive.
          You’ve been in my backyard – seeing 70mph is a surprise.
          The major difference in our optimistic milage systems is that one of us has the fuel gauge mounted incorrectly.
          22 years. Tempus effit? The self-locking jam nut on one of the three gear cluster shafts unlocked itself. For some reason I bought another one of those turbo-skates – then threw them both away when the California smog standards changed in ’98/99.

          I popped up a calculator. By not renting/subscribing the latest PDC for $6,650 and applying the savings, my very unsporty Mondeo only sets me back $185/month – including fuel and insurance… And I don’t feel probed. – Bill

  5. Dieter Schlaepfer

    Steve,

    >> Only because of Autodesk’s unreasonable and artificial restrictions on resale of licenses, except in the EU where the courts have quite rightly thrown those out.

    Ok, we’re now moving on to my second point: business model, terms and conditions, and their consequences. What business model can maximize Value—both to Autodesk for its investment and to Autodesk’s customers? This dynamic also applies to other software producers and companies in general, of course. But you touch on some important points here.

    What’s Value to a company?

    In my opinion, it’s the ratio of utility to investment cost and includes scope to an extent (which is why smartphones that do everything poorly have been successfully displacing high-quality special-purpose devices). Investment cost comprises mainly company time and resources that could otherwise have been used to generate revenue. This investment usually exceeds the price of the software by far.

    Autodesk customers have customers as well, and want to deliver good value at a good price to grow their revenue and ensure longevity. These relationships form a chain that extends to a consumer of that value and beyond. Important factors include eliminating non-essentials, structuring the company for higher efficiency, and creating products and services with a close fit to the majority of customers’ most important requirements. Fulfilling these goals is not easy to do well.

    Not considering price for the moment, what factors would you say are the most important for most customers of software?

    Bill,

    >> The software value instantly plunges to zero upon registration.

    I don’t agree. The software value is in its utility.

    >> “Cost disparity” is kinda hard to swallow since currently there appears to be zero physical deliveries of anything with many (not all) software products. No disks, no printed documentation, nothing but bloated downloads and eMailed receipts.

    When I started in CAD/CAM/CAE, software and hardware were bundled together. We made money on hardware margins since any fool could see that a physical tape for a tape drive cost only maybe $30. Unbundling the software from the hardware was a huge, scary step at the time, because most customers expected to see physical hardware.

    >> It’s not the cost of the software – it’s value received and perceived . . . [snip] Ribbons, silly schemes, unfinished commands, enforced punishing unknown “rental” futures, and online anything does nothing to improve my output.

    Exactly. Things that you can’t use or won’t use are wasted on you. You’ve just made the third point on my list: Quality of Fulfillment.

    >> I would never entertain any form of cloud-like participation without an eight figure performance bond for myself – and at least nine figure bonds for each client, and especially never where I’m contractually bound to never let a single file out of my direct control. I feel that the continuous degradation of Autodesk products will cause professional damage. And not just to me. Can’t guarantee safety if I can’t trust the software.

    You make the point of requiring support for secure installations due to the potential of espionage and malware. I’ve worked with people in secure facilities. This is the eleventh point on my list, which I’ll need to qualify as follows: The most secure facilities exclude internet connectivity, include strong encryption, digital signatures for critical data and executable files, 16+ character conforming passwords, data access control, and physical security. I agree that software vendors need to accommodate this level of security, and I don’t know what Autodesk is providing in this regard, if anything. If you send me (dieters@youknowwho.com) a personal email, Bill, I’ll follow up for you on this question.

    For customers who do not need this high level of security, keeping up with all software patches and upgrades for the OS, and all applications is vital. Cloud storage with strong encryption and access control typically provides a higher level of data security than local networks.

    >> As long as I’m wasting space, why did I lease a car? Out of pocket costs are similar. It’s a plug-in hybrid with a fairly comprehensive warranty. I don’t want to own this toilet beyond half the warranty period. $8K+ battery, $15K water-cooled controller, many unknowns. I expect rapid depreciation beyond four hears. Any hybrid surprises along the way will drop the value to zip. It gets about 36 of the advertised 45mpg, but my fuel costs are $80/month less than it’s predecessor.

    Good points all–thanks. My next automobile might indeed be leased as a result of this conversation–I’m seriously considering the option!

    Kind regards,
    Dieter

    • Dieter, I need to type less and not take advantage of Steve’s generosity. Maybe tomorrow.

      ** >> The software value instantly plunges to zero upon registration **

      * I don’t agree. The software value is in its utility *

      Sorry, that was meant to represent cash value.

      * Things that you can’t use or won’t use are wasted on you *

      It’s not that, it’s about poorly implemented changes that hamper my productivity.
      I dumped Inventor at 2009. The ribbon only interface turned an already sluggish and overly repetitive program into an annoyance. Granted there were some great features and huge potential, but why bother when 3D in vanilla ACAD was five times faster, wasn’t limited by a single part file system, and the UI and modeling structure weren’t designed by an apprentice cooper.
      The silly UI schemes showed up in ACAD 2009. Why? Where’s the off switch? Where’s the “Classic” scheme that showed up in a single beta and didn’t function. I edited a scheme file looking for flat boring neutral gray. No joy. The idiot MSI started screaming “Danger, Will Robinson. Danger.” and insisted on repairing the installation before the sky fell. The annoyance factor. Change without reason. Changes without considering the customer’s needs/choices. Much like the rental scam.

      * requiring support for secure installations *

      Simple. Include everything in the installation file. Ensure that the stand alone installation is just that.
      Include offline help identical in content and depth to the online version. CHM help files used to have an update link. Update in a HotFix. Compile the help system to a set of CHMs, including separate files for all standard documentation. All of ’em. Include PDF versions too. This is little more than printing copies of portions of the properly constructed help universe. Let the customer decide what’s displayed. The eight Mac customers won’t care about the CHMs.
      Include all of the textures in the installation file. There’d be plenty of room if all of the existing duplicate files were removed. I think I recall that the VBA enabler wasn’t allowed to be bundeled, but that could possibly be addressed.
      Install that puppy and request the authorization code from your phone.
      A nice secure installation attached to your cold Intranet.
      I won’t ever see it, but I imagine that the monthly subscription re-up is just another authorization code. Same over the phone procedure.

      * The most secure facilities exclude internet connectivity, include strong encryption, digital signatures for critical data and executable files, 16+ character conforming passwords, data access control, and physical security *

      Problem is that five of those systems require computer and/or human interaction. Computers sometimes misspell ones and zeros. SSDs can get an itch. And Lenny might make a boo-boo. Again. Massive redundancy may result in massively compounded errors. A tiny electronic sneeze just as thirty independent back-ups are written might cause a problem. Heads/tails? It’s a never ending tail chase. However, a single CAD driver on their dedicated machine attached to their dedicated server addresses is gonna present a pretty reasonable security scenario without outside evil intervention.

      ** So, Bill, why are you so anti cloud? **

      I don’t buy into the sunshine and unicorn hype. Trains. I envision a dilapidated box car parked on a siding somewhere. Smelly old straw lies scattered across the flooring. Tossed haphazardly on the straw are thousands of old brown paper sacks. Most have sticky notes, old Christmas labels, or sloppy Sharpie notations attached. Some of my data should be in one of those sacks, but could be someplace else. I request a file. My telegraphed message starts tippity-tappity printing in the nearby train station. Ole Jake tears his copy from the teletype, stares at it for a bit, reaches over to retrieve his spectacles, and then reads my request. Jake yawns and stretches, reaches down to give Old Blue an ear rub, stands up and shuffles toward the door. Jake grabs the match perched behind his good ear, lifts the glass on the signal lamp, flicks the match head with his thumbnail, twice, and lights the wick. Jake stumbles out to the platform and starrs to swing the lamp back and forth while he watches the line of old train cars over yonder. No response. Again. Jake sets down the lamp and wanders back inside the station. He grabs the shotgun, rifles through his desk for a pair of shells, and returns to the platform. BLAM – BLAM! Jakes blows out the lamp and returns to his desk. Old Blue thumps her tail. After a spell Clem drags himself into the station. Jake holds up my telegraph. Clem sighs loudly and grabs the yellow sheet. Old Blue growls quietly. Clem limps toward the abandoned rail cars, he has a bum ankle. Some say it’s the drinking. Imagine Clem’s delight when he finds car number WZUD37TN nearby. Great! This work builds a powerful thirst. Clem reaches up to remove the stick from the lock hasp. The last lock rusted away. He struggles to slide the door open, a handful of sacks fall to the ground joining others. The far door is wedged half open. Clem grunts his way up into the box car. He kicks the bags around. Surprisingly he finds my sack. Clem roots around to see what’s in there, shrugs, and semi falls out of the box car. Since he might return later, Clem leaves the doors open and limps back to the station. Jake and Old Blue are gone. A handful of strangers are milling around. Probably waiting for a passenger train. Clem tosses my brown paper sack toward the desk, turns, and limps back to his shack. A powerful thirst.

      In shorter words, we have no idea where our files live, who has access, or how the files are handled. Do Cloud providers graciously provide those fat performance bonds to back up their promises and good old boy guarantees? Of course not, didn’t you read the EULA?
      Cloud rhymes with Ponzi.

      On a happier note…

      Sorry to sound so negative. But I am. Huge ACAD improvements from 1986 through 1992. Absolute magic up to 2000. Pretty stagnant through 2008. Then continuous annual disappointment.
      It’s March. Upgrade time. So? Happy days are gone.

      It appears the corporate drive is to dispose of all existing customers and market to a fresh bunch of potential customers who simply don’t know any better. Instead of a loyalty program rewarding long term users, we’re expected to pay nearly full retail every three years.
      I’ve skipped the last two releases since loyalty requires reciprocity, and there is that subscription thing.

      It’s sad. Autodesk used to be my friend and partner. Now shes that girl with bad makeup who cuts herself and hangs out nearby, hoping I’ll go somewhere without locking my door.

      Littlr anger issues there, Bill? Yeah…

  6. DR >Not considering price for the moment, what factors would you say are the most important for most customers of software?

    Depends on various things, but functionality, performance, efficiency, reliability, continuous and ongoing availability are among the factors.

  7. Dieter Schlaepfer

    Steve,

    >> Depends on various things, but functionality, performance, efficiency, reliability, continuous and ongoing availability are among the factors.

    Yes. To that list, I would also add low disruption. As I’ve probably mentioned before, my neighbor is an architect who uses AutoCAD. He’s customized AutoCAD beyond recognition to match his workflow exactly—it’s amazing. His answer to my question of why he skipped releases, was that it was too expensive, and then he quoted me a figure in the tens of thousands. When I expressed surprise, he said that the price of the upgrade was tiny compared to the cost of integration (two weeks of otherwise billable time). Ohhh.

    Related to low disruption is bang-for-the-buck. How much complexity, integration, and learning is required compared to the gain in efficiency?

    Anyway, each of these factors can be tracked, recognizing that different companies may have somewhat different priorities.

    Oh yeah. Risk is also a factor, especially with newer technologies. That’s why when I first started in CAD/CAM/CAE, I found that most large companies temporarily ran the software in parallel to their existing workflow. They often said things like “Computer-Aided-Design is not as efficient as our established methods, but we need to become familiar with it, so when the time comes to make the transition, we’ll know what to do and the transition will be smooth.”

    Bill,

    Thanks for your perspectives—I enjoyed your story.

    I agree with your observations on the detriment of incomplete or poorly implemented features and the advantage of being able to selectively implement features (having an off switch). These factors are related to low disruption and high quality. As major features become complete to some level of functionality, all you need are incremental improvements at the pain points that have emerged. I’m not sure about the desirability of a single large install over several small installs.

    The fact is that the major cloud technology vendors cannot afford performance bonds and instead talk about “shared responsibility.” Still, a major failure on their part would result in dramatically reduced business—or no business. For example, read what Amazon S3 does when they decommission one of their servers. The same issues would emerge for CAD software if the power went out intermittently four hours per day. Reliability is the issue here.

    A lot of what you’re describing falls under document management and workflow (which includes notifications and approvals)—a huge topic in itself which mushrooms in large projects. There’s also access control software to determine who has access to what. Finally, I’d like to reiterate that encrypted files are probably more secure in the cloud than on local networks, but you’re right that there are vulnerabilities to local screw-ups as well. Many companies store copies of their critical data offsite in a physically secure facility as a result.

    Finally, I would say that I feel that it’s an important responsibility of software vendors to always invest in the future for their customers, especially in an environment of rapid technology change. It’s inevitable that some “unicorns” emerge. In some cases you barbecue it, in other cases, it might be ok after you saw off the horn and put a saddle on it. 😉

    Dieter

  8. 50 years ago, IBM didn’t sell computers. They rented them. You could only lease the computers and their software (on punch cards). Few companies could afford renting a computer, anyway. Competitors appeared, selling computers and the software, IBM was forced to follow, and microcomputers rise from the end of the 70’s allowed many companies and people to buy computers and software. At the beginning of the 80’s, CAD systems existed but were very expensive. AutoCAD appeard, riding on the microcomputer wave, and basically killed the expensive CAD systems with a (comparatively) cheaper alternative. Microsoft, Autodesk, Adobe and others built their empires upon that new computing freedom. Now, they are trying to take freedom and convenience from the user’s hands, for the sake of profit, using ridiculous marketing lies. It’s Autodesk’s turn to be killed by open source software. It might take time, but it will happen, for the user’s good.

  9. Steve is without doubt correct on number 12. Each release for the past half dozen or so has seen fewer and less impressive feature sets. Is there nothing left to add to AutoCAD? No, that isn’t the problem. I don’t know what the problem is, but as long as I have to pay you $XXX to use the software (or it stops working), how is that any incentive for you to give me a better product? Sorry Dieter, your argument is totally 180° off.

  10. Dieter Schlaepfer

    R.K.,

    You seem to be saying that

    1. The majority of AutoCAD customers want impressive new feature sets in every release.

    2. That subscription-based companies such as Netflix, Verizon, and Better Homes and Gardens would be significantly more responsive to customers if they switched to a one-time payment model of, let’s say, $10,000 and then it’s free after that.

    Would that be correct?

    I’m not going to argue with you, but I just want to be clear on what you’re saying.

    Thanks,
    Dieter

    • Can’t speak for R.K, but from my perspective 1 is nearly right, although I would rather say “genuine improvements”, whether that comes in well-designed and fully baked features, greater performance, fewer bugs, reduced impediments to productivity, etc. You know, the sort of thing that Autodesk used to do in the 20th century and Bricsys does in this one.

      With 2, we don’t need to mess about with analogies that conflate software licenses with incompatible products such as 100% consumption-based entertainment services. We can see what Autodesk has done, and is doing, and we can read what your co-CEOs have said they’re going to do to us. Any analogy you pull out of anywhere is always going to be defeated because we can always refer you back to reality.

      The reality is that Autodesk is providing less and less as it demands more and more, and your bosses have told the stock analysts that this will continue and expand. If you want to convince people that a huge rental-based reversal of that trend is just around the corner, good luck with that. You’ll need more than strained analogies.

  11. Dieter Schlaepfer

    Steve,

    No analogy is intended in my post to R.K. The question in my second point is meant to be purely generic. I’m asking R.K. about his perspective on the dynamics of responsiveness by companies that choose a subscription model in comparison to a purchase model. That’s all.

    If you want R.K. to evaluate a different type of business, how about this:

    Company Alphanet leases network servers and support on a monthly subscription.
    Company Betanet sells network servers and support for a one-time lump sum.

    Disregarding price for the moment, I would ask R.K. which of these two companies are likely to be more responsive to him as a customer and why.

    Dieter

    • Again, no logic-chopping, strained comparisons, theoretical evaluations or hypotheticals are required. We can see the stark reality of how responsive to customers all-rental Autodesk is, right now. Should we be impressed?

      If you insist on a comparison, we can compare that with the reality of how responsive to customers perpetual-committed Bricsys is, right now. Autodesk should stand out like a beacon of customer responsiveness in comparison, correct?

      I’ll leave R.K. to comment on that and your points as he sees fit.

Leave a Reply