What I was surprised to learn at SolidWorks World 2014 about CAD data management

Well, it’s been almost two weeks since Solidworks World 2014, and after enough times of saying “At SolidWorks World I learned...”, or “I talked to a guy at SolidWorks World who said…”, I thought I should probably write a post about it. But be warned: this is not your typical SWW14 recap.

http://michaellord.me/2014/01/29/solidworks-world-2014-general-session-day-2-photo-journal-part-1-sww14-solidworks/

Photo from michaellord.me

Given that as Director for Products at GrabCAD I’ve been working pretty heads-down on our GrabCAD Workbench for the past two years, I spent a lot of time at #SWW14 attending sessions on data management and talking to other SolidWorks users about data management. What I learned may surprise you.

Data management is boring

There’s just something about saying “data management” that turns off a circuit in most people’s brains. The sessions on modeling tips and tricks were always better attended. People have real, serious problems with having duplicate versions of the same file, sending the wrong versions to manufacturers, overwriting each others work, etc. but it’s not a problem anyone really wants to solve. As one person told me:

“We have thousands of files from 20 years of projects that are duplicated across network drives, multiple hard drives, etc. We hardly ever re-use parts from old projects because no one’s sure where they are or which one was the final version. Even our CAD admin quit after just 4 months because she couldn’t take what a mess our files were in. My boss doesn’t really care, and it sure isn’t my job to deal with it. I’m here to design.”

The “easy” options aren’t great

I attended a session called “File management for beginners” that attempted to teach tips for file management. Some of the key lessons from that session:

1. “Copy everything to a 2nd hard drive so you always have a backup. Periodically zip up all your files and burn on to a DVD so you have a 3rd backup.”

Because I have such great memories of waiting 10 minutes while burning a 700MB CD back in 2000 that I’m really looking forward to doing it with gigabytes of data. This is 2014, who burns discs anymore?

2. “Come up with your own complex part numbering system, preferably with a bunch of numbers and letters for categories, color, purchased/made, etc. Keep track of what the next number is in an Excel spreadsheet.”

Example:  100-1000-P-0-A  {category}-{number}-{purchased}-{color}-{revision}

5 years from now when you’ve left the company, isn’t everybody going to wonder what the hell all these numbers mean?

3. “Keep track of revisions by either exporting an eDrawing for each revision, by using a separate folder that only contains revisions, or by building your own Access database.”

Have you ever built an Access database? I have and it sucks, I 100% do not recommend this.

4. “Distribute files by sending eDrawings or using a Pack-n-go. Send them via an FTP, Dropbox, Box, or even a flash drive

This is pretty common, but installing eDrawings, setting up an FTP site, or waiting two days for a flash drive to arrive in the mail is a headache nobody wants.

Most people fall into one of three categories

I noticed that while everyone is different, there seemed to be three general types of behavior. I present The Innocent Bystander, The Getting-Educated Worrier, and The One Who Has It All Figured Out.

THE INNOCENT BYSTANDER:

A lot of people I talked to don’t use specialized tools to manage their files, just a shared network drive and good karma. They try to do some things to minimize issues, but there’s no way management is signing off on buying a big PDM system that costs thousands of dollars to implement. As they don't generally have an IT department it’s not really in anyone’s job description to deal with the problem of data management.

And yes, the reason I make these sarcastic comments is that I was the Innocent Bystander at my last company, Gemvara. I worked at www.gemvara.com doing CAD, rendering (most of what you see on the site), and also trying to manage the thousands of CAD models we had. I felt pretty good when I worked with IT to get our network drive doing automatic nightly backups. I thought my part numbering system was genius. I thought it was pretty cool when I spent weeks building an Access database so everyone could mark when they were working on a specific model. I thought it was a pretty clever to copy all our files to Box so we didn’t have to mail a flash drive to our manufacturers. And then I realized that all of this was taking time I should be designing!

THE GETTING-EDUCATED WORRIER:

I found a different group of people, usually at slightly larger companies, that knew this wasn’t the best way. They are researching PDM and PLM systems and learning all the features of those systems. They are trying to put together a picture about how this would get implemented into their company and how it would work. The Getting-Educated Worrier has a pretty good idea of what’s up but still has questions like:

  • Who’s going to administer it? What people need to sign off on the decision to buy this? Who needs to get trained in this?

  • Workflows seem really useful, especially in ensuring you’re releasing the right version for tooling and manufacturing. What’s the best way to set them up?

  • How quickly can I get “where-used” reports and good search set up?

  • How do I integrate with an ERP system? Not for today and not for tomorrow, but I’m sure this will be important some day when we get SAP.

  • How do I keep BOMs and versions correct?

At my last company, opening a 2nd office in a different state started the shift from the Innocent Bystander to the Getting-Educated Worrier. Pulling large models from a network drive over VPN is like pulling teeth without novacaine. We got demos from the big vendors and despite how customizable all these systems were, they seemed to have a thousand features that we didn’t need and precious few that we did. Plus they were expensive! So ultimately we decided to build our own custom system - essentially copy the aforementioned Access database into a real software system. Building that system took more than two years and turned out to be a real headache. And yes, for those who are curious, this is precisely why I came to GrabCAD - to save the Gemvara’s of the world from repeating these same mistakes.

THE ONE WHO HAS IT ALL FIGURED OUT:

Yep, this is the person that everyone else is jealous of.  Years ago, their company realized that this could become a serious problem, and they actually put time + money to solve it before things got bad.  This is the person who:

  • implemented a solution in 1994 before it was even cool

  • put in a dedicated T3 between different company sites so engineers don’t need to get a cup of coffee while waiting for a file to pull across from another site

  • has nightly/hourly replication between archive servers so files are always up to date every morning/hour

  • wrote their own web portal in 2004 so engineers at the company could pull files without being in the office, back when “cloud PDM” was still just a party joke

  • has integrated to more ERP systems than you’ve ever heard of

And yes, I did talk to this person and he’s done a damn fine job. This guy is like Tom Cruise in Top Gun. He was the envy of everybody at the lunch table. He did admit “I’m not a CAD guy anymore, I’m really an IT guy”. So it’s not all roses, even for him.

Conclusion

SolidWorks World was amazing - I talked to a huge number of people at all different stages of dealing with data management. Fewer than I expected were happy with where they were, or what they saw as their options.

Of course I’d love to tell you GrabCAD Workbench will solve all of your problems, but it won’t. You’ll still need to figure out a strategy, you’ll still need to get people on board, you’ll still need to make changes as you grow. But if you’re an Innocent Bystander, I highly recommend you think about getting going before your data problems get worse. If you’re the Getting-Educated Worrier, I invite you to worry about a few less things (hardware, upfront costs, and ongoing maintenance to start). And for the One Who Has It All Figured Out - go have a beer, man, you deserve it.

And congratulations to SolidWorks for throwing such an amazing show, and building such a passionate and amazing set of customers!

  • http://www.grabcad.com/hardi.meybaum Hardi Meybaum

    Grant, I really like this post. Very well written, there is definitely an open slot in the marketing department when you get bored with the Product job.

  • Lou Timperio

    Grant,
    Yes , great post .
    Well written and interesting.
    You are right about the classes,everyone likes the tips and tricks better than the data management. I will have to check out the workbench soon.
    Have a beer..
    Thanks
    – Lou

  • http://www.jvh.com Ed Rose

    Hmmm, I quickly skimmed your article and thought it was interesting.
    When looking at data management, you might want to try a new idea. It is so simple I wish more people could grasp it. A part number is just a number. If you are building intelligence into your part number then you really don’t have a data management system.
    I struggle with this all the time when trying to teach people how to use Autodesk’s Vault product. It is a database that tracks all your file properties so you can search on those properties to find anything.
    Want to find a file that had widget in the description, was made by Joe, between the dates of 12/12/12/ and 12/19/12? you can do that with Vault and it is easy.
    It amazes me how difficult it is to get people past the 8bit DOS thinking of how to structure your data in a 64bit relational database world.
    If you want to see some really exciting stuff, check out Autodesk University next year or check out the video clips online from this last year. It might give you a new perspective that I have been trying to explain to my customers, “Engineering is the start or the center of all design communications, the best solutions will focus on how to effectively electronically communicate design information in and out of engineering”.
    Just some rambling thoughts.
    Oh, and I am glad you had a good time at Solidworks World.

  • http://www.textualcreations.ca/ Don Cheke

    Great article – thanks.

    I am the ‘librarian’ for one of my CAD clients and we went through the ‘getting-educated / worried phase a couple years ago (I am still worried). We had looked at SolidWorks PDM Enterprise when we started looking and nobody wanted to spend the $20K it would have cost us to set up and get it implemented. As such, I was tasked with setting up something like you describe (minus the access database)and to ‘manage’ things. I am in a different country than my client and this makes things more difficult since I can’t just pop in to the factory whenever I want to collaborate face to face. I feel that we have many issues with how we are doing in this regard – mostly with having someone check and approve drawings, but that is a whole new story of its own.

    The question I want to pose to you is: Can a manufacturing company start to use something like Workbench (which seems to be fairly new), put all the time and effort required to implement it – and be confident that Workbench will be around next year – or five or ten years down the line. I know that none of us has crystal balls to know what the future holds but there is some security (if only imaginary) in going with something created by a provider that has been around a good many years. Can you comment on where a user of Workbench would be left if Workbench were simply to stop being available or the makers decided to go a different route.

    Thanks,
    Don

    • http://grabcad.com Grant

      Hi Don-

      You ask a great question about longevity and corporate support. I think this is a question for all providers, not just younger companies like us. There’s a fair amount of discussion in the CAD space right now about how the big CAD companies plan to migrate users to new kernels and what that means for all the existing data. So ongoing support is an important topic.

      For Workbench the simplest answer is that you still have your data. One of the great features of Workbench is our desktop sync, which replicates CAD files to the desktop of everyone on a project. If the cloud project were to suddenly go away, that CAD file is still on your desktop. You don’t lose any data, it isn’t trapped in some database you can’t get access to. It’s right there on your PC.

      The other benefit of a cloud solution like Workbench is that the “time and effort” to implement is actually quite short . Check out this quick video that shows how you can set your project up in less than 3 minutes. This isn’t a three month ordeal that involves a dozen departments.

      And of course I’ll add that we don’t plan to take Workbench away – we’re committed to helping engineers build great products faster and we’re seeing huge numbers of people signing up!

  • http://designandmotion.net Scott Moyse

    Spot On! It’s tricky all right.

  • Rosanna

    Incredible! This is really nice and informative post. You are having great idea to express your view. I loved your way of expressing about good data management with example of Garb CAD work bench video. I saw this video; this video will really help me about better data management with my employees.

  • Dale Stoner

    I just had to ask you if you ever made a “pivot row.” This is something I found in the 90’s helping my son implement a database for customs forms for a tomato and pepper importing company. It is a way of accessing an EXCEL spreadsheet from Access. My technical communications time was in the 70’s moving from carrying suitcases of cards between airports and facilities to a dedicated data line and overnight disk to disk transfers. Yeah! controls hate child windows with string names that are not in arrays like xeryzx(a,b,c). Don’t forget ergnog when it is time to summarize in a single word all the one page memos.