User-Focused IT at the U.S. Patent and Trademark Office, with CIO John B. Owens
Scale and inertia are key challenges facing any Chief Information Officer in the federal government. For this reason, successful approaches to innovation and digital government must consider the reality of life in an agency. For this episode, we talk with John B Owens, CIO for the U.S. Patent and Trademark office. Owens explains how his IT organization ensures alignment with end-users, using approaches such Agile development and DevOps.
Video Transcript: User-Centric Information Technology at the U.S. Patent Office
(00:04) Government IT, federal government IT is big and it’s bad. I don’t mean bad as in bad, it’s like bad as in bad-ass stuff especially when you have a CIO who’s focused on Dev-ops and connecting with users and the business.
I am Michael Krigsman, episode 115 of CXO-Talk with my glorious and fabulous co-host, Vala Afshar.
(00:32) I think you should save the big announcement and your scream describing our extraordinary CIO’s who’s with us, John Owens, the CIO of the US Patent and Trademark Office. How are you John?
(00:45) Very good thank you. Thank you both.
(00:48) John Owen, CIO of the US Patent and Trademark Office.
(00:53) How are you all doing?
(00:54) We’re doing terrific. We promise we’re not going to talk sport John, so could you briefly talk to us about your professional background pleas for our audience.
(01:05) Sure, I started my career going to college for computer science and computer engineering at Clarkson University. I became a programmer, worked my way up through several businesses; General Electric, (Mark Myriad) at E-Systems an all part division all the government contractors for the military contracts.
(01:27) Then I joined the little itty-bitty company called America Online back in the day, became AOL LLC, and left there in 2008 and came to the US Patent and Trademark Office in hopes of making a better patent system by improving it’s IT systems
(01:47) Okay, so give us some context about the USPTO as we love to call it. Tell us about the USPTO, the size, the scope, what it’s responsible for.
(02:00) Sure well we have over 12,000 employees. The United States Patent and Trademark Office our background is actually founded right back in the constitution, where our forefathers talked about the protection and intellectual property and at that time patents, trademarks came a little later during I believe the 1850’s, and the basic concept that as an inventor you should profit from your ideas and someone shouldn’t steel them from you and profit from them without you sharing in that profitability.
(02:37) But in exchange you would then make your inventions public and other people would be able to build upon your ideas and make even better, and better and better things.
(02:48) So we handle both patent examination and trademark examination and of course through that examination with the dissemination of both patents and trademarks to the world is part of my job as this Chief Information Officer. Along with the operation and maintenance of all the public and internal facing IT to support that innovation that novation and actually the world brings here to the US for registration.
(03:20)John can you talk a little bit about the size and scope of your IT organization, and perhaps talk a little bit about major technology initiatives that you’re leading.
(03:32) Sure I have about 650 employees. I have over 620 openings, so if people are interested in joining us here at the USPTO, you might want to look on USA Jobs after this is all done because we do have openings.
(03:55) We are looking for other folks to join us with technical skills, particularly computer programmers and engineers just to put a little plug in for us.
(04:05) Did you say 700 openings?
(04:08) 720 positions and 680 something filled, so there’s a good number of openings there and we try to keep bringing in good talent.
(04:28) So it all started with a complete redo of the complete infrastructure here at the USPTO; started in 2008 ended in five years. We have the most modern technologies and infrastructure you can buy. Then in the last few years we’ve started initiatives to re-write all of the legacy applications, so of which were built back in the 1970’s on mainframes of COBOL and AGOL. And some of which were built during the 80’s and 90’s and we’re in the middle of a huge user centered design, agile development practices over the last five years I guess.
(05:08) And of course a focus on dev-ops and rapid delivery on those products. There are a few major ones patents end-to-end, which is a complete re-write of the patents environment and trademarks next generation. Trademarks actually had an electronic processing system before and this is a re-write on top of modern technology.
(05:31) Along with new feed processing systems, brand new web page of my USPTO experience, where you can customize whether you a customer or an internal user and customize your experience type of initiative. Think of it as a widget based environment, or you can add the widgets that you care about if you’re a trademarks person or a patents person or if you’re a big firm.
(06:00) And it’s all built on top of various cloud platforms. We’re running internal red hat cloud, based on cloud forms. We also heavily use the Fed-ramp, Amazon cloud and Google cloud to both host and disseminate our data as well as handle scalability for the end-user products.
(06:25) John, tell us how federal IT or projects in your organization are different from those in the private sector because I’m not sure that everybody has an appreciation for the fact that there are important differences, so maybe give us some of that context.
(06:44)Well to be honest, I’ve worked very hard to make them very much like the outside. We no longer do – well we do a little bit of waterfall type of projects, but all of our new projects are all agile.
(07:01) Because the United States Patent and Trademark Office is fee funded and we are giving exemptions under Title 35 to work around the far or certain parts of the far.
(07:14) We are a little different than most federal agencies. So no tax dollar goes to fund this organization; it’s all fees. And we have certain exceptions, particularly here around the sole source selection of technologies and services, so I’m not the atypical federal government agency.
(07:34) With that I also have certain leniencies in using more cutting edge technologies, although I have to meet all of the FISMA and federal regulations on you know security which is a primary focus, but I don’t have a lot of the same constraints. But we manage our environment through a series of project management best practices that we took right out of industry. In fact the whole dev-op’s movement, which I am a very big supporter of in the federal government, is right out of industry. So I would say that I am hopefully not that much different.
(08:12) That’s terrific to hear. Like yourself we’ve had extraordinary CIO’s on our show and we always like to ask CIO’s how they define success in IT. Can you share with us maybe the certain number of KPI’s or maybe some management philosophy, how do you define successful IT?
(08:33)Well, I take it a step at a time. Here at the USPTO it’s about quality and service availability. This is going to sound a little odd but I would like to have the failures of the United States Patent and Trademark Office IT systems make the front of the paper.
(08:55) That does sound odd.
(08:56) Now most people would say, well why would you want to be on the front page of the paper. I wouldn’t, but I would like those failures to be so uncommon that they’d be top newsworthy. That the system was so resilient and certainly the systems we’re building today, there is no single point of failure. They’re housed in multiple data centers.
They have 24/7 by 365 operations. They are built on you know Red Hat J Boss, you know, totally scalable type of technologies that are out there. There is service orientated architecture throughout and you know, so we’ve taken the best that industry has to offer and every piece of knowledge that we can gain, and a bunch of open source best practices and we’ve really turned that into a robust platform.
(09:44) And I would like the systems here instead of the situation that they’ve been in over the last decade, which is quite honestly fairly poor, to meet the quality level of expectation that people have of a top notch IT environment, which is full availability, full redundancy and it’s always there. It’s like when I was a kid there was a big push, when you pick up the telephone there was a dial tone, not the ding but that was the thing at one time.
(10:15) I think that it’s pretty much the same here. I want the expectation to be that the USPTO will satisfy all of your IT needs when it comes to patents and trademarks, and the dissemination of information and the submission of that data without interruption and that’s the picture of success.
(10:37) So you’re very focused on dev-ops, you’ve mentioned it several times and you have a dev-ops conference going on now as we speak, so let’s talk about dev-ops. So to begin give us a, for those who don’t know a very very, very brief primer of dev-ops, what is it, why do you care about it so much? Why are you so passionate about it?
(13:47) This is a very dramatic cultural shift for IT in the public sector and especially in the private sector, so how do you get your IT department to have the skills and the competencies and culture of recognizing the primacy of that goal, of that responsiveness goal, how do you transform IT itself?
(16:27) John, you have up to 70 openings in IT, what do you look for in terms of skills, experience. Do you look for developers that so get from BitHub and embrace open source and believe in agile development processes? What are some of the things you look for when you’re looking to hire a great talent to help further cultivate this innovative culture, that seems to differentiate on qualities, serviceability, and speed? Can you share a little bit about of the type of talent you look for?
(17:03) Sure, you know there is a plethora of talent with various skills, but certainly over the last year or so it’s been more heavily hard-core computer science focus. We’ve done an incredible amount of hiring in other areas, though enticing computer scientists with degrees, with a love of agile and you know Java development. We’re writing the bulk of our code in Java and J-plos platform and so on and so forth, is all-important.
(17:37) But the most important thing that we really look for is not how many years of experience that they have, because we have a range of positions. But how in tuned and excited people are to tackle the difficult problems to really become a leader and say, ‘I’m going to tackle this and I’m going to solve it’, and work together as part of a team to make you know their ideas heard and make a difference.
(18:03) it’s those levels of motivation that we’re really looking for. So we have both, to put it in government terms GS2210 and GS1550 positions available, the 1550 positions being computer scientists, computer engineer positions, with positive degree requirements.
(18:24) This all gets back to the fact that I’ve been able to drive up quality in the federal government through contracting by providing developers on the agile teams that are Feds not just contractors. And making sure that they are the gatekeepers for the code reviews, because as we know when the government accepts code from a contract, and now the government is responsible. So the other thing that I did was the United State Patent and Trademark Office and my team, who I have encouraged and driven a lot of them from industry are the integrators.
(19:01) We don’t rely on a contractors to do the integration for us. We hire them for their skills and to build our teams, but we place Feds on those teams. We place scrum masters that are Feds on those teams, and it’s a continuous series of training and improvement and the skill sets of the employee that makes a big difference, for which I invest approximately, given budgetary constraints aside, $2 million a year for 40 to 80 hours of training per employee, which you know certification etc., which I very much believe in.
(19:42) Why the importance of including federal employees on these project teams rather than just relying on external contractorsas so many others do.
(19:56) Well, there’s a difference, and I don’t want to sound rude since I was a contract at one point, and again I did work for general electric, (Mark Meredith division?). But, the motivation of contractors is somewhat different than the motivation of a federal employee. I put them on the teams, not only as contributors but as a check and balance against the deliverables of you know what is being developed, code reviewers, you know the technical experts to give you the real quality of the code, the stability of the code. Whether or not the code, we have automated tools to help the code compliant with our policies and standards and coding standards and so on.
(20:45) Do we get the appropriate level of unit tests of the company, because we are really big on automated test as well, so we have a configurations management environment that allows us to do one button deploys to the cloud, including one button that deploys automated tests. So the product gets built, put on a brand-new virtual environment configured with a tool like Puppet, you know it dumped on their and then automated description test run and those description test of course produce a series of results. So we’ve also embrace that, and that is part of dev-ops by the way is the ability to do that rapidly without human intervention.
(21:27) But it’s really a check and balance, and I’ve noticed that you know, in the old days when a developer would come to one of my staff who didn’t know how to read or write Java, and they would make a delivery of course the conversation went like this;
The federal employee would asked, ‘hey, contract, bloody hell are you doing?’ ‘Oh great’.
‘Well your delivering me a product today.’ ‘Yup, here it is.’
‘Well is it good?’ ‘Oh of course it is.’
‘Is it bug free?’ ‘Oh of course, no bugs, never a bug.’ ‘ok.’
(21:57) Then we will install it of course and it would run in production and fail. And then you get to pay the contract again to fix it, it kind of looks like a Dilbert cartoon.
(22:07) By placing a federal government employee with the right skillset right there upfront, you’re able to do that evaluation and know what level of quality you’re getting, combined with agile development you know, where you are producing a product every 2 to 3 weeks.
(22:23) That’s very powerful because now you’re guaranteeing a level of quality that you are going to get when you actually produce the product and put it in production. And that’s a big deal. I wasn’t able to properly find a way to incentivize the contractors appropriate way to do that same thing of due diligence. If you have another method, let me know.
(22:48) Michael, what’s going on here, I mean are we talking to a start-up CIO? If we’re cloud computing, dev-ops, agile development, employee training, speed, quality, or are we talking to a large scale federal CIO. I mean I don’t know Michael what’s going on.
(23:08) Yeah, it’s pretty extraordinary to hear about this when talking about having the best infrastructure that money can buy when you talk about dev-ops, agile. When you talk about your attitude towards contracting, it really is like you’re running a much small organization. So is the fact that you are funded by your users as opposed to federal budget, does that play a role in this, or what’s going on John.
(23:40) I definitely think it does. The Patents Public Advisory Committee and the Trademark Public Advisory Committee, the number of people vary and not all of them show up all the time, but there’s about 10 or so people plus some full time USPTO members that have been appointed by the Secretary of Commerce on those two separate committees. We meet at least quarterly, and they review everything. They have all insight and full access as an employee would in everything that we do.
(24:18) We also meet bi-monthly on the telephone, give them briefings. So they represent our constituency who pay fees. And I can tell you that the lawyers that make up these groups and the users who make up the groups want quality. In fact the amount of money, they’re like ‘yeah, we would give you more money if we can get more’. Right more, better, faster and they do invest.
(24:54) Now I’m not saying we as a federal agency don’t, we also ride the tide the economics of the country, and of course we also drive the economics of a significant portion of the country through innovation. But there is some give and take there, but when we talk about oversight you know, I have those two, and of course I have the other plethora of government agencies that love to check in on us now and again.
(25:23) But that level of oversight by people that are actually making the investment, that are like you’ve got to generate us a result and now, right. Whereas typical federal government you know, oversight is like, ‘okay, well we’ll check with you in a year, have a nice day!’ These folks are like, ‘okay what did you do for me yesterday, and can we drive more faster, better, harder. Show me what you’ve done. You know, it can’t be paper, we want to see it, we want to touch it, feel it. And we want to hear from our constituents and part of those two teams sits the three unions that we have here at the USPTO of users.
(26:03) So they hear right from their own team members 0n (pre-fact and te-fact?) on how the users think that we are doing. So it’s impossible to hide anything, I mean everything is very widely in the open. And that really come a long way particularly you know we just launched the patterns end to end project, and it had been attempted three or four other times in the history of the agency and failed. This is the first time it has succeeded.
(26:34) We launched it, and we’ve trained about 4000 people and the users love it. In fact they are telling us what they want added to it, more than they are complaining about any particular bug or our bugs, there is always bugs no one is ever going to say that there is not. But, we are rapidly fixing those and we are listening to our customers and that user centered design interaction is just building, and I’m looking forward to doing the same thing with trademarks next generation towards the end of the year as well as the fee processing for the external groups, and next year we have got even more to give to the public.
(27:19) So this entire idea in connecting what IT spends with outcomes that are meaningful to users as opposed to just process milestones, where we’ve done this test, we’ve done that test it’s such a strange idea for a lot of IT. it’s hard to grasp that one.
(27:50) Two weeks ago we had Mark Schwartz who is the CIO of the United State Citizenship and Immigration Service, and he talked a little about oversight and he said that the job of oversight can be the easiest to remove obstacles you know rather than an impediment to innovation and efficiency. It sounds like in your role that oversight doesn’t slow down innovation and efficiency. In fact it encourages as Michael said, outcomes tied to tech investments.
(28:18) Well I can’t say that it can’t be used to slow things down. I remember when we first started off patents end-to-end, there was a brief point in time where we had more people investigating, patents end to end is a project than I had people working on the project.
(28:40) So I’m not saying that it can’t go too far, anything can go too far. But I think that when you have the right level of oversight, particularly by the people that are paying the fees, you know to provide the feedback, they are a little more careful on watching and learning, and listening and producing results on a tighter ongoing basis. I mean, this isn’t a once a year review and they walk away, this is a continuous conversation, and every couple of months and I get together at least every quarter. And then the yearly report that goes to Congress as well as the president, where they talk about the successes and failures and the things that they have recommended that we change.
(29:33) I think they each treat it like it’s money coming out of their own pocket because, and largely they represent companies and themselves who do have money coming out of their pocket in funding the agency.
(29:45) So, that level of oversight can be very constructive. I’m not going to say that all oversight is that constructive, but at least in this case I do think it is.
(29:57) Let’s talk about digital government because you’re right in the thick of it. so to begin, when we talk about digital government, you can even bring in open data if that interests you. But when we talk about digital government, what exactly do we mean?
(30:14) Well I think the term ‘digital government’s’ used to describe a lot, and I think if you had asked five or six people that question, depending on where they’re from you would probably have seven or eight answers. But let’s go with the basics.
(30:29) For use, it’s about fully embracing the electronic world, right. We would like all the processing of patent and trademarks to happen electronically. I certainly would like to get rid of you know legacy stuff like fax machines and snail mail but that would never happen.
(30:50) But it’s about fully embracing that but also fully embracing it from the beginning of the process to the end. So that doesn’t just mean electronic intake, but it means electronic processing, electronic publishing, electronic disposition, and then electronic hosting of the patents and trademarks to the public in a way that they can consume them in a meaningful way.
(31:16) And that’s actually a lot more complicated in retail than you might think. So we have a lot of great intake systems today for both patents and trademarks, which we are going to revamp to make even better and friendlier. But sometimes the people in the middle still rely on faxes and paper trails and they don’t rely on email correspondence or as much video chat but we offer those things but they don’t use them as much.
(31:44) We have to break down those barriers and make them more accessible, you know that’s very important to us.
(31:50) But we’ve had recent changes over the last few years, for example we had the e-publishing of trademarks which the EEOG, which is a product that’s brand-new over the last two years.
(32:08) We used to publish every I think it was every Tuesday two big multipage volumes to the government printing office of new updates to trademarks, then it was a huge deal. Now, you can go onto our site, login with your iPad or you know your tablet or computer, and while you are on the train or subway it you can log in and look at this week’s publication or all the publications back for the last couple of years. Search them, correlate them, tagged the ones that you want to investigate or talk to. Even click and send this, hey that’s mine, I don’t like the fact that someone is doing this, and we have made all of that easy and accessible.
(33:01) So that’s the one part of the digital government. The other part is mostly about data. You know, people talk about big data, but making all of the data are available but not only in bulk, but through a series of API’s and then build upon those API’s a series of systems, which allows users to interact with them, no matter what device they’re using or operating systems through the web.
(33:25) And that is something that we’ve done a little bit on over the years, all of the data here at the USPTO is public. We’re hosting it ourselves, we were using several other companies, first Google and then Re-technology to host it for us. We’re hosting it ourselves this year, and then we have plans over the next few years to produce this series of rest-based API’s to allow access into that data that is now hosted in the Amazon and Google federal clouds.
(33:57) And then building some user interfaces on that, for example for public patents search, to allow people to look at all of the patents at all of the repositories that we have here from all of our country, but also our partner countries which is a big deal.
(34:14) And John why is this so hard? Why is this difficult to do? Give us some insight for people who are outside the government and simply say, ‘this is not a big deal. I’ll simply go to the website’, and who don’t have an understanding in what’s actually involved. So why is this so difficult?
(34:32)Well, as far as the USPTO goals, particularly for patents, the examination system that I inherited when I got here was built for about 2500 examiners, and we have nearly 8000.
(34:46) And the public facing environment that allows the user to query patents that are called public pair or private pair if you just wanted to look at your private submissions pre-publication, were tied to the examination system. So every person in public that logged in to look at something took out the same resources out of the same system that an examiner would. So it doesn’t take a huge amount of time to realise that an examination system built for 2000 people plus the public soon, when we’ve hired 8000 since it was originally launched until now, it doesn’t handle the load of people to well, particularly if a couple of data miners were to go after it, which is why we had to put in capture for the public.
(35:43) So one of the reasons that it was so difficult because everything was intertwined. You know, the systems were still built based on the old automated information system with a ton of single points of failure, that you know the pair system resides on you know one or two computers that are not redundant, they’re just handled separate parts and a fixed data store, and that was it and that’s all I had.
(36:13) So I’ve had to keep all that up and available and operating, and tried to scale it as best I could, both by stabilizing its software, as well as buying legacy hardware to shore up those platforms that quite honestly hadn’t been produced in years.
(36:32) At the same time, we write it right, because you know, not on that system but on some trademark systems, just as an example you know how expensive it is trying to find COBAL and AGOL programmers and then the EMT staff that you have to keep on staff to resuscitate them. Because let’s face it, if you’re going to lose one you’re not going to find another and – well that was supposed to be funny.
(37:01) We have a question from Twitter, Vala and I we’re pretty… well actually I’m going to take that back, Vala’s pretty dry. I know me and him are going to have words after the show. But we have a question from Twitter, from Ryan Faye, who asks, and he wants to know, what safeguards the Patent and Trademark Office plans to use and to keep all of this data secure without limiting accessibility?
(37:37) That’s a good question. So we do keep the data of record here, and when we publish it out to the cloud we publish it one way. So the API’s we have will push it to the cloud, we use the cloud like a giant cache think of it like that.We also have a backup to alternative sites and we employ the best, and the brightest in the federal government security and we meet the standards, well published through security audit and so on and so forth.
(38:08) For the first thing is, is that the only people that can really change it are those in-house, and we have several layers of security to make sure that you don’t have access to it unless you’re here and you have the role of being able to change it.
(38:23) So even when someone comes into the USPTO and says we would like to change the assignment of a patent, or a patent that was sold from one company to the other, they actually have to go through an electronic process which involves someone verifying it and making the change today making a safeguard.
(38:42) Because no one wants to wake up one day and have all of the patents owned by Bill, right I mean that’s just not a good idea. So that’s the first series of safeguards. But you have to counter that with the fact that except for a very small number of patents that are security related, that are kept as paper and managed and reviewed by the military examiners, these are things that we wouldn’t want to publish.
(39:16) Everything in the patents world, unless it’s in that group or a small group because of court order, they are all published after 18 months and every trademark is published the second you hit the enter button on submittal.
(39:30) So the bulk of our information is public, but the focus that I really have is protect the stuff prepublication for patents. The stuff that secured get turned into paper and handled through our military connections. And the rest is open to the world and all I have to do is to make sure no one has the ability to go in and change it. And so, that is a much less complicated model than some folks have.
(40:01) John, you said that you work with the best technology; does that mean that you work with start-ups as well?
(40:07) Sometimes, you know we have worked with some smaller companies and I want to call them quite start-ups but smaller companies. I know that on some of the contract that I’ve had they were small companies at the start and large companies by the time we were done with them.
(40:25) We also sometimes do of course RFI’s and RFP’s for technologies. There has been a long-standing need to find a technology that can look at pictures and identify things, and that technology is coming along. So we have certainly done that investigation for a number of years, and called up people, asked them to come out and demo their wares and talk to us about their technologies and what they do.
(40:55) So you know although we are exempt from some parts of the far, we do have to follow others, so there are importance as to getting procurement involved early and dotting the I’s and crossing the T’s, because you know I don’t look good in orange!
(41:32) Well you know we’re almost out of time, so finish us out with some advice from your perspective on how IT leaders, either in the government or outside the government can remain relevant in this digital age. What are the keys to a CIO and IT remaining relevant?
(41:35) Keep producing. If it doesn’t work, change what you’re doing and try again. If you feel again, that’s fine just keep trying but you’ve got to produce. You’ve got to be in tune with the customer, you’ve got to be in tune with the business and you have to produce a product and a quality in a timely manner that the customer wants.
(41:58) Embrace ideas that you’re not familiar with, use a centered design, agile development, dev-ops. Look to produce a quality product, you know EVM, which the government regularly uses is a very poor indicator in my opinion of quality, because it doesn’t account for quality.
(42:16) And you can be on budget and release a product and still produce something that no one wants to use and that’s a serious problem. I think we have a lot in the federal government to learn from industry, and it pains me greatly to know that federal government spends about $80 billion a year, and we are not the leaders in IT. We should be generating these ideas. We spend more money unless your Google than almost everyone else combined. And if you think about it that’s a lot of practice and we should be producing a heck of a series of good products and services for the people in the United States, and in my case the world.
(43:03) But we’re not going to get there if we continue to do things the old way, and in my opinion one of the greatest things that happened since I’ve been here was formal federal CIO Vivek Kundra allowed us to use agile, and that shook things up and we need more of that. We need more people saying, we can try things and work with this stuff that the industry does, and if it works continue to refine it, and use it and adapt it to the federal government. But it’s all about producing that quality product in a timely manner and listening to your customers and producing something that they want to use.
(43:44) And if you want to remain relevant keep up your technical skills and produce, and that’s the best way I knew to be relevant in private industry and it’s the best way that I know how to be relevant here.
(43:55) Okay, great lessons on relevance, well Vala I’m afraid sadly it’s that time.
(44:02) That was a very fast 45 minutes john. Thank you very much and congratulations on the first dev-op conference in the Washington D.C. area and thank you so much in taking the time during the conference to be with us on CXO-Talk, you were fantastic.
(44:16) Well than you to both. I appreciated it.
(44:22) We have been talking with John Owens, who is the CIO for the US Patent and Trademark Office on episode number 115 of CXO-Talk. I’m Michael Krigsman and I wish my co-host, Vala Afshar a what – a great weekend.
(44:02) Thank you Michael.
(44:43) Everybody come back next time. One last thing, subscribe to our newsletter. It’s great, you’ll never miss a show. Subscribe now. Thanks everybody, and John Owens thank you so much, bye bye.
E-Systems – bought by the Raytheon Company?
General Electric www.ge.com