Tips, Tricks, & How to: Become an Elite DevOps Organisation (Cloud Next ‘19 UK)
How's. It going all. Right. Nicole. How are you today I'm good how, are you I'm wonderful, wonderful, so happy to be here it next you asked me - yeah let's talk about tips. And tricks on becoming an elite DevOps, organization. Sounds. Good to me all right let's do it I mean. But. Like elite, devops, organization. What, does that even mean. We'll. Get to that in a minute first. We want to remind you that your feedback has really greatly appreciated. We appreciate, first that you're here, and, listening. To the content that we're sharing maybe maybe, it's sport spurning some questions that you have in your mind how can you apply what, we're about to talk about to your own organization. But, also it, would be really grateful we would be really grateful if you share feedback on the, talk itself so, the, surveys will open up in 20 minutes but in 20 minutes time you're gonna forget that we mentioned this because you're. Going to be learning about becoming, an elite DevOps performance, organization, we're so exciting, yes so. We'll remind you again at the end yeah, that's, a good plan yeah okay. All. Right so first hello my. Name is Nathan Harvey I am a developer advocate working, for Google cloud the, place where I focus, is on the DevOps and sre community, so I love to come out to your team's and understand. How are you doing DevOps within your organization. How can we help you how can you take those DevOps, and sre practices. And principles, and, apply, them to what you're doing on Google cloud, and throughout your organization, and I'm. Nikole for scrren I have, been so. Sad man I've been a developer for. A really. Long time a minute, for, a minute and then I've been doing research, for about a decade if, anyone here is heard of the state of DevOps report or the book accelerate, yeah. All, my people okay so that's me I've been leading that that, research in those investigations for. For. The state of DevOps report for six years and the related research for over a decade now so. Speaking. Of if anyone, wants to check that out it's at cloud google.com. Slash, DevOps. So. That's. That research and when we were talking about elite performers, we, dig into that, in the. Research, so. The. Report has always focused on four, key metrics, and being. Able to develop and deliver software with speed and stability and, how, that enables. Us to deliver, value.
And It's interesting because it's not just our research, that talks about that other researchers have been finding it too so Erik Brynjolfsson at, MIT. James, Besson out of Boston University and, for. The last few years others. Across the industry have been using these four key, metrics, speed. And stability to. Track. Software. Delivery and even, technology, transformations. Or DevOps or, whatever. Word you want to use in your organization, find and replace write digital, transformation. To. See how well their organizations, are doing so. These, four key metrics I keep saying speed and stability this. Is how we track them so its lead time how long it takes to get from. Code. Commit to, code running in production sorry so when you said code commit that's, committed. Where, Code. Commit the first time you commit your code so it's not just like, we. Commit to the idea of code. Only. I love, that question the. Reason we talk about code commit like code, committed, into version control is because, that's. Where so code commit to code running in production that's. The most, predictable. Part, of. The. Code process. This, is the part that's most predictable. This, is the part that we, can really squeeze, the most efficiencies, out of this, is the part that's very highly automated all, right, if, we. Back, that up and talk about, ideation. And, together. Things that we can think about right so it's like art the first idea right, the first time we talked to customers products market fit which features, are we going to build still. Super, important, however. That's. Kind of an R&D process, right, I expect. This to be highly. Variable so. I still. Want to measure it but, I want to measure it very very, differently, and so. Still. Important. But it's just different, okay, so so, that's not right. So it's that code I've written the code I've committed, it to the repository. Yep timer starts, timer starts, and when does that timer end just so I'm clear when. It hits production, okay, or. Some. Will say like a pre prod environment, so if I'm developing. An app I'm building an app I'm putting. It into the App Store the. App store isn't gonna let me, push. Code, every. Minute or every hour also. As an end user I'm gonna be real pissed if you make me update, all, the, time so, if I have like a pre prod environment, and I just push that once a month that's. Fine okay that that's acceptable okay. The next one is deployment, frequency deployment frequency how often do I push code, those. Two are gonna be highly correlated right those are pretty similar, now. Those are my speed metrics my stability metrics. Are changed. Fail rate when, I do. Is push code yeah so, this is when I push, code to production and then expletive. Right. Related. Okay it's not just if it comes all the way down it's, if. It requires any kind of attention right get, a human involved get a human involved yes right don't have to patch it do I have to pay attention to it don't have to give it's love and. Then. When. You do, expletive. How. Long does it take to get it back up now. In, the best-case scenario. That.
Lead Time and that time to restore service are going to be very very similar because. In the highest performing teams their, regular change process, and their emergency, change process, are, almost the same in. The. Lowest performers, we see them differ. A lot because. Right. Sometimes, an emergency, we skip, tests. Because. You just have to push that change out because we have to react to an emergency however in. Emergencies. We probably don't want to be skipping tests. So there's, an emergency, I skip the tests I push to production that's, never caused anything, to get. Yes. Now. These. Are our four key metrics I will. Note that last, year in 2018, s report and the Shi 29 teens report we also talk about service, availability. We. Found that, service, availability, is a, better, predictor of organizational. Performance, things like profitability, productivity. Market share effectiveness. Efficiency customer, satisfaction, right. So what's, availability. Making. And keeping promises to our customers, so this makes sense right so sometimes. The joke is day one is short day two is long so. It's how well we can keep our services, up yeah. Okay. So. Skippy. With the clicker sorry so. Some. Of the key findings that we have one, is that the, industry, continues to improve particularly, among, elite. Performers, so. People. Keep asking me is the stuff ops thing still a thing yeah are we done it's we've been talking about DevOps for 10 years and frankly, Nicole I've been doing DevOps since before there was a word for it a hipster, right. It's. It's been around forever is this real or sometimes, like when I go to enterprises and talk to executives, they're like is this a safe investment, is this, real can I still do this thing do y'all ever hear this like is this real can, actually put money into this it's, real, analysts. Have been saying that DevOps has crossed the chasm for. A while now so, this is what we found when we looked at the 2018 report, there in gray elite. Performers, were seven percent of our population there, were a subset. About. 48% of high performers, this. Year the elite performers, broke out into their very own group, and they almost tripled. Now, they're 20%. Now. The. Low performers, shrunk. They went from 15%, down. To 12%. Now. Take a look at that medium performance, group it got bigger some. Were, from low performers, getting better some. Were from high performers, falling, because remember that high performance group used to be high Plus elite. So. Why. III. Think I know why because I, set. Out to do the DevOps and I. Made it I've done the DevOps and so now I can sit back and relax right, yes so. Some is from people, who've decided they have arrived they are, they are done and they.
Stopped Getting better and then. The. World keeps moving on, the. World progresses, another. Is because. Organizations. Struggle with complexity, and they don't adapt, they don't get better right. So that absolutely, happens so the interesting thing is that this analysis, confirms, what many and met what many other analysts, have said right so DevOps, does cross the chasm. Great. So. You. Know this word DevOps has been around for a long time you say it's here to stay but what exactly do we mean when, we say the word DevOps. Let's. See I bet. In this room we could come up with roughly. 350. Definitions, of the word DevOps well at, least at, least easily, easily, and the truth is DevOps is very different, we're depending, on where you work the teams that you're working on and the things that you're trying to accomplish but. At the end of the day we provide, this one definition, or a sort, of a framework for it it's a cultural and technical movement, that's focused on delivering value by. Simultaneously. Improving velocity, and stability, and, DevOps, really, encourages, cross team collaboration, shared. Incentives, and responsibilities. And an. Environment. That celebrates, learning, I like, it it's pretty good. So. We. Also in this report found that delivering, software quickly, reliably. And safely is at, the heart of a technology transformation. And it's. At the heart of organizational. Performance so it is definitely true we're all becoming technology, organizations, we're all technology-based. Companies. And so getting better at delivering. Software safe, and and, fast, is important, to how we work yeah absolutely and, so. When, we look at this report Nicola you talked about the four metrics but you kind of boiled them up into two top level things we, have speed on one hand and stability on the other and so, I, face. This all the time with, people that I'm working with and in fact organizations. I've worked in before what. Should we do we have to decide do we want to be fast or do, we want to be stable. And that's. What your research bears out right fast or stable, I love, you mentioned this because for years, there was kind of this like rumor that was whispered in dark alleys right like you had to have one or the other but we. Actually, don't. See trade, offs for six years in a row we see that speed and stability go together and many. Times I say that people think that that, only means that they go together at the top-end speed instability, or only things that elite performers, get together, speed. Stability also, happened together at the low end so, like, low performers, suck at all of it together so. What that means is that when we make smart. Investments in, technology, and. Process. And cultural, capabilities, and metrics and monitoring capabilities, it drives, performance.
Along. Both speed and stability they. Enable. Each other they move in tandem and the statistical, analysis, shows us this and. I'll say that we do this analysis, completely, independently, we start over every single year just to make sure like. I'm not missing, trends, so. When I compare, the highest. Performers, these elite performers, and low performers, this is what we say, we. See 208, times more, frequent, code deployments. 106. Times faster, lead time from commit to deploy, over. 2,600. Times faster, recovery. From incidents and, seven. Times lower change fail rate and. I think that paints a really compelling, story in, terms of what, it enables us to do because we, are technology, organizations, now that's, true but I can also take this right to the CEO and talk about these metrics, exactly. And sometimes, when. We talk about these metrics to our CEOs, or our CIOs, or our CISOs, we can say. You. Know it's about delivering, features, but it's also about patching. Our infrastructure, it's. About turning. Over. You. Know. Patches. And an IP right. Because, that makes us it makes it harder for us to hack it's. About keeping up with, compliance and regulatory changes it. Really is this holistic, picture for, what our organization, can do and. You. Know if we want to do these amazing m/l, and AI plays, for our companies we. Have to be able to deliver, them, and that's. What this enables, for organizations. So. You. Know again for bonus points we have availability, and again it's about making. Whatever. It is that we build, available. And accessible, to. Our customers, in reliable. Ways and. It's. About delivering, and also, about how well teams, are able to measure and capture. Availability. Targets and then learn from outages. So. Again it's kind of this holistic, picture of. Availability, yeah, and this is where we also start to get into some of the overlaps between DevOps, and sre and how, they actually work together and, so we can look at some of the SRE practices, that we have within Google that we've really started to share with the rest of the industry and hopefully you're, starting to pick up some of these practices as well of, course we have the fifth nine lounge where you can go and talk and learn more about Sree, but, at a very high level sre helps, us to measure the availability. Of our services, and when, we take an SRE approach, to this there's, a couple of things that are important to keep in mind the, first thing that's important to keep in mind is that we want to take an approach to measuring our availability, from, a customer's. Point of view I know, in, my career I've been woken up many times because, a cpu load was, spiking and a, customer, no customer, cared no customers, cared about that at all so, the, sre practice, really has us looking at what are those critical user journeys, that our customers, are on and how, do we put accurate, and, responsible.
Metrics, Around those, so, we start with a service, level indicator, and it's exactly what it says it's a level of service that we're providing to, our customers. What are those indicators let's, start measuring those once. We understand, what those are we then set service, level objectives, so we have a metric, let's set a target, for that metric and then the other S word that you maybe have heard of before is the SLA, the service, level agreement, now, typically, this is a business, agreement that your company has with, some of the businesses, that it might do work for and I don't know anything about your, business, but, I am pretty certain that your business, is not in the biz this of paying, your customers, and that is usually the result of missing your SLA, you're giving refunds back to your customers, as a technologist. I want to stay as far away from that as possible I do not want to be responsible for giving money to our customers, and I, want to make sure that we're giving them the best experience that, we can to, meet our service, level objectives, which, are tighter than our service level agreements, yep absolutely. So. I, like. This one yeah, for the first time you see that industry, matters so you've been doing this research for, six years plus and this. Year industry, matter as well I'm. Gonna just make some guesses here in a corner financial. Services in government look. These are highly, regulated industries. Clearly. That's going to have an impact on performance and they're going to stand out in the research you sound like all the people sliding, into my DMS all the time saying, like help me tell, my story because we, were highly regulated and, there's no way that all of your research says that high performance is possible and. I. Love that guess but. The. Industry that was statistically. Significantly. Different is, retail. Huh. Retail. I wouldn't have thought that, retail. Would be slower. Retail. Isn't slower retail, has higher performance. And it's, the only one that's different technology. Isn't different really. Better or worse and. None. Of the other industries, are different either now this only shows five we actually, tested, about, a dozen industries, that just made the slide look super messy so. How, does retail, get it right like what are they doing within retail, that helps them become elite performers, here's the thing that's super interesting I loved that you asked that because if we think back over the last decade.
Or So, there's. A trend that's been happening that people have called the retail, apocalypse. It. Has been so. Unbelievably. Cutthroat. To be a retailer. It's. So competitive, that if you can't keep up you are out of business, so. Many retailers, have been shuttering. It is a sounding familiar at all, right, so. What retailers, have had to do is, embrace. Cutting-edge. Technology. They, have to do a/b testing. To understand, customer, behavior they. Have almost no margins, so they have to automate, to. Embrace efficiency, and also. Have. Consistent. Repeatable. Auditable. Processes. They. Have to also be, in the cloud to, cope with huge swings in demand so in the, u.s. we have Black Friday it's. The day after our Thanksgiving. Yes when everyone. Swarms. The internet and buys we do all of our Christmas shopping it's basically, 24, hours there. Is no question of if you're in the cloud it's like. Of course you're in the cloud yes and then, everyone says oh we're highly regulated. Retailers. Have to process credit card transactions, of course. They're regulated. Yeah. What's. Our next finding, Nicole okay. The. Best strategies, for scaling DevOps focus. On structural. Solutions, that build community so, often I get questions saying okay, okay okay but how do i scale, I'm in a large enterprise I'm not in this cute little tiny start-up I have, to make. The devops work with more than five or ten people right so we did this analysis this, year and and. Here's what we found we. Have to have community, structures, because. This, makes. Your. DevOps, investments, and practices, resilient. To things like product, changes, or reworks I don't, know about you but like big companies love I think they like Rea orgs more than they like money. So. Things, like communities, of practice right. Because these are resilient. Although these also still require resources. Grassroots. Efforts right, DevOps, it. Started, at the grassroots, it's. Grown up from there I also, love this idea of communities, of practice I see so many large organizations that. Build these centers, of excellence and, I. Really. I really feel, bad about those centers of excellence. Imagine. If your organization, has a center of excellence but. You are not part of it where. Do you sit. You're in you're over here in the, lake of mediocrity, you're. Not in the center of excellence no that's terrible. Also, the center of excellence we tend to look. To the center of excellence to pass down the excellence, from on high, they sit, in this little ivory. Tower, figuring. Out what is excellence, and how do you do it and they push things, along where as a community, of practice is, very, inclusive, everyone, can be part of this community we, are truly going to figure out what works best not, by talking, about it not by reading a paper but by going and doing the work so that we can figure out what works best in our organization, and then use our grassroots, to help spread that throughout, the organization, which makes it more efficient, I mean has anyone here heard the term which is almost a cliche by now breaking down silos. Yeah. No no who here hates just raising their hand. So. The, challenge, is that by creating a center of excellence, did.
We Just make one more silo. -. Kind of yeah right so, the challenge is that you've created one more silo which then many, times makes, it harder to disseminate, the knowledge that even if you have experts, it makes, it so hard to get all that information out right, and and, now we have research that even backs this up and it shows exactly what. The high-performing, and elite performers, are doing that, are helping them to stand out as they spread these practices, throughout their organization, yep so, we also see that proof of concept is a template and proof of concept is a seed and sup being a great strategy, so what's. The difference between those two yeah, proof of concept as a template is where you, have like, little Tiger teams building out pocs and then, you kind of stamp them out right so you you end up replicating, them out POC. Is a seed is you have that POC, team and then, you break off you kind of break up that team and embed them in other teams so, there are a few different ways to kind of do that and we've seen, that work several different places we, wondered if there were differences there weren't there, could be, cultural. Preferences. Depending on your organization, sure I'm, sure you've seen that absolutely a different place absolutely so. What, works and what doesn't as we just mentioned communities. Of practice. Bottom-up. Or grassroots, training. Centers are also called dojos, a few, places yeah. And those we, find that they just they don't have consistent, results right right some of them can work really well the, challenge, again is that you may have inadvertently. Created. Another separate, group and separate silo, they, can be. Impactful. Hmm, we see them in the highest performers, so high and elite performers. But. We only see them show up about nine percent of the time it, takes. Significant. Significant. Investment, and buy investment, I mean money, training. Resources. Staff, because. If you do that you have to take people, out of their original work environment, build, things build, things that are meaningful and real and impactful, and then find ways to embed them back in their normal environment and, still. Continue, on all of the work that they're doing. Yeah. So that's how we scale, and our next finding, what do we have Nicole, cloud. Continues. To be a differentiator, for elite performers, and it drives high performance, I mean. Obviously you've. Got to be in the cloud in order to get this stuff done so in, theory obviously. But I'm a researcher, so I need to test everything, and. It's. Also interesting, because I. Trust. No one and when people say they're in the cloud I went, out and a whole bunch of people said they were using the cloud but it didn't make a difference. Some. People said I know shock shock face this is my surprise face some. Executives, said that they were making investments, in the cloud that didn't make a difference and some, developers, said that they went to the cloud but their jobs didn't change I'm like. But. How did you go to the cloud and nothing changed, so, what does it mean to be in the, cloud I love. That question because. As. A researcher. I have to have a way to measure things to. See what the difference is and what we see is that ends up being the key difference. Is how. You actually, do, cloud, and then, we see a huge, difference. So. How. You actually, do, cloud is what matters, so there, are five key. Practices. To, doing cloud correctly. And I didn't make this up this actually comes from NIST National, Institute of Standards and Technology. So. These are the five one, is on-demand self-service, right, I can get it anywhere and all, the time all the time and self provision, so if it ends up hiding behind like, a ServiceNow ticket, for someone to approve. You. Get no points doesn't. Count number. Two broad. Network, access not only can I launch things anytime I want or need to I can do it basically from anywhere if I need to I can pull, up my phone and launch a service yeah exactly, three, is resource pooling right. So we get to share the resources we know that. My, stuff might run on some of the same hardware as your stuff but it's isolated. We're pulling those resources, together yep exactly. My favorite rapid elasticity that's right what means we go bigger and bigger and bigger and bigger and bigger and bigger bursting. Like magic up yes and down, all right we should get smaller too.
Which. Is important because we have a measured service and we, care about how much money we're spending on the cloud and we want to be able to look after that and spend the appropriate, amount you only pay for what you use okay. So, here's. What we found of. The, respondents, that said they were in the cloud only 29%. Of, them were, doing all five. And. Elite. Performers, were twenty four times more likely to have met all five characteristics. Of cloud computing so. That ends up being the key difference it's. Kind of like DevOps right you can ask people if they're in their cloud and you're, gonna get, five. Different answers for what it means to be doing cloud computing, right it's. Sort, of like that gym membership that you pay for but you got to actually go to the gym if. You want to reap the benefits of cloud computing. You. Need to, actually have. Processes. And technologies, architected. To let you leverage, them the right way. And. You. Know there are clear wins that you get from cloud, not only the fast auto-scaling, you get that cost visibility, that we talked about but, another great thing that you get is security, yeah, have someone, that is not within your organization, that is looking, after securing, that in first now, this, does not absolve you from any security, sins you still have to care about security, within, your own applications, and infrastructure, but, depending, on the services, that you're using that security. May be built right into the service yep absolutely and of course with Google cloud we offer a number, of different cloud-based services. That, you can take full advantage, of you, can do, much, less with your infrastructure, give, more of that to us let, us worry about securing. It, and. You've. Probably heard, but we recently G, ade cloud, run so, if you've been container, izing, your workloads, and want to get to a place where you don't have to think about the infrastructure, at all with, cloud run you can actually give us a container and run it in a fully service. Operational. Environment, that means we're handling all of the underlying infrastructure, the security, of that underlying infrastructure, and so forth you need to build that container hand it to us we'll scale it up we'll scale it down we'll, manage all of that for you and you only pay for what you use and you pay for we like that managed service piece absolutely.
And We. See customers all the time including, this particular, research report, the. Broad Institute looked. Into, how was their cloud spending, spent, happening, over time and the, more cloud, resources they, took advantage of the, more they were able to cut their costs significantly and the thing I love about this example is that they were able to adopt an iterative approach right, so they kind of did it a step at a time and a piece at a time was, something else I love about the cloud so. Another thing that we found is that. Productivity. And. Increasing. Productivity can, actually, drive improvement, and work/life balance, which. I really, love because some, of this is about performance developing, delivering software but some of it really, is individual, right like some of it is about making our own lives better and I, want to be more productive but I don't just want to burn myself into the ground I want. To make my own life better and by improving. Our productivity it, can, reduce burnout, and organizations. Can make smart, improvements. To. Help with that so there's a lot more detail in the report, but. Some of this comes down - easy-to-use. Tools to make CI, and, CD butter and effortless, right so one, of those is you. You know is your tool useful does it actually help me do my job which sounds. Obvious I don't know about you but like I've had tools that. Don't. Help me very much I just fight with that tool all the time so having good tools, good, developer, tools good IT ops tools are, super. Important, another. One is that it's just easy to use right. That. One is super important another. One, that I love is how, well it's integrated. Into the other things that I use a. Nerd. A nerd note about integrated, yeah this actually throws all the way back to my PhD, in my postdoc, we find that like how, well the tools integrated to the rest is super, predictive of performance, for technology. Professionals, yeah and speaking of integrated, I know like I like to stay in the same set of tools I don't like to move from, one IDE, over, to a web browser over, to this over to that just to get everything, that I need done and so we're making a lot of strides within, Google cloud itself to, help you with that as well, so as an example cloud. Build is integrated, right into github, so, that without actually, doing anything you, can change the code in your repository push. That change to github cloud, run can arts sorry cloud build can pick that up create, a new container for you and orchestrate, all of that you don't have to change the environment that you're in it's fully integrated into your workflow nice, love it. Yeah, and there is a right this is our sixth finding, here from the report there's a right way to handle change, approval, process, right, and.
It Leads to improvements, to both speed and stability and I know, you. Know change. Approval process, it's, all about making sure that we're safe, so, I have a change control board that looks after the changes that are coming in make, sure that they get approved, I know, because. I can see the whole landscape I know what's going to happen and so I can. Be the judge of whether or not this change is safe does anyone here have a change approval board where they work. Change. Approval boards, are very well intentioned. The. Challenge, is that they so often insert. Long. Delays, and what it is that we do change approval boards and time. Delay, right it's a huge thing and the. Challenges, that we also know that delay. We. Don't know we had heard, stories that delay. Can. Be problematic, right we know that speed and stability go, together so, this year we explicitly, tested. It and we actually find, that. Heavy. Weight process. And heavy, weight change approvals, actually. Lead, to instability. And. Having. A clear, change approval process, drives. Better. Performance. Better. Speed instability and reduces. Burnout, so. We do have some, great tips, and tricks in the report as well, as guidance about how change approval boards can, help. Streamline, their, change approval process and adopt, a more strategic approach. Yeah. I think that's really important also that it, helps to reduce burnout, so, I, think that when you have a clear, understanding of. How does this change that, I've just made how, is it going to get all the way through to production do I have, transparency. Into what that looks like with, a change control board often I'd make my change to the code and I I, don't, know it goes off to the change control goes into the magical black box of doom, right. We don't know what happens, right but. By adopting risk, based approvals, it. Can make things a lot better it's. Great okay, but. Like hmm. I I, keep like popping, on my research hat so what, what can you all do like how how. Can we apply this, and. What it is that we do I, have. An idea yeah. We. Can go over to cloud.google.com, slash. DevOps and I'm just gonna do that definitely, I'm yeah, questions, oh this. This. Section. Of the cloud.google.com. Site. It's at slash devops if i didn't mention that has, a lot of great resources for you the, first thing that's listed on the page right now is a video and.
I Don't know about you but if you wanna see Nicole on. A trap here they put me on a trapeze yeah and, in a hot-air balloon you can go check that out all about the that's. Not the useful, part skip. That yeah. So, then you can come on here and download the state of DevOps report yeah the 2015, report absolutely. The thing that I'm really excited about that I want to share today though is this DevOps, quick check so, you can, go back into your organization. And answer. Five, quick. Questions to. Sort of assess, where. Does your organization sit. So, we're gonna do that right now I have to crowdsource the answers so I'm. Gonna call on a couple, of people if you, get called on please realize, that whatever, answer, you give this entire audience, is gonna come back with you to your office and. Audit, it for truthfulness so. Sir, in the white shirt right here for. The primary, application, that you work with what is the lead time from the time you commit until, that time that. Changes, in production. Less. Than one hour I have all right less than one hour, let's. See that all right our second, question the deployment, frequency, and I'm gonna croute I'm gonna I'm gonna do the right thing and crowdsource, this from a bunch of different people so, the narrator completely, not the right thing a false, answer. No. All right normally you would do this assessment per, team because, teams, operate, at different Cadence's, right, right. So, deployment, frequency, you're. Sitting right next to our first answer would you mind telling us how, frequently. For your primary application. Is code moving, into the production environment. I'm. Said less than one day. Multiple. Per day okay great. So on demand we'll, call that multiple, per day okay. So. Now we, have your application you've. Realized. Through. Your monitoring, through, your users that, there's an outage how. Long does it take to fully recover, from that outage, or from that incident and, your choices are more than six months one to six months one. Week to a month one day to one week less than one day or less than one hour. Just someone shout out an answer for me. Less. Than an hour, alright, I think. We might find some elite performers, here Nicole what do you think now alright. The, fourth of five questions your change, fail percentage. So, when, you push code into the production environment how. Often, does that require manual, intervention, which, may or may not be preceded, by an expletive yes. Nicole. Pick someone to give us an answer here. Sir. What's. That. Zero. Zero, to fifteen percent all right the, final question that we ask is what is your industry, now, this, allows us to make comparisons right. And this also shows. Us all of the industries that you've looked at through part of your yeah right yeah as you mentioned more than we're on the previous slide, because. We know retail is the one that stands out most I'm just gonna select let's do it all right, so. Now with, these five questions, answered, you can view your results and, what, you can see of course everyone. Gave the, best answers, possible here. Just about which is great but, you can see your, overall, ranking, across, all of the research and then so, the blue bar is where you are the, black line is where the, rest of the research respondents. Are and we go so typical, industry performance yep typical, industry performance use the word typical, because that's like median. We don't use averages, sorry stats hat is on yep, and then we can click over to your industry, to the retail, side and I, just want everyone to watch this is the magical, part of the demo or you keep, your eyes on the black bars and I'm gonna click your industry, oh we. Got better did everybody see that I'll just do it again in case you missed it watch watch the black bars and they, copy each alright so, you, can compare, where, do you sit not only across the, entirety. Of the, reality but also within yours, so like how good, are you is in terms of a performer, we're, about to roll out a classification, so you can see if you're a low/medium hire an elite performer but, then the big question remains, how. Do I get better what if I come back is a lower performer, or a medium performer. How. Do I get or even if I'm a higher and, performer, how do I get better yeah, and so you can see exactly where you fit here and then there's a couple, of things that I also want to share with you that answer exactly, that question now do you, get better on this, page we also have a collection, of software, delivery capabilities. Or DevOps, capabilities. If you will and these, capabilities, you'll, see are falling, across four, different areas. So, we start off with some of the technical, capabilities. Things, like version, control do, you put all of your code and configuration.
Into, Version control if you, don't please. See me afterwards I want to have a discussion with you and I want to teach you some get but. There were other other technical, capabilities. Here as well right or even see I like, CI is one of my favorites I will mention really quickly these are all capabilities. That we have found that are statistically, significantly. Predictive, of improving, your ability to develop and deliver code with speed and stability right. Will you click on CI for me well, I click on CI a I certainly, will so so. Often when I meet with customers or. Enterprises. Or clients, or like, bump into someone and they'll say oh well I'm doing CI and then. They'll find out that CI is something, that like in some of our door assessment, work CI. Ends up being a constraint, for them and they're super surprised, and then, we'll roll through this and like. If you scroll down a little bit you'll say like in order. To be doing CI that, is, significantly. Predictive, of performance, you. Need to be doing, like. Four or five things, so. Some. Of these are. Really. Helpful to help outline, right so come and pick common, pitfalls are you, actually doing, these things in bold so. These. May be super, helpful and kind of interesting, in terms of doing, a quick check because CI is one of these like. Magical, industry buzzwords, that so many organizations have decided are a. Personal. Or like an organizational. Initiative and then. They will redefine, it to. Only cover, like two of these things or three of these things but, not all of these things. So. Right, here on this page you. Can see all. These technical, capabilities. And as you saw we have an article that helps, you take, this technical, capability, back into your organization, we, saw, there how, do you fight objections, to CI what do you need to measure but, it's not just about technical, capabilities, there are also process, capabilities. You see here right. Here the streamlining, of change, approval, again. That's research, that comes directly out of the report other, things like working in small batches and team experimentation. And, even. Cultural things absolutely, there are gradual, things here like. Like. Westrom, like the westrom organizational.
Culture Right if you've not read Ron West rums research. It's cool you don't have to you can come to this page and we'll tell you all about that research and how to implement it within your own organs, to. See ecology transformation how it helps and you should totally read Ron West rooms actually as well, it's so really good there's also measurement, here as well yes forget, measurement metrics and we have a couple of different capabilities. There. As well so from, this page you can get not, only a great assessment, of sort of where to use it as an organization, or where does your team sit but, you also get insight, into all of the capabilities, here, and before. We go back to our slide I will say a couple more things are coming so the 2019, report just came out we don't have like I were, working on building out the the deep dives into the lake code maintainability, was in the 2019, report it's not here yet stay tuned, another. Thing another, thing you can do from here is you can share the results, I find. It fascinating to, go into an organization and, ask a team, who. The, entire, team is working, on the same application, I ask them to take this quick check and then, I compare, the results, across, the individuals, of the team I have, a theory, but, Nicole won't let me say. That it's finalized, yet because, I haven't tested it well enough but, I think the higher up the organization, you go, I think, the rosier, the picture gets oh no that's, totally true always it executives. Tend to overestimate their technology, capabilities. By 2x yes yes, that's, real, which is why when the, door assessment, is, live so so our assessment, only, collects data from dev test ops QA InfoSec and we don't collect data from executives, because they. All. Right Sarah their estimates, are Rosie, yes, so we definitely want you to check out cloud.google.com. Slash. Devops you can find this information here all of these capabilities but, still how do we put it to work so I will, mention we, just, showed, you how. To assess how well you develop and deliver how. To improve, it because you've got all these capabilities but. Then the next key question, is where, do I start, there's this big giant list of stuff that is predictive. But, it's a giant list. So. Oftentimes, people will come and they'll say I know. Maturity. Model who. Here. Loves or who's executive slash manager, slash boss loves maturity models, everyone. Right because. It gives you this nice rosy. Guided, path to, Nirvana, and it's always five levels, why because, they're made up and everyone has five fingers right. It's fine the. Challenge, though is that, they. Don't they're, not great, because all teams, are different right, like I can't rule out one maturity, model that will be the same for cloud native versus. Mainframes. It's, gonna be different so, what we need to do is identify your constraint. Take. A look at those capabilities, and think, what's, my biggest headache what's, holding me back the most that's. The rough version you can also take very. Formal. Versions of it like value, stream mapping that will be a process based constraint. Approach or the. Door assessment, takes a capabilities-based, approach but, you can sit down with your teams and just say of these capabilities, what's, our biggest problem, what slows us down the most what do we swear out the most, start. There because. That will allow you to accelerate, your transformation. You're removing the roadblocks so. Focus. There. Dedicate. Resources which, is either money or time or focus, or attention, once. That is no longer your constraint. Identify. Your next constraint. Sometimes. That makes executives. Or bosses little bit nervous because you haven't given them a three to five year roadmap, but. What you have done is built. Them a customized. Roadmap, with, a year-long. Six months two-year long time, frame and, you're. Then training everyone, you're building, institutional. Muscle on how. To build, your own roadmap, everyone. Is now the experts, in how, to build a roadmap how to identify constraints. How to think strategically and. You'll. Get better. Exponentially. Faster and it, works for anyone whether they're at, the very top end of the spectrum and they're already elite performers, or.
They're. Early on and they're, just getting started, right and think back to those high performers, from last year that maybe are medium performers, now, potentially. The reason that they're now medium performers, is because they stopped iterating, they stopped that practice, of continuous. Improvement so. It is truly a continuous. Improvement and continuous learning journey, that, we're on yeah and we have more details in the 2019, report about how to do this including tips about how to attack, this from multiple, areas in the organization, there, are some things that you can do at the management level and things you can do at the team level even. More optimal, performance yep. So, please do go download the report cloud google.com, slash DevOps give, that report to everyone, in your organization it, is full, of great information for. Any level, within your organization, CEOs have read this yeah sis admins that we keep in the dark and in the data center have read the oh no no no, no, I used to work in that dusty corner I know it's not very fun down there is it yes with. That we would like to thank you and remind you to please provide, us some feedback on the talk and the things that you're learning today throughout the application, thank. You very much thanks so much for joining us. You.