Empower Your Business with Lead Nurturing and Management in IT Architecture Documentation

How to create outlook signature

[Music] hi everybody as Mary said I am a consultant so I work for consultancy which means that I probably get to see more organizations and their approaches to developing software than people who do not work at a consultancy and I also talk to people a lot my coworkers who also again see a different set of organizations so sometimes it feels like being at a conference every week and that's great for discovering similar patterns but also of course similar challenges and one of those challenges that lots of larger organizations have had over the past few years is that of how to coordinate autonomous teams so why do we want autonomous teams it's been this great or it's been this buzzword over the past few years and it's of course not that new of a trend it was already hidden in the second page of the agile manifesto the principles page and there's one principle that says the best architectures requirements and designs emerge from self-organizing teams so in a way I think this term autonomous teams is probably kind of a Renaissance of that idea of self-organizing teams and why autonomy what are we hoping to get out of that a few reasons one of them is motivation so what autonomy is a powerful factor in intrinsic motivation and also helps with creativity and people taking initiative and then another reason is that we want some flexibility so different problems needs different approaches to solving them sometimes so we want to have a certain amount of variety where it makes sense and then it's really helpful if the teams who are close to the problem are empowered to decide for themselves when they want to deviate from maybe some central norms and they have to be they have to feel safe to deviate from that so you don't want needless variation but you need flexibility sometimes to solve things in different ways and finally and this is I think the original reason why we want autonomous teams is scaling in the sense of organizational scaling so the development and the design process they just don't work anymore as the the group who is doing it is getting larger and larger if you have everybody in the same group so if you want to build more you need more people and then you need to coordinate the work between those people so you almost have to split them up into smaller groups because otherwise it doesn't work anymore and autonomy here is also means decentralization in a way because the larger you get the harder it is for a central group to make all the decisions because they just won't have enough information anymore to make good decisions so autonomy is a solution to scaling organization scaling it's not just the magic solution though because with that necessity come new challenges and those challenges come when we combine that with architectures so some of those challenges when you have lots of autonomous teams in your organization and it comes to architecture is that you risk duplicate work or even worse inconsistent duplicate work so because maybe teams didn't know they were both working on the same thing there's risk of that needless variation that I just mentioned you have the extra effort of coordinating all the work and the dependencies between the teams and also you often have challenges with staff mobility so if you want to move people around between teams and you have that unavoidable level of variation across the organization then it becomes a bit harder so those are just a few of the examples of challenges when you have autonomous teams and you want to have a good architecture in the sense of an architecture that supports your organization's goals and that makes you ultimately move faster so when you're in charge of leading an organization like that with all those autonomous teams and you want good architecture what do you do kind of say oh we need some governance for this and might be a bit of a cheesy approach but I did actually look up the word govern to govern in the merriam-webster dictionary and I was very glad because I found out that there are actually two different categories of what that word could mean so one of those categories is meanings like manipulate rule control or restrain and another category of definitions of the word are to guide have decisive influence serve as deciding principle or exert a determining or guiding influence so you kind of have this dichotomy of ruling and guiding and I think this also reflects how different organizations are interpreting what architecture governance means for them so to kind of you kind of have to find your flavor of what you need in your organization and what probably most of you want is kind of a mix of like it being guided autonomy so you don't want to be too prescriptive but at the same time you want to avoid chaos so you need some of those rules sprinkled on top of that so how do you do that I don't have like a definitive comprehensive answer for you but I have a few building blocks that I have in mind when I think about this and I want to walk you through them and also give a few examples of things that I've seen in organizations so I want to start with a reminder of something that might seem obvious but I already just mentioned that ultimately we don't want architecture for architecture sake or to kind of satisfy our feeling of everything being our need for everything to seem neat but ultimately architecture needs to serve the organization's goals or the business's goals and the the developers on the autonomous teams they are shaping the respective pieces of the architecture and they need to have a similar understanding of how their piece can ultimately contribute to the organization's goals and this might seem obvious but I'm still seeing that a lot that developers are not really asking for that context that they don't really seem that interested in it and on the other side the business side or the product side they make assumptions that developers don't really need to know that much about this and I also find that in myself I'm often so into the tech that I kind of forget to ask extra questions and when I make myself ask them I almost always find out new things that I didn't know before and I regret that I didn't ask those questions earlier so you need to make sure that both of those sides kind of push and pull context about the organization's goals so if you are somebody with relevant context think about how to push it to the team's a very simple way at least if you're poor located is to visualize things in the team space like this example here on the slide where there were the dis business units three main goals in order of priority that we constantly had up in the team space and could refer to you could also work on a team or business units mission statement maybe for the next quarter that also really helps show the other teams what you're working on and also check with stakeholders that this is actually what's needed right now and if you do need more context on the business or the organization which I would argue as developers and especially as lead developers we should all have then try to consciously pull it more and remind yourself of that like the minimum should be that we all understand either how the business we work for makes money or the product that we work for and where the cost for that comes from or if you work for a nonprofit organization of course what is the mission and how does this product contribute to that because otherwise you might be ending up with something like this so you might be solving imaginary issues it's kind of thinking a bit about that when you're talking about performance before because that's also something of course that goes into that like that we all think about what we're fixing those things for so the techies need to understand the organization's contexts and goals to be able to move in the same direction together but on the other hand also the business or the product side needs to have a certain level of understanding of the architecture and what we need there to help us define and prioritize our architectural requirements on that context so what do I need by mean by architectural requirements it's basically what is also referred to as ill-at-ease or some people called non-functional requirements or cross functional requirements when I did a dry run of this talked to some of my colleagues before one of them was really adamant that I should not use the term architectural requirements because they are also product requirements and I would agree but I think in this context it's worth pointing out that they are also slightly different so the trouble with with them is that there are a lot of them so this is just like a brief excerpt of that here you can actually see a lot more so you need to find out which ones are actually relevant for you and that might also be different on different levels in the organization so some of them might be very different for each of the autonomous teams and they actually decide what's important for them and some of them might be relevant on the overall organization level or business unit level and teams need to coordinate to reach them together and also how to measure them together and some of them might be highly relevant somewhat relevant or not relevant at all and this is also whether it's that duality comes in again of the ruling versus guiding because for some of them you might want a lot more ruling like it's something like compliance for example and for some of them maybe you just want a guidance based approach or you just tell a team okay this is totally your domain you figure out what performance requirements you have for your particular product and then you need to prioritize those as well and that's where it becomes tricky if there's so many of them right and one thing you can do for this some of you might know this exercise the trade-off sliders it's often used to help all the stakeholders prioritize the competitive priorities of things like budget scope time and quality and you can also do that with your architectural requirements for example if you're building a web application that needs some offline capability where you need to download some some data onto the user's machine or device then that might that offline availability might be kind of in tension with things like data privacy or portability so you can use the slider to have a discussion around that that those things you cannot have them all 100% at the same time and then prioritize them so and then one way to achieve alignment that I and also my and my colleagues have seen in larger organizations over the past years is to create a set of architecture principles for the engineering organization and that's what I want to spend a larger part of the talk on now so what is an architecture principle I like this definition by Owen Woods that says it's a declarative statement made with the intention of guiding architectural design decisions in order to achieve one or more qualities of a system so this is a long sentence so I try to break it down into some boxes so it's a declarative statement and the statement is its purpose is to help us guide our decisions so if the principal doesn't help us guide our decision-making then it's not very useful for us and ultimately we want those decisions to lead to system qualities that we want and those are basically those architectural requirements that I just talked about so it's always helpful to be very aware of what those requirements are for us and I want to give an example and don't get too hung up on the content of the example so we want to stay on the meta level of how it's actually structured so this is an example from out in the wild on the internet the title is smarts in the notes not the network and then there's a bit of a description that says we want systems to be as decoupled and cohesive as possible and not centrally choreographed so this is definitely a declarative statement check but as a developer just seeing this I think would not really help me make a decision yet because it's not a lot of information I would still ask myself what does that actually mean for my daily work so that's why there's usually bit more than just the title and the description so this is the full content of that particular principle so you have that declarative statement at the top then you have something titled rationale so why do we want this so in this part you can put a lot of those qualities of the system that you hope to improve with this so it's almost like an iPod hypothesis you think that by having this principle you will improve these qualities of a system in this case it was things like we want to evolve over time and we want to change quickly and we think that will help us with that and then with implications and examples you can make you can talk more about how this can guide your decisions or the actual practices on the ground and this structure I don't know if some of you recognize it it comes from an architectural principle definition from TOGAF know if does anybody know what TOGAF is I have very few hands so it's like an open framework for enterprise architects to do enterprise architecture and this is one of the pieces in there that one of them that I actually find quite you also one of the few ones that I know personally after so this example was from John Lewis a software engineering principles John Lewis is a department store chain in the UK I was not involved in creating these at all I've never worked with them they're just out there on the Internet and lots of organizations are actually putting their own principles out there and you can get inspired by them and that's basically what they did they took that TOGAF structure they came up with 18 principles and yeah are using those in their daily work so that's one way to do it another structure I've seen is what I call them I had the three column approach so this is this is a graphic from Sam Newman's book building micro services and what you do here is you have those three columns and on the left you list your strategic goals so basically those organizational goals I talked about before and then in the middle you have your architecture principles and on the right you have design and delivery practices so basically bringing those principles to life having kind of practical guidance for performing your tasks and so it's kind of rationale and implications put a different way so your rationale is always that you want to achieve these goals and the implications are what you actually do on the ground an example might be that your strategic goals are you want to enter a new market and you want to focus more on customer centricity and you think that maybe being more data-driven will help you with that and then you come up with a snappy title for being more data-driven like fact over opinions and then there might be some practices that you think will help you with that like hypothesis hypothesis driven development and data democratization so these are just a few architecture principles that you can find out there on the internet or that I've seen at organizations so some of these might seem very obvious to you right so you might ask yourself and those just like good practices that I find on the internet anyway like yeah we want our application to be maintainable we want to scale horizontally we want it to be production ready so it's kind of like da you know that's already all out there and all those books and blogposts so why should you bother with your own principles I think one advantage of it is that it helps give you focus I actually think 18 principles is maybe even a bit too much because for me it should not be an attempt at having a comprehensive live list of principles because that comprehensive list is out there on the internet but you should think about what do you want to focus on and where do you want to get better in this moment I actually think success criteria for a principle is that you feel like you can throw it away because it's become business as usual and then maybe you can replace that one with another one so maybe have five or so and really focus on them and see if you can get rid of some of them another reason why you might want your own is to achieve consensus so if you want those principles to be guiding for teams to still take their own decisions then they have to have a certain level of abstraction they can't be too specific so there's lots of room for interpretation so the principles or the journey to finding them can be a great tool to have those conversations with each other and figure out what do you actually mean by them so you have people with different experience levels different previous things that they have done in the organization so you can get them all together and talk about this and actually also learn from each other and figure out what it means for your particular context and finally it might be helpful for change management so to kind of get an official blessing for some things for example depending on your organizational culture you might want to make it explicit that failing fast is now really something that is wanted so if previously there's been a culture where people have kind of been punished for failing and now you actually want this whole experimentation and failing fast then you might actually need this official thing where management says yes this is what we want and this is we understand that this is what it will look like and you kind of get this safer environment for people to point to when they then actually try that and say you know this is what we achieved with this or this is what we need to change about it so then if you decided ok let's try that let's put together our own five principles how do you actually find them and I've seen three approaches to this overtime the first one was kind of top-down so there was this architecture board of people who got together I don't know every two months or so and they came up with these principles and the great thing about it was that there was there were very experienced people in the room who had first-hand knowledge of the business context the bad thing about it was that they weren't directly involved in the development in the teams and they also did not document rationale for why they chose these principles so that made very hard for people to accept them and not question them the second approach I've seen was kind of a step by step approach and in this organization it was really helpful that they had a mission at that time so they had a big reap lat farming project on the way and this mission kind of drove the creation of their principles so they had an architecture guild and got together and whenever they got they came across the next challenge in this reap lat forming they thought about you know should this be a principle or how could we make this more general guidance for the future so that was very helpful for them and then the third approach was also kind of up front but then not make it so make it very collaborative with everybody and also make sure that it's not set in stone but that you iterate on it and that it's not like a fixed thing and I want to give you one example now for a workshop that you can do to come up with an initial list of principles oh that was the laser pointer instead of the arrow so some of you might recognize this this is a this is a common format for retrospective where on the right you kind of say this is where we want to go this is our light tower and then the boats are kind of you and the team or the organization and then there's wind moving those boats forward as anchors holding them back and there's risks on the way that you don't know if they will manifest yet but you think that maybe there's something there to look out for so you start with the goals and requirements and those are basically those organizational goals and the architectural requirements so ideally those things are already well defined when you have this workshop but even if it's not defined you don't have to wait for it to be really efficient you couldn't also start with some generic ones like faster time-to-market it's something that I would say almost everybody wants right so you can start with something like that and also have this workshop with business and tech people together like it shouldn't just be the tech people and if you do this and it's not so well defined yet it will also put some pressure on on the business or product side to be more explicit about what those goals are so and then wind anchor and risks those are basically like a SWOT analysis right so they're your strengths weaknesses opportunities and threats it just doesn't sound so Consulting need when you when you put it with wind and anchors and cliffs right so you you kind of you know do your your common thing that you would do in a retrospective you brainstorm around these areas and then you cluster and maybe based on those clusters you can brainstorm some principles like what could be principles that help us get away from those anchors and what could be principles that can strengthen our wind and focus especially on the ones with cross team relevance because those are the ones that are important for you to coordinate those autonomous teams and then once you've brainstormed those principles you can split up into smaller groups and fill out each of those principles more so this is basically another version of the name statement rational implications that you can use for people to work together and fill this out more and then once you have those principles how do you make them more relevant to people again the implications what does this mean for my daily work for the practices I should use and one of the things you can do here to structure those principles is to create your own technology radar as you can see here in the background this is a Solano's version of it from last year who here knows about the technology radar formats this is actually so full disclosure this is a format that my that comes from my employer that works and we usually do it for for the whole market like what we see at our clients but if you do it for your own organization it's actually slightly different so it's basically a format to organize and document the concrete practices and technologies that you are using in the organization or that you want to use or not use in the organization and you have these four rings where you put this so these green dots represent your technologies or practices and the things that you put an adopt I think it's probably not readable in the back things that have proven to work within your enterprise and that are well supported so you put things in there that are have been in production for a while that you want everybody to use ideally if they have a similar problem to solve because you already have a lot of experience with it then in trial you put things that are promising and you're experimenting with them and one or more teams so maybe there's one team who already has it in production and they are playing around with and you're collecting experience and then assess are things that you're evaluating for potential experiments that you're researching but they're not in production anywhere yet and then holds as things not that you want everybody to drop immediately because the whole things are often the cash cows write the pieces of legacy that are actually still running your business but you want to highlight that they you want to deprecate them in the future so you don't want anybody to start new projects with it so having a session and putting this together and then constantly maintaining it is a great tool of learning from each other as well on the team and getting an overview of what other teams are actually trying out and note how you should you can also use this to determine where you need a lot of guidance and documentation so the further in the middle something is the more you should think about are you giving people enough tools and guidance to use this are you making this easy to do because that will really help you get people to adopt what you want them to adopt and another way to structure this if initially maybe you don't want to come up with a whole radar but you have at least an idea whenever somebody starts a new team or there's a new product line that you want to start what are sensible defaults I really like this term of like having a default so that implies that you can do from the default but it's like something sensible to start with and then a team can see what else they need so let's say you have looked at those building blocks as well and you have lots of energy and good intentions and came up with the radar and some principles but how do you not actually keep the momentum going and keep it alive how do you know if what you're doing is effective and how do you know that you're doing autonomy right that you're actually getting out of autonomy what you want so the state of DevOps report has some answers on that the state of deference reports I hope that some of you have heard about it and also the accompanying book accelerate both excellent excellent things to read it's basically research that over the multiple years that has found relationships between the things that we do in software delivery and how those things impact our performance our software delivery performance but also ultimately the organization's performance so they've actually found a predictive relationship between those two things and good news they found that autonomy is indeed one of the predictors that influences our performance in a positive way and they are saying in there that autonomy predicts trust and voice so Trust in the sense of people trusting their leaders and thinking a fair and honest and trustworthy and voice meaning that people actually speak up and are not afraid to say the wrong thing and being punished for it so if you think you're providing autonomy but you don't see these things you don't people trust the leaders and you don't see them actually voice dissenting opinions as well then maybe you're not doing autonomy the right way because as an example for the principles if you regularly go to the teams and ask them to justify how they're following the print following the principles and maybe even consciously or subconsciously punish them if they're not then they're not going to trust your voice their opinions they're just going to play hide and seek with you and Trust and voice also ing to the research positively influence a healthy and highly cooperative organizational culture which in turn again is good for your delivery performance so autonomy also has an indirect effect on that as well and another thing that they say that has a positive influence on this highly cooperative culture that you need is to have a climate for learning so helping each other do the right thing and making it easy to do the right thing and for me this is actually the key or the success factor in doing things like principals right it's kind of the attitude that you take towards these tools are you seeing these as tools for telling people what to do or are you seeing them as tools to foster a learning culture because you can use tools like principals and other things I mentioned as part of that learning culture so if you talk a lot to each other about these things then it helps you learn from each other and it can also help you reflect on the progress and ideally you have measures in place to measure some of these things and then help you evolve the principles if you're seeing like they're not actually getting you anywhere and you can also use them as tools to record the whys so why are you doing things the way that you're doing them because if you don't know why you're doing it then you might not then people might spend a lot of energy on questioning all the time why it is this way why is it that principal not that one I've seen it differently in my previous job so that's a lot of energy that you could be spending on other things if you had a good way to document how you how you did that workshop for example to get to the principal's what are the anchors you want to get rid of and so on so I heard pet qua say in an interview recently that these days everybody's doing architecture all the time especially in a fast-growing organization and I think to scale that you have to find tools to help people learn from each other as much and as fast as possible and yeah these are some tools that I found useful for this and I'm looking forward to the office hours to maybe hear some other things from what has worked for you thank you [Music]