<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by David Thomas on Medium]]></title>
        <description><![CDATA[Stories by David Thomas on Medium]]></description>
        <link>https://medium.com/@david.thomas1?source=rss-433c2ef9fdc5------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*i3tH9pwVttBLz_gv</url>
            <title>Stories by David Thomas on Medium</title>
            <link>https://medium.com/@david.thomas1?source=rss-433c2ef9fdc5------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 11:10:33 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@david.thomas1/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[A year of organisational design]]></title>
            <link>https://medium.com/@david.thomas1/a-year-of-organisational-design-db5eef9b8483?source=rss-433c2ef9fdc5------2</link>
            <guid isPermaLink="false">https://medium.com/p/db5eef9b8483</guid>
            <category><![CDATA[organisational-design]]></category>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[uk-government]]></category>
            <dc:creator><![CDATA[David Thomas]]></dc:creator>
            <pubDate>Tue, 16 Sep 2025 13:20:21 GMT</pubDate>
            <atom:updated>2025-09-19T11:32:23.347Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FsZcCLkGr61QSLIPHRPDiQ.png" /></figure><p>A year ago I started focussing on the organisational design aspect of my work. In my career as a delivery manager I’ve always tried to remove blockers for my team, but also remove the causes of those blockers. I’d always aim to leave the organisation in a better place than I found it. The past year has seen me spend less time looking at team level blockers and more at organisational level causes of those blockers.</p><p>As this was a learning journey for me I wanted to share my thoughts and what I found along the way. To this end I shared three posts outlining <a href="https://medium.com/@david.thomas1/organisational-design-part-1-4ab4f055eb3b">my view of organisational design</a>. I also started <a href="https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-1-2f4d2de8d7ab">blogging weekly</a> (ish) about my experiences of organisational design in practice.</p><p>Originally I thought I’d be writing about making a series of small changes, how I was rolling them out, and the difference it had made. This isn’t exactly what happened. Organisations that need large scale transformation struggle to quickly roll out a series of small incremental changes to improve the way they work.</p><p>In the end the year split into three chunks:</p><ol><li>A waterfall attempt at changing our governance processes</li><li>A more agile attempt at changing our governance processes</li><li>Setting up a delivery group</li></ol><h3>An overview of the year</h3><h4>Waterfall organisational design</h4><p>After volunteering to support our digital department’s transformation programme, I was assigned to the governance workstream. This made sense as I had a lot of experience with our governance processes and had worked to improve them previously.</p><p>I threw myself into the work, organising workshops and designing what an improved set of governance and assurance processes would look like. My focus was a high level vision for governance and assurance processes, as well as a set of low hanging fruit that we could focus on first. I was immediately hit with a large amount of nervousness. What if the changes make things worse? Where is the evidence backing any change? Have we got enough senior support for changing things?</p><p>This nervousness meant the series of incremental changes developed into a larger proposal to be signed off by senior leadership. This proposal wasn’t huge, but still ended up in rounds of discussion and consultations. Eventually after 5 months the proposal was overtaken by my next piece of work and was shelved without ever being put in front of senior leadership.</p><h4>Agile organisational design</h4><p>In February I started a new piece of work as part of the workstream responsible for rolling out delivery groups, the new delivery model we were moving to. This piece of work was to develop a governance model for delivery groups. So far so familiar.</p><p>I was given a lot more scope to shape this piece of work, and set up a multidisciplinary team to iteratively and transparently develop a new governance model. As delivery groups hadn’t been rolled out we couldn’t make changes iteratively, but at least we could design the model iteratively.</p><p>As a team we developed a draft set of assumptions and principles, and ran a large scale consultation session to get feedback. Based on this we developed a draft governance model and ran a similar consultation on that. We spent the next few months making changes to the model every fortnight and running another consultation session.</p><p>It was intense managing this work alongside our day jobs, but as a team we were really proud of the <a href="https://defra.github.io/delivery-group-governance/">final (draft) model</a> that we produced. I say draft because 3 months later this still hasn’t been in front of the final board it needs to go to for sign off. Getting something in front of a board is a challenge, let alone getting it signed off by them.</p><h4>Setting up a delivery group</h4><p>Towards the end of my time developing the delivery group governance model, I was appointed as a lead for one of the new delivery groups. This meant I got to design my own portion of the organisation.</p><p>This coincided with a big slowdown in posting updates about my progress. This wasn’t because I was suddenly terribly busy, doing my day job and developing a governance model was harder, but because there seemed to be fewer big things to post about. When I had control, organisational design stopped being a series of big hurdles to jump over, and more a series of small conversations to have. Gentle nudges to encourage people to feel empowered, or conversations with our heads of professions to make sure we got the right people in key roles. Having said that, while it has been less post-worthy, it’s been hugely rewarding.</p><h3>What I’ve learnt</h3><h4>Agile transformation has to be agile</h4><p>This isn’t really something I’ve learnt. In fact it sits front and centre in <a href="https://medium.com/@david.thomas1/organisational-design-part-1-4ab4f055eb3b">my description of agile organisational design</a>. But this last year has really hammered it home. Designing an agile operating model upfront and then planning how to transition to that model misses the point of what agile is. There isn’t a correct agile operating model. For an operating model to be agile it has to be able to adapt.</p><p>A waterfall approach to transformation is likely to fail for the same reason waterfall projects often do. If making a change is a big thing which requires careful planning and execution, changes need to be big to justify that work. Which means you need more careful planning. Which makes for bigger changes. Until ultimately the changes are too big to contemplate and are shelved.</p><h4>Transparency is hard but important</h4><p>Being really transparent about the changes you’re making to an organisation is important. While some people may be concerned about sharing ideas that aren’t fully formed, this gives people the chance to feedback early. If people aren’t involved in the changes being made then they will naturally fear the worst about them. This reduces morale far more than sharing early ideas.</p><p>Transparency helps bring people along with the changes that are being made. By being part of the wider team making a change, they are naturally more invested.</p><p>People are often resistant to having change imposed on them. Transparency brings different people from the organisation together. Colleagues who are resistant to change may be swayed by seeing that change benefits their colleagues, rather than just being imposed on them by management.</p><p>Despite these benefits, transparency is hard. You can’t just hope that everyone will know what you’re doing. And even if you tell them, people will be focussed on their day jobs so may only hear partial messages, or not understand the language you use. Having specialists in your team to communicate changes can help with this.</p><p>The most important part of transparency is that it means transformation work can begin to affect culture before it has been officially rolled out. Transforming culture is more important and trickier than transforming structures and processes. By talking openly about better ways of working the organisation can start improving immediately.</p><h4>Multi-disciplinary teams work well for transformation</h4><p>Understanding current processes and their limitations, understanding agile ways of working, understanding people’s needs and communicating effectively are all different skills. While everyone will have all these skills to a greater or lesser extent, creating a team made from specialists in each of those areas will make transformation quicker and more effective.</p><p>There is a lot of work understanding the different perspectives and complex processes that make up a large organisation. Being able to share this amongst a team is really valuable.</p><p>Finally, organisational design can be emotive. People are often resistant to and nervous of change. Having a team to support when those emotions are directed at you is invaluable.</p><h4>Senior leaders aren’t in control</h4><p>One of my more surprising realisations is that senior leaders aren’t in control of the organisations they lead. My mental model of a hierarchical organisation was people at the top, passing on instruction to the next layer down, which passes on to the next later down, until things happen somewhere near the bottom.</p><p>In reality senior leaders are too far removed from the work near the bottom to make coherent decisions about it. Any decisions they make will be interpreted through several layers before being implemented. This meant that I was often trying to persuade senior leaders to sign off on changes to processes that they didn’t know existed.</p><p>If people felt more empowered many changes could be made more quickly and effectively without senior leaders knowing or needing to worry. The only impact they would see would be a more effectively running organisation.</p><h4>Sign off is painful</h4><p>One of the impacts of the lack of control senior leaders have is that sign off can be painful. Senior leaders aren’t keen to sign off on something they don’t understand, and people are nervous about putting anything that is less than perfect in front of senior leaders. This means I’ve been working on organisational design for a year, a lot of which has required sign off from our senior board, and none of which has made it to that board yet (edit: this was signed off 3 days after publishing this).</p><p>But taking a transparent approach to organisational design work has meant many of the changes proposed have happened anyway. If it hadn’t been for that it would have felt like a long and fruitless year.</p><h4>Strategy isn’t more important than implementation</h4><p>Typically senior levels of an organisation will have a more strategic focus. A hierarchical model therefore makes strategic work seem more important than implementation work.</p><p>The closer I’ve got to more strategic work the more I’ve realised that while it is just as important as implementation work, it is in no way more important.</p><p><a href="https://www.flightlevels.io/what-is-flight-levels">Flight Levels</a> nicely separates work into implementation, coordination and strategy. These levels could easily be mapped to the different levels of a hierarchy. But at the same time Flight Levels explains that strategic and co-ordination work shouldn’t be seen as more important than implementation work, but all three are crucial and need to work together.</p><h4>Data isn’t the be all and end all</h4><p>Possibly the most controversial of all the views I’ll share regards data, and how it should be used to inform changes.</p><p>When it comes to organisational design bigger changes are less likely to need the backing of data than small ones. To explain this, consider improving the performance of a car.</p><p>If your car is on fire by the side of the road, you don’t need to gather data about its exact temperature or current top speed in order to make a decision. You need to buy a new car.</p><p>If you’re in a Formula 1 pit lane you’ll gather a huge amount of data, but the decisions that data will inform are relatively small. You might want to change the angle of a spoiler or the number of litres of fuel in your car’s tank at the start of a race.</p><p>The reason for this, is that even with big changes you do typically have data. Is my car on fire or not is a data point. Just not one that people usually feel the need to track. Similarly if your organisation hasn’t delivered anything for the past year that’s unlikely to be a good thing. And while we could gather data on exactly where in the process things are going wrong, if things have gone that wrong it’s likely to be a big change to fix.</p><p>Big changes are also harder to measure. We could have baselined the time to launch our services before and after launching a new governance model. But given the rate at which we launch services it would take a long time to see any difference. And if you measure something like that over a year, you’re as likely to be measuring the impact of a different external environment as anything you’ve done to change things.</p><p>So use data to make tweaks when your organisation is performing well, but trust your instinct and good practice when making the big changes early on.</p><h3>A few other things I’ve learnt</h3><p>As well as learning a lot about organisational design — I also learnt a couple of other things while posting regularly over the last year.</p><h4>Blogging is hard</h4><p>When I started writing I thought each post would take a few minutes to write at the end of the week. It was important for me to write in a short, simple and easy to read way. This meant that writing a post that would take three or four minutes to read, was taking an hour and a half to write. It didn’t help that writing was the last thing I did on a Friday, when my brain was fried and I really wanted to head downstairs to have dinner with the family.</p><p>I chose to write in a factual way, despite the fact that what I’m writing is a set of opinions. When I write I naturally qualify everything with “usually”, “normally” or “probably”. When writing these posts I chose to remove all of that. Not because I’m sure of everything I’m saying, but because keeping those words would make the posts longer and harder to read.</p><p>I also chose to stick to the subject matter. As a personal preference I like to read things that are direct and to the point, without pop culture references or anecdotes about my life. This has the benefit of making what I’m writing more accessible, but removes any sort of connection a reader might have to the author, which will be important for some people.</p><p>I also totally appreciate the irony of writing this while talking about myself, but yeah…..</p><h4>Drawing is fun</h4><p>When I started posting a year ago I considered myself totally non-artistic, and the idea of sharing anything I’d drawn was pretty terrifying. On the other hand I wanted to illustrate my posts, and drawing stick figures having conversations seemed like a good way of doing it (albeit one completely ripped off from <a href="https://xkcd.com/">XKCD</a>). To start with drawing the cartoons took as long as writing the posts, but with practice I could draw them in a couple of minutes.</p><p>Off the back of enjoying drawing those cartoons (and some lovely positive feedback) I started drawing daily at the start of this year. I kept this up for several months, and became really proud of what I could produce. I now regularly share things I’ve drawn with friends without any sense of embarrassment.</p><p>If you’d asked me a few years ago whether I’d ever be able to draw competently I’d have laughed at you, and if I’ve gotten nothing else out of this process I’ll take that.</p><h3>What next</h3><p>In the short term I’m going to focus on my role as a delivery group lead. That should definitely provide plenty of opportunity to scratch my organisational design itch.</p><p>I’m keen to get involved in wider organisational transformation, but I’m wary as taking on too much in the past has left me feeling burnt out. But from different funding approaches, to what delivery groups will look like next year, there are plenty of things that hold an interest for me.</p><p>In November I’m going to be speaking about the agile transformation approach we took to develop the governance model at <a href="https://bristol.ripple.events/">Ripples: Bristol</a>. It’s a conference I go to most years (under its previous name Agile in the City) and I’m a big fan, so I’m excited to be speaking there.</p><p>I’ll also keep writing, but only when I think I’ve got something interesting to say. I’ll give you a shout when that happens!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=db5eef9b8483" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Agile organisational design in practice — week 45]]></title>
            <link>https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-45-bb0e76468ff9?source=rss-433c2ef9fdc5------2</link>
            <guid isPermaLink="false">https://medium.com/p/bb0e76468ff9</guid>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[uk-government]]></category>
            <category><![CDATA[organisational-design]]></category>
            <dc:creator><![CDATA[David Thomas]]></dc:creator>
            <pubDate>Fri, 18 Jul 2025 14:20:13 GMT</pubDate>
            <atom:updated>2025-07-18T14:21:15.791Z</atom:updated>
            <content:encoded><![CDATA[<h3>Agile organisational design in practice — week 45</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Jj6qsRbjzXe7DNdbjbPe1g.png" /></figure><p><a href="https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-44-52af71c4748b">Last week</a> I looked at whether to try and structure teams in a consistent way across the delivery group. This week I’ve been looking at how work flows through the delivery group. I’ve been taking inspiration from <a href="https://www.flightlevels.io/what-is-flight-levels/">Flight Levels</a> to guide me on this, which has made me think about a gap (or next iteration) of the <a href="https://defra.github.io/delivery-group-governance/">governance model</a> I’ve been developing.</p><h4><strong>What do I mean by visualising work</strong></h4><p>I’m focusing on what happens to a piece of work before people start work on it. Within the delivery group we have multi-disciplinary teams, so once a team is working on something, visibility of the workflow isn’t needed at a delivery group level, as it is all contained within the service team. However at a delivery group (or organisation) level it is interesting what happens to a piece of work before it is worked on by a team.</p><p>In a very hierarchical organisation this might consist of a senior manager telling a middle manager some priorities. The middle manager then passes on a subset of these priorities to their junior managers, who then instruct the team. In a flatter organisation with a less centralised strategy this could simply be every team figuring out their own priorities.</p><p>In our case we have an annual budgeting process known as business planning. From this we get priorities for the year. These priorities turn into a set of projects which build services. Services generate their own work through bugs and feedback in lives services, as well insights gathered by our researchers.</p><figure><img alt="A flow chart showing how work flows through a delivery group prior to being started" src="https://cdn-images-1.medium.com/max/1024/0*SROG58X6tAZScuvp" /></figure><h4><strong>Why is this important</strong></h4><p>It is important to know the steps a piece of work goes through before it is started, so that people know how to influence this. If someone has an idea that could help an organisation, you want it to be clear how they should raise this to an appropriate place, so that a decision can be made as to whether this idea should be taken forwards.</p><p>If you’ve visualised something you can improve it. If it is taking a long time between someone coming up with an idea and someone starting work on it, then knowing what that process is, is the first step to improving it.</p><h4><strong>Flight levels and the governance model</strong></h4><p>Inspiration from this came from rewatching this overview of <a href="https://www.youtube.com/watch?v=6JtRPjTMEWQ&amp;amp;t=4s">Flight Levels</a>. At around 33 minutes in the video Klaus Leopold moves from talking about the different flight levels and the activities taken at each, to how work progresses across those flight levels.</p><p>While developing our governance model we focused on governance being about individuals, and the decisions they are able to make. This is important, but showing how governance works from the perspective of a piece of work is important too. This feels like a second iteration we could make to the governance model.</p><h4><strong>What’s next</strong></h4><p>Next week I’m going to take a closer look at one of the areas of work within my delivery group, and how that should be structured to improve delivery. I’ll let you know how I get on.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bb0e76468ff9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Agile organisational design in practice — week 44]]></title>
            <link>https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-44-52af71c4748b?source=rss-433c2ef9fdc5------2</link>
            <guid isPermaLink="false">https://medium.com/p/52af71c4748b</guid>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[uk-government]]></category>
            <category><![CDATA[organisational-design]]></category>
            <dc:creator><![CDATA[David Thomas]]></dc:creator>
            <pubDate>Fri, 11 Jul 2025 15:39:27 GMT</pubDate>
            <atom:updated>2025-07-11T15:39:27.594Z</atom:updated>
            <content:encoded><![CDATA[<h3>Agile organisational design in practice — week 44</h3><figure><img alt="A person talks to a hand holding a pen which represents the author of the comic. The person says “You know how all your stick figures look the same?”. The hand says “Yeah”. The person says “Is the strategic flexibility worth the local inefficiency?”" src="https://cdn-images-1.medium.com/max/1024/1*tcCterXeCSjO9ra6wWkaPQ.png" /></figure><p><a href="https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-43-223bbc2eb9d0">Last week</a> I worked on developing a set of strategies for the delivery group I’m now leading. This week I’ve been thinking about whether or not to get teams to work in a consistent way across the delivery group.</p><h4><strong>Context</strong></h4><p>There are two main programmes of work within the delivery group. One of them has had its service live for the past few years. The other has two services, one which went live last year and the other which will go live next month.</p><p>One of the programmes has a few internal staff members in product and delivery roles, but is otherwise outsourced to a supplier. The other is mainly internal staff members but with technical roles filled by a different supplier.</p><p>The first programme has a live service team to handle issues raised by users with a separate team focussed on delivering new functionality. The second has two teams aligned to the different services they provide. Live support is handled within the delivery teams.</p><p>There is no overlap between users of the two services and they’re aiming to do unrelated things.</p><p>In short, there are two programmes working on unrelated things and working in very different ways.</p><h4><strong>How could this be more consistent</strong></h4><p>The outcomes the programmes are trying to achieve can’t be brought together, but the way those programmes work could be. We could have internal teams working across both programmes, or a single supplier on both. Or we could have a single live service team across the programme, or make all delivery teams handle their own live service support.</p><h4><strong>The advantage of consistency</strong></h4><p>If these two programmes worked in a more consistent way this would allow more flexibility across the two programmes. Currently if there is a gap in an internal team, we can’t fill that with a supplier from the other programme, as that supplier is only contracted to work on that programme.</p><p>Similarly if priorities change and one programme becomes more important, it is hard to shift a team across because of the different team structures.</p><p>If we had a number of teams that were all internal or from a single supplier, then we could move people and teams based on the current highest priority and where gaps are.</p><h4><strong>Problems with enforcing consistency</strong></h4><p>Enforcing consistency means limiting people’s ability to do what is best in their area. It is the opposite of agility and empowerment.</p><p>A live service team may make sense for one service, as it has a higher volume of support issues, but not in another which has just launched and wants direct feedback while developing a service. A supplier may have particular knowledge of the technologies used in one service but not those in another.</p><p>Or it could simply be that people on one set of teams work better in a certain way, whereas people in a different team work better in another.</p><p>This means it is important to allow people to find their own way.</p><h4><strong>How to balance this</strong></h4><p>The way to balance the benefits of consistency with the importance of local flexibility is to provide a default way of working, but allow deviation from this where required. This is what I aimed to do with the sensible defaults in the governance model I developed.</p><p>The strategies I wrote last week followed a similar model. A good way of working, but with no push to implement them immediately if that would cause disruption. And no push to enforce them ever, if they weren’t right in a particular instance.</p><p>This means over time it is likely that the delivery group will move to a more consistent way of working, but I won’t be pushing for immediate change.</p><p>It’s also interesting to note that consistency and flexibility feel like opposites in this instance. If all teams work in a consistent way there isn’t room for teams to be flexible in how they work. But the first advantage of consistency is that there would be more flexibility across the delivery group.</p><p>This means that consistency makes for local inflexibility but strategic flexibility. The question is whether that strategic flexibility is worth the inefficiency caused by local inflexibility. The value of that trade off will determine how much you encourage consistency.</p><h4><strong>What’s next</strong></h4><p>Next week is looking quiet, as the summer begins. But I’m going to spend some time looking at how work begins in the delivery group. I’ll let you know how I get on.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=52af71c4748b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Agile organisational design in practice — week 43]]></title>
            <link>https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-43-223bbc2eb9d0?source=rss-433c2ef9fdc5------2</link>
            <guid isPermaLink="false">https://medium.com/p/223bbc2eb9d0</guid>
            <category><![CDATA[organisational-design]]></category>
            <category><![CDATA[uk-government]]></category>
            <category><![CDATA[agile]]></category>
            <dc:creator><![CDATA[David Thomas]]></dc:creator>
            <pubDate>Fri, 04 Jul 2025 15:54:45 GMT</pubDate>
            <atom:updated>2025-07-04T15:54:45.365Z</atom:updated>
            <content:encoded><![CDATA[<h3>Agile organisational design in practice — week 43</h3><figure><img alt="A person sleeps on a desk while another person presents in front of a complicated looking slide. The presenter says “As you can see on slide 94 our technology strategy leverages our vision…”" src="https://cdn-images-1.medium.com/max/1024/1*QwKZitRj23izt_8S5tp3ug.png" /></figure><p><a href="https://medium.com/@david.thomas1/transparent-transformation-a-retrospective-1d4f36a0fa5c">Last week</a> I wrote a retrospective on the transparent transformation approach I used for developing a new governance model for digital delivery. This week I’ve been focussing back on the delivery group I’m leading. In particular I’m developing a set of strategies for how the delivery group will work.</p><h4><strong>What is a strategy</strong></h4><p>A strategy is a long term plan or approach to tackling a problem. Much like governance there is a lot of misunderstanding about what a strategy is. It is common for a strategy to be presented as a slide deck that has been signed off by senior leaders. However this is not the strategy, just a presentation of it.</p><p>As long as you are working on something, there is a strategy for it. The strategy might not be coherent or documented, but it is embodied by the people leading the work. At worst if they appear to have no strategy at all, then the strategy is for everyone to work things out on their own, but there will usually be something more than that.</p><p>This means to start with a strategy should be documented, rather than created or signed off.</p><h4><strong>What does a strategy look like</strong></h4><p>For a strategy to be useful, someone should be able to read it and understand how they should change the way they work in order to follow it. This means a strategy should be short, and someone should be able to disagree with it.</p><p>It should be short so that it is possible to read and remember. If a strategy is ten pages of text then someone will struggle to remember it when doing their job. If a strategy relies on a series of complex diagrams, people will struggle to apply that in their day to day work.</p><p>A strategy consisting of a set of clear and simple statements about what a team is trying to achieve and how, will help people shape their work to fit that.</p><p>It is important that a strategy is something that someone could disagree with. If your strategy is too vague or exclusively positive then it won’t change anything. “Deliver great digital services” might sound nice, but it won’t change what anyone does.</p><p>“Deliver digital services focussing on quality even if it costs more” is something someone could disagree with. They could argue that we should focus on reducing cost at the expense of quality. This means that someone currently focussing on cost at the expense of quality, knows they should change how they work.</p><h4><strong>What have I done</strong></h4><p>I wrote down the approach I would like to take to running the delivery group in a number of different areas and shared these with service owners, product leads and delivery leads across the delivery group. I’ve broken the strategy down into 8 areas:</p><ul><li>Wellbeing</li><li>Delivery</li><li>Resourcing</li><li>Team structure</li><li>Live service support</li><li>Funding</li><li>Governance</li><li>Technology</li></ul><p>These aren’t going to be the right areas for everyone, but they’re the areas in which I felt I could document a clear approach, and in which it was important for my delivery group.</p><p>Each of these sections was around 150 words long, and consisted of a series of short paragraphs describing how we would aim to work in the delivery group. I also wanted to be clear that I didn’t want a sudden change as a result of publishing this strategy. Instead we’d prioritise stability and move in the direction of the strategy gradually.</p><h4><strong>What’s next</strong></h4><p>Because of my delivery background I’ve focussed on developing a strategy for how the delivery group will run. But this is only one type of strategy. The next step is to develop a strategy for what the delivery group is trying to achieve. To help with this I’m meeting the director in charge of the policy area I’m working with. I’ll keep you posted on how I get on!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=223bbc2eb9d0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Transparent transformation — a retrospective]]></title>
            <link>https://medium.com/@david.thomas1/transparent-transformation-a-retrospective-1d4f36a0fa5c?source=rss-433c2ef9fdc5------2</link>
            <guid isPermaLink="false">https://medium.com/p/1d4f36a0fa5c</guid>
            <category><![CDATA[organizational-change]]></category>
            <category><![CDATA[uk-government]]></category>
            <category><![CDATA[agile]]></category>
            <dc:creator><![CDATA[David Thomas]]></dc:creator>
            <pubDate>Fri, 27 Jun 2025 12:07:54 GMT</pubDate>
            <atom:updated>2025-06-27T12:07:54.555Z</atom:updated>
            <content:encoded><![CDATA[<h3>Transparent transformation — a retrospective</h3><figure><img alt="A person says “That went well didn’t it?”. Another person responds “Yeah but we’ll do better next time”." src="https://cdn-images-1.medium.com/max/1024/1*UjUxbGtiCCJHa0ua_EAcKA.png" /></figure><p>Four months ago I started designing a <a href="https://defra.github.io/delivery-group-governance/">new governance model for digital delivery</a> in my department. I wanted to do this quickly, transparently and to consult as widely as possible. Now the work is finishing, I wanted to share what worked well, and what didn’t.</p><h3>Background</h3><p>I work in digital delivery in a government department. We’ve moved from having a single large portfolio of projects and programmes to having eleven delivery groups. These delivery groups are aligned to business areas, and will have standing teams to deliver the services required in those areas.</p><p>To support this move from projects and programmes to delivery groups, I was asked to develop a new governance and assurance model.</p><h3>The team</h3><p>Using the skills and experience of other specialists in my department would make the model better, easier to understand and get wider support for it. So the first thing I did was set about building a team.</p><p>Our department piloted a single delivery group for two years prior to rolling out the approach more widely. Including one of the people who worked on that pilot meant we could learn from their experience. We couldn’t simply copy the model they’d developed, as it was done in their specific context, but building on what they’d done was invaluable.</p><p>It was really important that people understood the model we were developing. For this I got help from a couple of content designers within the department. Their day to day job involves writing for GOV.UK in a way that anyone can easily understand. They also make sure that content isn’t written just for the sake of having it, but that it serves the needs of a particular user group.</p><p>The content designers quickly pointed out that if they were to write content to serve the needs of a particular user group, they would need to understand the user groups and their needs. Our user researchers were the perfect people to support this and two joined the team.</p><p>While I have a lot of experience in the assurance processes around delivery, I don’t know so much about the running of live services. Bringing in a subject matter expert from our live services team plugged this gap.</p><p>Seven people felt like quite a big team for this piece of work. But with all of us working part time on it, there was a lot of work to go round.</p><h3>The approach</h3><p>We developed a series of iterations of the governance model, holding a consultation meeting every fortnight on each one. The aim was to always have the latest version of the governance model visible.</p><p>Before creating the first iteration we developed a set of assumptions and principles which would underpin the governance model. I knew a fair bit about how delivery groups were intended to work, but I wanted to share my assumptions before going too far down the wrong track. The principles I shared were largely based on <a href="https://agilemanifesto.org/">principles from the agile manifesto</a> and <a href="https://www.gov.uk/service-manual/agile-delivery/governance-principles-for-agile-service-delivery">UK government governance principles</a>, which I hoped would be uncontroversial.</p><p>We ran a consultation session to share these principles and assumptions. We advertised as widely as we could across the department. In the end about 70 people attended. People were pretty happy with the principles, and they corrected a couple of our assumptions.</p><p>Based on these principles and assumptions we developed the first iteration of the governance model. We ran another consultation session, iterated the model, and then held another consultation a fortnight later. After five iterations we decided the model was in a shape we were happy with. There would still be a need to iterate the model but the more intense work to develop it was complete.</p><p>The consultation meetings consisted of giving a quick overview of the work we were doing, before presenting the latest iteration of the model. This took about half an hour and was followed by half an hour of discussion on the model. We shared the model in advance of the meeting through the meeting chat, which gave people the chance to digest the model in advance. This also provided a forum for a more in depth discussion than would be possible in a consultation session.</p><p>We tried to reach as many people as possible through the consultation meetings. To do this we advertised through various internal channels, and welcomed anyone who might have an interest in the shape of the governance model, or who might need to use it. We also held targeted consultation meetings with specific interested groups. To keep people up to date we shared progress through weeknotes and presented at town halls and team meetings.</p><p>To complete the work we aimed to present it to the board which sits across delivery groups and hand it over to them. But more on that later.</p><h3>What went well</h3><h4>Teamwork</h4><p>Working as a multidisciplinary team made a huge difference to the quality of this work. Working with content designers did slow down the production of the model, but this was because they were ensuring what we developed was accurate and useful.</p><p>Having support from user researchers made a big difference to the quality of the consultation. It was a huge help to have someone tracking feedback while I was presenting. More importantly though the researchers were willing to engage with people offering more challenging feedback. Change can be stressful and so emotions can run high. As the person responsible for the content this didn’t encourage me to engage with everyone. Having someone more neutral, and someone with expertise in listening to people regardless of the situation meant that many more voices were heard.</p><p>Finally the involvement of other SMEs was vital. Whilst I have a lot of experience of governance and assurance it is largely limited to delivery. So the support of people who had already tried a similar governance model and those with experience of running live services was vital.</p><h4>Transparency</h4><p>Being transparent from the start worked really well. There were some drawbacks that I’ll cover later, but overall it’s the approach I’d always take in the future when developing any change like this.</p><p>The biggest positive is the momentum it gives the change. By talking openly about the work from the start, people become part of the change. A more traditional approach would be to develop a change, consult on it and then implement it. The approach we took meant that aspects of the change started being implemented from very early on in the change process as people got on board with what we were doing.</p><p>Being transparent brings together people who are positive about change with those who are more concerned. People are more easily brought round by their peers, rather than an external group who are imposing change on them.</p><p>Being involved in the development of the change means that people can shape it in a way that makes sense to them. Again this makes people more bought into it.</p><h4>Sharing the model early and providing a forum for discussion</h4><p>Before each consultation meeting we shared the model in the meeting chat. The original aim was to give people the chance to read the model before the meeting, so that they had a better idea of what the meeting was about.</p><p>But what this achieved was giving a forum for people to discuss what we were proposing in far more detail than we could have achieved in the meeting. In a meeting of 70 people it is hard to have back and forth discussion. In an instant messaging chat there is more space and time for this.</p><p>Different areas of our department work in quite different ways, and often don’t interact with each other. This means they have quite different ideas about how digital delivery should work. Bringing them together like this allowed them to share their differences and for us to develop a model that worked for as many people as possible.</p><h4>Empowerment</h4><p>There were a few senior sponsors for this work, but they empowered us to develop the governance model in the way we saw fit. We had their backing, but they were happy for us to carry on without checking in regularly. They engaged through consultation meetings like everyone else.</p><p>Having senior support was really important. Without it none of us would have felt comfortable running this project. It is also unlikely people would have engaged to the extent that they did.</p><p>The empowerment they gave us was vital. If we had to check in with them regularly to make sure the model we were developing was ok, then we would still be working on it. Senior people tend to be busy, and they tend to have questions. Delaying consultation meetings so that we could organise review sessions with our sponsors, or so that we could make changes they wanted would have delayed the whole process.</p><p>Allowing our sponsors privileged access to make changes would remove the transparency. There was a clear link between feedback provided and changes made to the model. If changes started being made that we couldn’t justify then people would lose trust in the process.</p><h3>What went badly</h3><h4>Being the only bit of transparency</h4><p>Transparency is great, but it’s also hard. There was less transparency about the rolling out of delivery groups than anyone would have liked. This meant that our work became a magnet for questions about delivery groups in general.</p><p>We were frequently asked questions that were outside of our scope. People were often frustrated that we couldn’t answer these. To handle this we were very clear about the scope of our work at the start of every consultation meeting. But this didn’t stop all the questions.</p><p>Having said that, providing an offline forum for people to discuss delivery groups helped people. Worries could be shared and questions could be answered. This didn’t impact us in the same way that out of scope questions in an online consultation session did.</p><h4>Not being full time</h4><p>Everyone involved dedicated less than half their time to this project. This meant we moved more slowly than we would have liked. We also couldn’t keep to our original commitment of always having the latest version of the model visible. Too often things were in too much of a rough draft state to share.</p><p>Most seriously, juggling multiple pieces of work impacted the wellbeing of the team developing the model. People were passionate about this project, and so wanted to do it justice. But this led to people overstretching themselves.</p><p>Personally I’ve only started to notice the level of stress I’ve experienced over the past few months now that it’s reducing. It’s been different to stress I’ve experienced before as it has come from overwork on something I wanted to be doing. But switching off in the evening has been hard, and I’ve felt really exhausted at times.</p><h4>Being sensitive to people who are impacted</h4><p>One of the biggest challenges is that change won’t affect everyone in a positive way. We’ve talked about streamlining governance, but this will impact the people responsible for running our current governance processes. We were also challenged to be careful with our use of language on a couple of occasions as people on consultation calls felt we were dismissing the work of certain groups.</p><p>The problem with this is that trying to be sensitive and careful often means saying nothing. In this case everyone ends up being worried, because they don’t know what is coming next. Being transparent allows people to plan properly, and shape change where it affects them, but does also mean having sensitive conversations in the open. There’s a balance to be struck and we didn’t get it right all the time.</p><h4>Failing to get a wide spread of people</h4><p>We wanted everyone who was interested in this work to be involved and reached out through our networks to achieve this. Unfortunately the approaches we used targeted people in certain areas. Many of them were people who already thought like us.</p><p>As there was a lot of change going on people were invited to a lot of meetings to discuss change. So people who knew the team running this project were more likely to prioritise our meetings.</p><p>This issue partly stemmed from not having user researchers in our team at the start. They would have taken a more considered approach to recruitment than I did.</p><p>We don’t have easy access to a department wide mailing list, but if I did this again I would have approached our comms team to make sure that everyone was at least invited to the consultation meetings.</p><h4>Lack of familiarity with transparent and iterative working</h4><p>Working transparently and iteratively was new to a lot of people. This meant that when we presented draft governance models, some people took this to be what was happening, rather than a draft they could influence. This provoked some of the more challenging responses.</p><p>I think we did a good job of explaining the approach we took. This problem is worth being aware of, but should get easier the more transparent your department is.</p><h4>Struggles with scope</h4><p>As previously mentioned we were frequently asked questions that were outside of our scope. This ended up taking a lot of time from our consultation meetings. Part of this came from a wider lack of transparency, and part of it from a lack of clarity over what governance is.</p><p>Being really clear about our scope at the start of every meeting helped, but ultimately didn’t make questions go away.</p><h4>Struggles with handover/sign off</h4><p>The one area which we still haven’t got anywhere is with handing over or signing off this piece of work. The board we’re intending to hand the model to hasn’t sat for the past two months. Speaking directly to board members it is clear that they don’t all agree on a governance approach, making it hard to get anything signed off.</p><p>Having said that, the approach we’ve taken feels like it has made sign off redundant. Only time will tell but the governance model has got momentum of its own and lots of people are already talking to us about implementing it. It would be nice to be able to hand it over to someone who can iterate it, but that might have to stay with us for now.</p><h3>Overall</h3><p>The best thing you can say about this approach to transformation is that it has worked. It’s been hard, and at times was a bit rocky, but the work we’ve done is already being used. It hasn’t been as quick as I would have liked, but it has been quicker than change usually is in this department. So were I to try and change something of this scale again I’d use a similar approach.</p><p>Next time I’d want a team that was more full time. This is tricky as people can’t just pause their day jobs. But having people experienced in the department leading the transformation is an important part of the process. There isn’t an easy answer, but it’s definitely something to work on in future.</p><p>I would set up a dedicated chat channel for the transformation. Having a chat associated with each consultation meeting was great, but it meant conversation was lost in between meetings.</p><p>Having a better approach to getting a wider consultation group at the start would be a lesson for next time. I think this is something our comms team could have helped with and I regret not approaching them about it.</p><p>Having said all that, as a first go at a transformation of this scale, I’m really proud of what our team has achieved. We did something different to our department’s usual approach to transformation, and it worked really well. I’m pleased with the outcome, and we’ve learnt a lot to do an even better job next time.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1d4f36a0fa5c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Agile organisational design in practice — week 41]]></title>
            <link>https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-41-11b5ae10bcd8?source=rss-433c2ef9fdc5------2</link>
            <guid isPermaLink="false">https://medium.com/p/11b5ae10bcd8</guid>
            <category><![CDATA[organizational-design]]></category>
            <category><![CDATA[uk-government]]></category>
            <category><![CDATA[agile]]></category>
            <dc:creator><![CDATA[David Thomas]]></dc:creator>
            <pubDate>Fri, 20 Jun 2025 15:45:57 GMT</pubDate>
            <atom:updated>2025-07-04T16:10:25.339Z</atom:updated>
            <content:encoded><![CDATA[<h3>Agile organisational design in practice — week 41</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dnWq4xibFwMjeiyhBRkgSw.png" /></figure><p><a href="https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-34-649875bb907c">A little over a month ago I wrote</a> about experiencing delays in sign off for the governance model I’d been helping develop. The month since has been filled with a few holidays, me getting up to speed with my new role of delivery group lead and handing over my previous role. The board which is meant to be signing off the governance model has been cancelled a couple more times. Most importantly however we have published the <a href="https://defra.github.io/delivery-group-governance/">final draft of the governance model</a>.</p><h4><strong>Publishing a draft</strong></h4><p>Transparency is a key part of my approach to organisational design. From the start we shared our work as widely as we could. It was hard to share externally to our organisation as we didn’t have a platform to do this on, or at least not one that supported the constant iteration we were after. Now that the model is relatively stable, publishing it on GitHub makes perfect sense.</p><p>There are no big secrets in what we’re doing. We want people outside our organisation that we work with to understand how we work, and we want to share how we work as an example to others.</p><h4><strong>What’s in the draft</strong></h4><p>The draft contains a governance model based on <a href="https://www.flightlevels.io/what-is-flight-levels">Flight levels</a>. In this model we’ve moved away from group decision making through boards, to empowered individuals. We’ve also focussed on empowering people to make decisions at the right level. So in our model people working at a strategic level wouldn’t be able to overrule people working at an implementation level. Instead the aim is for people to follow the strategy because it’s the right thing to do.</p><p>The core roles in the model are business/service, product, delivery and technology. Business people or service owners are responsible for setting outcomes and providing funding to achieve those outcomes. Product people are responsible for deciding what to do in order to achieve those outcomes. Delivery people build the teams to do the work. And finally technology people decide on what technology is best suited to build with.</p><p>The assurance model is centred around</p><ul><li><a href="https://www.gov.uk/government/publications/digital-and-technology-spend-control-version-6/digital-and-technology-spend-control-version-6">spend control</a> — permission to spend based on the approach you’re taking and how you’ve performed so far</li><li><a href="https://www.gov.uk/service-manual/service-assessments/how-service-assessments-work">service assessments</a> — assessments at the end of each phase to ensure you’re meeting the <a href="https://www.gov.uk/service-manual/service-standard">service standard</a></li><li>operational service design review board — a process to ensure that the service being built will work effectively when fully operational</li></ul><p>On top of those three required things there are other suggestions on assurance that you may want to do within delivery groups. These cover agile approaches to change management and architectural assurance, as well as things that wouldn’t always be considered as assurance but fulfil that role such as private beta and show and tells.</p><h4><strong>More sign off delay</strong></h4><p>The board that is meant to be signing off the governance model hasn’t happened at any point in the past couple of months. This is due to members being unavailable. I’ve had positive discussions with a few of the board members and so am going to aim to hand this over or sign off the work by correspondence next week.</p><h4><strong>What’s next</strong></h4><p>Next up I’ll be looking forward to the feedback we’ll hopefully be getting on our governance model and the approach we’ve taken to developing it. I’ll be using this to feed into a bit of a retrospective I’m planning around the transparent transformation approach. I’m also carrying on trying to get to grips with the work in my delivery group. <a href="https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-43-223bbc2eb9d0">I’ll keep you posted</a> on how I get on!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=11b5ae10bcd8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Agile organisational design in practice — week 35]]></title>
            <link>https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-34-649875bb907c?source=rss-433c2ef9fdc5------2</link>
            <guid isPermaLink="false">https://medium.com/p/649875bb907c</guid>
            <category><![CDATA[uk-government]]></category>
            <category><![CDATA[organisational-design]]></category>
            <category><![CDATA[agile]]></category>
            <dc:creator><![CDATA[David Thomas]]></dc:creator>
            <pubDate>Fri, 09 May 2025 15:04:21 GMT</pubDate>
            <atom:updated>2025-06-20T15:50:49.474Z</atom:updated>
            <content:encoded><![CDATA[<h3>Agile organisational design in practice — week 35</h3><figure><img alt="A person says “Can I get my paper signed off?”. Another person standing in front of an empty table and chairs replies “Sure — the board next meets in 2035”" src="https://cdn-images-1.medium.com/max/1024/1*dn00_GxUiV6pU9UfzJM7_g.png" /></figure><p><a href="https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-34-149195255b53">Last week</a> I wrote about starting up a new delivery group and our progress towards getting a new governance model finished. This week I’ve found out what our delivery group is expected to work on this year. We also had the board meeting for signing off the governance model postponed.</p><h4><strong>Delayed sign off</strong></h4><p>This week the board meeting where we intended to hand over our governance model was postponed as several board members were unavailable. The next meeting will be in a couple of weeks. This is a good example of the problem with escalating sign off to boards.</p><p>If an individual had been empowered to sign off our work, then that individual being unavailable on a given day wouldn’t have had a serious impact. They simply could have signed off the work the next day. But because a board has to take this decision then that decision is delayed by two weeks because of a clash with a conference.</p><p>It would be even better if sign off was not required at all. As a team we’ve been the closest to this work for the past few months, and are best placed to say when it is ready. This would have avoided any delay at all, as we could release the governance model at the time of our choosing.</p><p>In reality this is quite close to what has happened. The majority of our governance work consists of sensible defaults. That is, things that you should do unless there is a good reason in your content to do otherwise. By developing these sensible defaults in the open, the latest version of them has always been visible. This means that members of other delivery groups have already started to approach our team for support in implementing them.</p><h4><strong>Delivery group first steps</strong></h4><p>There are two programmes of work that exist in the delivery group I’m going to be leading. They have both been running for several years and are currently relatively stable. This means that they already meet one of the aims of a delivery group — to have stable teams working in a consistent business area.</p><p>So instead of leading any sort of dramatic change, I need to help with iterative improvements in the short term, and ensure that stability continues in the long term.</p><h4><strong>What’s next</strong></h4><p>Next week we’re looking at writing the final bits of content for the governance model. Once that has happened we’ll hand it over so that it can continue being iterated longer term. I’ll let you know <a href="https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-41-11b5ae10bcd8">how I get on</a>!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=649875bb907c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Agile organisational design in practice — week 34]]></title>
            <link>https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-34-149195255b53?source=rss-433c2ef9fdc5------2</link>
            <guid isPermaLink="false">https://medium.com/p/149195255b53</guid>
            <category><![CDATA[organisational-design]]></category>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[uk-government]]></category>
            <dc:creator><![CDATA[David Thomas]]></dc:creator>
            <pubDate>Fri, 02 May 2025 15:54:47 GMT</pubDate>
            <atom:updated>2025-05-09T15:06:11.038Z</atom:updated>
            <content:encoded><![CDATA[<h3>Agile organisational design in practice — week 34</h3><figure><img alt="A person holding a page saying “Governance” says “I’ve finished”. Another person responds “Great, now go and implement it”. The first person replies “Err…. OK!”" src="https://cdn-images-1.medium.com/max/1024/1*D9ijTY7SG6Tqk3t86dKS3w.png" /></figure><p><a href="https://medium.com/@david.thomas1/two-people-agile-organisational-design-in-practice-week-30-2149ad95f8a8">A month ago</a> I wrote about the importance of avoiding escalations if you want a flat hierarchy. During that month I was appointed to the role of delivery group lead. This means that I’ll now be responsible for implementing the governance model I’ve spent the last few months helping to develop.</p><h4><strong>How to implement a delivery group</strong></h4><p>The most important thing to start with is to find out what the delivery group I’m leading is responsible for. The organisation I work for has lots of small services, many of them legacy, which means this isn’t as easy a task as it may sound. I’m currently working on one of the large services within the delivery group which gives me a good start. But the first job is to speak to the director responsible for my business area and understand what it is they want the delivery group to achieve.</p><p>Alongside this I’ll be immediately looking at the people required to work in the delivery group. The governance model we’ve developed has roles split into strategic, coordination and implementation. From my perspective if I can fill the coordination roles and the other strategic roles, then those people should take care of everything else.</p><p>Once I know what the work is and there are good people to complete the work, the rest is a case of giving those teams time to mature.</p><h4><strong>Funding</strong></h4><p>The one other thing that is on my mind longer term is funding. Currently our work is funded as projects and programmes. We’d like to move to a model where the delivery group is funded and works on the highest priority items. We’re just into a new financial year, so this doesn’t need resolving now, but will be another thing high on my list when I speak to our business director</p><h4><strong>Finishing off the governance model</strong></h4><p>Over the last month the team I’ve been working with has produced two more iterations of the governance model. We now have sensible defaults for governance, sensible defaults for assurance and required assurance. The sensible defaults are for delivery group leads to implement as they see fit within their delivery group. The required assurance is required across all delivery groups. The required assurance mostly consists of <a href="https://www.gov.uk/government/publications/digital-and-technology-spend-control-version-6/digital-and-technology-spend-control-version-6">spend control</a> and <a href="https://www.gov.uk/service-manual/service-assessments">service assessments</a>, which as a department we’re mandated to do by the Government Digital Service.</p><p>We’ve also added a series of case studies showing how the governance and assurance processes work. This is particularly important for governance. Our current governance processes are very process heavy and typically involve multiple boards. Moving to a role based governance model with individuals taking decisions is quite a big shift, so it is important to be clear how this will work.</p><p>Finally we’ve tapped into our wider network of designers to produce some great visuals to help describe how the governance model works. The one thing that has been hammered home to me over and over again is how important a team is in completing a piece of work like this.</p><h4><strong>What’s next</strong></h4><p>Next week we’re speaking to the board that sits across delivery groups to present the governance model. The intention is to hand the work over at this point, or at least agree what needs to be done before it can be handed over. I’ll let you know how I get on! (<a href="https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-34-649875bb907c">Spoiler alert: badly</a>)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=149195255b53" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Agile organisational design in practice — week 30]]></title>
            <link>https://medium.com/@david.thomas1/two-people-agile-organisational-design-in-practice-week-30-2149ad95f8a8?source=rss-433c2ef9fdc5------2</link>
            <guid isPermaLink="false">https://medium.com/p/2149ad95f8a8</guid>
            <category><![CDATA[organisational-design]]></category>
            <category><![CDATA[uk-government]]></category>
            <category><![CDATA[agile]]></category>
            <dc:creator><![CDATA[David Thomas]]></dc:creator>
            <pubDate>Fri, 04 Apr 2025 16:43:06 GMT</pubDate>
            <atom:updated>2025-05-02T15:56:10.115Z</atom:updated>
            <content:encoded><![CDATA[<h3>Agile organisational design in practice — week 30</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*g2Uzxp7amHkHGp5SYWeP4g.png" /></figure><p><a href="https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-29-fa3d0c72c9fa">Last week</a> I talked about handing over the work I’m doing on governance rather than signing it off. This week the team I’m working with on developing the governance model held another consultation. This one was led by my colleagues in operations. I’ve also been thinking about escalations, and how and why they are the key to a flat hierarchy.</p><h4><strong>Escalations and flat hierarchies</strong></h4><p><a href="https://medium.com/p/22b3711b4a99">I wrote a couple of months ago</a> about how to act like you’re not in a hierarchy, even when you are. I talk a lot about pushing decisions down, but really it should go one step further and not have hierarchy as an escalation route at all.</p><p>I’ve been running a small programme with three teams and a total of 25 people for the last 18 months. In that time I haven’t taken a decision in response to an escalation. So I wanted to explore the kinds of escalations that occur and how to avoid them.</p><h4><strong>Two people can’t agree</strong></h4><p>The most common type of escalation is that two people can’t agree, and so ask someone more senior to mediate</p><p>This issue indicates that there is not a clear separation of responsibilities. It should be clear whether a decision is a product management decision, a delivery management decision, or one that falls to another speciality.</p><p>I came across this issue most often on a programme where policy, delivery and operations all worked separately. They each believed that they were in charge of how the service should work. If they disagreed then this was quickly escalated to each of their directors.</p><p>Escalating decisions like this slows down decision making. If more senior people receive a lot of escalations then they can be slow to respond to them. It also prevents them from focusing on strategic work.</p><p>Senior people are also likely to make worse decisions in this situation. Because they are not close to the decision they will just “make a call”.</p><p>Someone just making a call doesn’t resolve the issue. A product manager who is in charge of the direction of a product can expect everyone to follow their vision. If three different people have three different visions, the ones who lose out are unlikely to change their vision just because they’ve been overruled.</p><p>To prevent this from happening everyone should have clear responsibilities. This does not necessarily mean a RACI matrix. If people are using standard roles such as delivery and product management then this should be obvious. People in different roles or newer to these terms may need a conversation to make it clear.</p><h4><strong>We need sign off to proceed</strong></h4><p>It is common to need sign off from senior people to proceed from one stage of a delivery to another.</p><p>This tends to slow down delivery, as getting agreement from those senior people can be slow. If the senior person does not agree to the plan to proceed the team can be left in limbo while they replan how to proceed.</p><p>It also leads to poor decision making. As the senior people are further away from the work being done they have a less clear view of exactly how the work should proceed. But by providing a plan on what will happen next you’re inviting a response to that.</p><p>As well as slowing down delivery at certain stages, this means that a delivery that is off track will only likely be raised at those points. This means something can head in the wrong direction for a long time before being corrected.</p><p>Instead people working at a senior level should be able to provide appropriate input at any time. Without that input the team should be free to carry on.</p><h4><strong>This decision is too big for us</strong></h4><p>Sometimes a decision is of a scale that people don’t feel comfortable making it without more senior cover. This indicates that people do not feel they have appropriate support from senior management.</p><p>This has happened recently with changes to governance processes. The people who owned them were not willing to make big changes to them without consulting more senior people.</p><p>The senior people can’t make better decisions, they can only feel more comfortable carrying the responsibility. But more junior people should be made to feel comfortable by being empowered and supported.</p><h4>Things are going wrong</h4><p>Things are often escalated to senior people when they’re going badly. Even here it is difficult for the senior person to resolve. They usually won’t have the time or focus to properly resolve any issues. And without fixing the underlying problems that have caused things to go wrong, things will simply go off track again as soon as senior attention is elsewhere.</p><p>Instead people working in the team where things are going wrong should be empowered and supported to fix things.</p><h4><strong>This needs strategic input</strong></h4><p>Sometimes people need to escalate decisions that fall out of their areas. They may be going outside of an agreed budget, or trying something that falls outside of the agreed remit of their work.</p><p>In this case an escalation is reasonable as it is not a decision that the person escalating is in a place to take.</p><h4><strong>Be your grade</strong></h4><p>I was recently asked at a talk on flat hierarchies about how to handle the situation where your area of work is trying to act in a flat way, but you’re working with people who are much more hierarchical. In response to this I suggested people should simply “be their grade”.</p><p>To explain this imagine a product manager in a delivery team working with a policy colleague. Digital act as if their hierarchies are flat, so the product manager feels empowered. The policy colleague disagrees with the product manager’s vision and escalates to their director. A good response to this escalation from the product manager’s director would be to attend escalation meetings, but not try to resolve things, or even back their product manager. Instead, just being there and supporting the product manager’s right to make the decision should be enough.</p><h4><strong>What’s next</strong></h4><p>Next week I’m on leave, but I’ll <a href="https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-34-149195255b53">update on progress with our governance model</a> when I’m back!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2149ad95f8a8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Agile organisational design in practice — week 29]]></title>
            <link>https://medium.com/@david.thomas1/agile-organisational-design-in-practice-week-29-fa3d0c72c9fa?source=rss-433c2ef9fdc5------2</link>
            <guid isPermaLink="false">https://medium.com/p/fa3d0c72c9fa</guid>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[uk-government]]></category>
            <category><![CDATA[organisational-design]]></category>
            <dc:creator><![CDATA[David Thomas]]></dc:creator>
            <pubDate>Fri, 28 Mar 2025 16:00:37 GMT</pubDate>
            <atom:updated>2025-03-28T16:00:37.166Z</atom:updated>
            <content:encoded><![CDATA[<h3>Agile organisational design in practice — week 29</h3><figure><img alt="Person 1 holds a pen over a page titled “Governance”. They say “Here’s your governance model”. Person 2 says “I want it pink and in comic sans”. Person 1 says “Good luck with that”" src="https://cdn-images-1.medium.com/max/1024/1*Bx91iejnnen4r-eJksjiGg.png" /></figure><p><a href="https://medium.com/p/e969a04d7813">Last week</a> I worked on updating the <a href="https://www.gov.uk/service-manual">service manual</a>, and sharing ownership of the governance model for delivery groups that I’m helping develop. This week I’ve been thinking about how to approach sign off for the governance model, and what sign off means when things should evolve iteratively.</p><h4>What do I want signed off</h4><p>The governance model is broken into two parts — governance and assurance. We’ve described the model as being sensible defaults for governance, but this isn’t entirely true. The governance part of the model is up to the delivery group lead to implement, and so is a set of sensible defaults. The assurance part of the model sits outside delivery groups, and so is imposed on them.</p><p>There isn’t a need to sign off sensible defaults for governance. It is up to the delivery group lead to set up their delivery group as they see fit. This work is to provide them with guidance. As the assurance model will be changing across the department, this does need senior buy in.</p><h4>Sign off shouldn’t be final</h4><p>The sensible defaults for governance need to be iterated as long as they exist, and so need long term ownership. This means instead of sign off we need to handover the defaults to a group that will maintain and share them.</p><p>The people who own our assurance processes want senior support before making big changes to them. Once we’ve done that the detail of these assurance processes needs to be owned by the teams that run them. This means again that we’re not looking for sign off for a full detailed assurance process, but just a set of big changes to that process</p><p>So instead of looking for full sign off for our work, we can instead hand over most of it, and get specific sign off for some of the bigger changes we’re looking to make.</p><h4>The advantage of handover versus sign off</h4><p>Sign off is difficult to get. It is difficult to know who to get sign off from, and when you do find out, people usually want changes before agreeing to something. That’s because sign off is seen as a one off thing. The opposite of iteratively developing something</p><p>By handing something over there is a clear indication that work is expected to continue. It just needs a permanent home. This also provides a clear end date for the team working on this governance model.</p><h4>What’s next</h4><p>Next week we’ll be running our third consultation on our governance model. This time focussing on the work of our operations department.</p><p>I’m also hoping to complete a second draft of the organisational design page for the service manual. This will be available to share more widely for consultation.</p><p>And finally I’m going to talk with our sponsors about the proposal to hand over our work rather than look to get it signed off. As ever, I’ll let you know how I get on!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fa3d0c72c9fa" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>