Maps
Liam Maxwell
- Gov must move silos -> platforms
- Introduction of open standards and data
- Coding in the open means sharing
- Common standards means common tools
- Outcome based teams
- Change needs spend controls
- Square of despair: procurement, security, legacy, capability
- Institute for Government encouraged tyranny of agile
- How can we get to AGILE(tm)? Doing it properly
- Mapping tells us which techniques to use
- Cassian Young: how would you do this in a dept. and how would you use a map?
- Mapped incident management using maps
- Identified utilities, things that could be SaaS
- Identified areas of custom
- What's up with hosting?
- Gov spending £1.6bn per year on hosting
- Lots of hosting comes with added lock-in in contract
- Gov lock ins broken by Crown Hosting, the "Hotel California" where you check in your legacy but where it can never leave
- Depts can then move on
- Used open to unlock change
- Competition could then drive technology model
- High degree of open source
- High engagement from SMEs
- What is a minimum viable agency?
- PaaS, Registers, Open source -> multidisciplinary teams -> GOV.UK, supported by user research, continuous delivery
Hadley Beeman
- Architect, Open Data and Standards, W3C
- Talk about what standards are, and how they are made
- W3C speak for standards is "recommendation"
- Standards building is a game of werewolf. Everyone has their agenda. Balancing the needs of peoples agendas are key to good outcomes.
- Standards push things towards commodity, but some things might be best left in product/commodity
- Private browsing is different in each browser. Complexity does disservice to users. Hampers innovation for the rest of the web.
- Browsers prefer to innovate on their own.
- Some features of private browsing, for example, may be in different places on the x-axis of a Wardley map.
- WebRTC emerges from proprietary tech like Sype, Facetime. WebRTC makes these things interoperable (???)
- Presentation API: 1st + 2nd screen. Presentations + speaker notes, Game + controller, media "flinging" (Chromecast)
- Vehicle information service data: various vehicle metrics - technical and infotainment systems. Control you car from your phone, have a virtual manual on your phone (see your oil level)
- WebVR
- Community groups in custom->product space, working group & recommendation track in the product->commodity space
- Good standards start with implementations
- EMEA. Challenge we were faced with was that the big players were going to do a thing anyway, possibly with less transparency, less concerns addressed.
- Rec. track is when things start to get very stable, and you can start to build things on top
Julie Pearce
- Openness, Data, and Digital, Food Standards Agency
- Delivering new tech + digi services for FSA
- Ensure the food you eat is safe, and is what it says it is
- Vast estate outsourced to Crapita
- About half were "shadow IT"
- Brexit would require vast changes. No idea what would be required.
- Decision to crack on and get out of Crapita and work it out as they went along.
- Undaunted by perception of "many dragons' being there, they went and mapped it all
- Lots of duplication, lots of complexity, many things too far left
- Plan to start at the bottom: uncontroversial as far down the value chain, things that won't be affected by EU exit
- Lots of savings made by moving infra
- Phase of tidying up
- Untangling Capita contract using mapping
- 90% of what we do is exactly the same
- Design everything in future as if we are the most boring, bog standard regular, or other organisation
- Next stage map leverages commodity in bottom right
- Modern, simple, cheaper things that work well. People don't mind they are not "special", they like the better tools.
- What about the uncertainty? Well, on the basis we're 90% the same as anything else we'll be able to adapt.
Paul Shetler
- Hundreds of sites in Australian govt
- Using engagement as a metric, but... nobody wants to engage with you. People want to get shit done.
- Asked agencies to do 3 things: 1) start with user needs, 2) experiment and deliver the smallest improvement quickly, 3) agile + multidisciplinary teams
- Timeboxes mitigate ocean boiling. Apply strict timeframes, force descoping.
- Break the big system down. Where are the bottlenecks?
- Iron law of bureaucracy: organisations want to maintain themselves in their current form
- Deskilled workforce: outsourcing and privatisation deskill government, reduce skills and breed incompetance
- Square of despair: inappropriate governance, procurement, IT, funding
- Australia context: digital skills shortage country-wide. It's a problem for the country.
- "Internet marketplace is red in tooth and claw" i.e. Amazon will eat everyone, Aus needs to look at industrial policy
- Concept of spinning out "labs" from your org. If your org is operating in the square of despair the product of your labs will die.
- Hackathons and design theatre: same problem. If you can't actually progress the outcomes they are counter productive.
- Lipstick on a pig front ends: that is not serious
- "Evolution not revolution" should be banned. This is a path to go out of business.
- 2-speed / bi-modal: you are creating new legacy, setting up warring camps. Winners and losers.
- Digital cultism / agile everywhere. It's not always appropriate. If it's not new, unusual, unknown - why would you use agile?
- You need the political will to change. Inertia is the name of the game. People have been there forever. You need political pressure to break inertia.
- Put a senior minister in charge. End of career not beginning. Fine with causing controvery, they won't be loved.
- Change the management
- If you're under existential threat, maybe get rid of the frozen middle.
- Work with human nature.
- Fix the basics. Answer the friggin' phones.
- I.e. NSW did it in Aus. Earned a lot of political capital with users.
- Change has to go into the back office. What is the user trying to get done? Help them.
- Automate everything. We are told people want a human touch. In reality people prefer a consistent, high quality outcome. Whether or not that's a human.
- Stop talking about "engagement"
- Square of despair driven by downskilled staff. You must radically upskill, from the bottom up.
- New digital professionals, recruit from the front line. I.e. digital academies in UK.
- Dismantle the square of despair. Funding is a huge problem. If you need a big business case for an experiment you're doing it wrong.
- Procurement in days if not minutes.
- Simplify and consolidate governance.
- Institute spend controls. Nothing is going to work without that. Gives you a map of what's going on, can stop bad behaviour, steer towards to right things.
- End the digital / IT / product split.
- Fit governance to the problem, method, aptitude.
Chris Adams
- Mapping by Example
- How do I talk about what's hapening in my org?
- Business model canvas? Product roadmap?
- Tells me the what and the now, but not the why and the future
- We are planning to detect icebergs by ramming them
- Useful features of maps: anchor: keep user in the picture, position: how are we set up to meet their needs?, movement: how might things change
- Maps help with movement in a unique way, are information dense, and explicit
- Mapped value chain for website company + added evolution
- It looks messy at first, that's fine
- Client doesn't say "good job on configuring Postgres Chris!". Doesn't happen.
- Will I get better at ops than Amazon? No
- Will clients expect us to be familiar with the things on the right? Yes.
- Growing Communities company in Hackney. Legacy website, lots of problems. What to do? Map.
- What the core user need? Veg boxes.
- Spent a couple of times looking into writing an open source thing but found out how hard it is
- "Mapunditry"
- How the battle of cloud was lost example. Map Amazon vs traditional.
- Another example: Why would FB send the GDP of Iceland on WhatsApp?
- Monetizing the social graph: the people who know each other
- WhatsApp was a messaging app choking FBs value chain: keep in touch with your family without eyeballs in front of FB ads
- How to get practice mapping?
- AWMUG (Amateur Wardley Mapping User Group)
- https://awmug.org
Liz Keogh
- Cynefin (sounds like "kernevin")
- Epiphany vs Apophany
- Apophany, Claus Konrad coined, means the kind of epiphany schizophrenics have: seeing patterns that aren't there.
- Example: a photo of the surface of Mars that looks like an elephant.
- Some conservation groups don't recognise the discovery that there are two species of African elephants, because their tracking systems can't cope
- Apophenia simultaneously extremely useful and highly hazardous
- We filter the data we observe
- Beliefs dictate how we filter data: this produces confirmation bias
- Pareidolia, actor-observer bias, Google effect... there are so many biases we suffer from
- Our brains are broken when it comes to understanding uncertainty
- There is an assuption that there is order and predictability
- Cynefin framework. Obvious things, complicated things, chaos, complexity
- Obvious things, complicated things => predictable outcomes
- Chaos => need to act quickly
- Complex => cause and effect can be correlated, but only in retrospect
- Obvious: sense, categorise, respond
- Complicated: sense, analyse, respond
- Chaotic: act, sense, respond
- Complex: probe, sense, respond
- Emergent outcomes, processes
- In the middle is disorder, we can Eapply our preferred pattern to disorder but it tends to chaos
- Systems are complex and adaptive, agents can change the system
- Estimating complexity
- 5: Nobody has ever done it before
- 4: Someone outside the org has
- 3: Someone in the company has done it
- 2: Someone in the team has
- 1: We all know how to do it
- 5 to 1 moves from complex to obvious
- Things that are 1s: should they be commodities?
- Safe-to-fail Rpbes: know if it's succeeding or not? Have lever to dampen or amplify. Coherence: a realistic reason to think the probes might succeed.
- Coherence: "Can you give me an example"
- BDD approaches: given x then y, z
- We often don't try things because there's a "but it won't work" voice in our heads
- Correlation in retrospect
- Second order effects: more police on the street = more crime reported because people reporting crime to them
- Cognitive Edge "Sensemaker", gathers
- Identify "dispositions" in systems, find ways to change from bad ones to good ones
- Where is change landing and succeeding? Listen out.
- Use existing disposition
- Complexity is dominated by relationships. Don't just think about nodes, think about relationships. That's what you are changing.
- Relentless positivity. If you want to get things going you need to be relentless.
- Get people started, break the initial inertia.
- We notice what we do wrong more than what we do right.
James Findlay & James Duncan ... Missed first half of talk
- Von Neumann & his contribution of computing
- ENIAC, for calculating the trajectory of artillery shells was being built
- Von N had been working on Manhattan Project, but worked out that ENIAC could be modified to a general purpose machine
- He got it to do a load of work on atom bomb things
- EDVAC paper proposed architecture which we still use today in x86 and similar
- Von N made use of commodity valves... as if he had a map
Adrian Cockcroft, @adrianco
- Mapping AI
- VP at AWS
- OSS programme at AWS
- Was at Netflix doing OSS
- Part of Cloud Native Foundation
- AI was a thing, and they identified MXNet as an academic project they might sponsor
- They had a go at mapping MXNet
- Improve docs, make easier to migrate from TensorFlow
- Move it from being a uni project to be incubated by Apache
- Simon W contributed to the maps, introduced concept of pipeline for research (custom) to pre-trained models
- Introduced plays such as investing in high level frameworks for MXNet, partner (and control supply) of best new GPU chips (NVIDIA) etc.
- Looked at every team in Amazon and what they are doing in AI: shops team etc.
- MXNet used by WolframAlpha, Pinterest, self driving car systems
- Future state map: commoditised everything, academia continues to feed the models
Tal Klein
- Mapping in the context of science fiction
- Think about Dr Evil's motivation in Austin Powers: world domination
- Strategy: sharks with frikkin' laser beams
- Actually sharks are not a commodity. To achieve sharks with laser beams he'd have to become the most ecologically friendly supervillain ever.
- There were a load of bad assumptions.
- We assume that maps solve our strategy problems. We can't draw maps to justify our conclusions.
- The Martian does it well.
- Ready Player One makes a fundamental assumption: everyone will be plugged into a VR framework. Mixed reality actually more likely.
- To take another supervillain... like Putin. Leverage commodity things, like social media, and add propaganda and fake news etc.!
- In the Punch Escrow he writes about a future where he has used mapping to try to work out what it will look like.