Earlier this year Nicola and I decided to move house. After 10 years on Electric Avenue we’ve moved to suburban West Norwood.
Mostly driven by the inevitable white goods purchases that go with moving home, I decided to join Which? AKA The Consumers’ Association. It turns out that it’s also a pretty interesting organisation.
The idea of a new organisation to advise consumers was originally proposed by a Labour Party researcher called Michael Young (who was also responsible for drafting the 1945 manifesto and helping found the Open University, AKA ‘University of the Air’).
The idea didn’t make the cut of the 1950 manifesto, but it lived on and Which? ended up being published from a garage in Bethnal Green in 1957.
It’s aim was to help citizens navigate the rapidly changing, relatively new and occasionally dodgy market and marketing of consumer goods.
History aside, I mostly joined for the cooker reviews. The do proper standardised physical tests of the sort that are never going to feature in an Amazon product review: things like test how equally toast browns or if your fridge will stay cold on a hot day.
As part of membership you get the monthly magazine. Flicking through it, the lack of reviews for digital services and issues about privacy/trust really jumped out at me.
They do reviews of tablets and phones, but they are mostly about build quality and value for money, but again, not really about trust of those devices. Are they full of security holes? How quickly are bugs fixed? Are the privacy settings understandable by an average user? How many dark-patterns do they use? Is the cryptography going to see off a casual hacker? How good is the data protection policy?
These issues are the modern equivalent to the ‘sharp trading practices’ of the post-war years that Which? set out to expose, but what institutions are there to look out for the average consumer of today’s digital services?
Luckily, this seems to be an issue that Which? have recognised. A few months ago the magazine published a call to members to stand for the ‘Which? Council’ (the management body of the charity) and specifically asked for members who have experience of digital to stand for election.
I decided to try and stand for election, but party due to moving house, partly because the process is quite complex, I failed to see the process through to the end.
However, there remains a real opportunity to help an organisation, that is trusted by citizens, make digital services better. So, below is what I learned about the process in the hope that developers, digital policy professionals and designers will consider engaging with the organisation and hopefully voting in or standing for election next year.
## How to stand to become a member of Which? / stand for election to the Council
Join Which?. This makes you an ‘Associate Member’. You get the magazine and can access reviews online.
Once you are a member of Which? you then need to apply to become an ‘Ordinary Member of the Consumer Association’ (the official name of the charity). This means you can nominate others for election, vote, attend the AGM and stand for election. To become an ‘Ordinary Member’ you need to print and complete this form (PDF) and send it to: Company Secretary (Council Nominations), Which?, 2 Marylebone Road, London NW1 4DF. (note: your details will be kept on a public-ish register of members and may be made available to other members)
To stand for election, you need to email councilelection@which.co.uk. The Company Secretariat team will then send you a ‘candidate information form’ (basically your CV for the election) and a nomination form.
You need to find 5 other ‘Ordinary Members’ to nominate you. There are apparently about 8,600 Ordinary Members, and the secretariat posted me a print-out of the contact details of the 30 or so who live in South West London, I guess so I could visit them to ask for a nomination.
The original forms must be returned before the AGM (which happens in the Autumn). Any ordinary members that attend the AGM can vote for a Council Member, as well as any changes to the constitution.
Earlier this year, when I was working with Jamie, Tom, Anna, Paul, Stephen and Adam on a vision for Government as a Platform, I got stuck on the Central Line on the way back from work and ended up trying to distill all the things the team were talking about. The list below was the result.
I’m posting it here because Jamie keeps on telling me I should (he’s normally right), and in case it’s useful to anyone who happens to find themselves redesigning a government:
Split data from services. Hold it in organisations with appropriate accountability (central government, local government, professional bodies) and make the quality of the data independently verifiable.
Services can be provided by any layer of government, and by commercial or third sector orgs. It’s OK when they overlap, complement and duplicate.
It is possible to interact with multiple layers of government at once while respecting their organisational and democratic sovereignty.
Build small services that can be loosely joined together however citizens like. Do not try and model the whole world in a single user experience, you will either fail or build a digital Vogon.
Put users in control of their data. Millions of engaged curators are the best protection government has against fraud, and that citizens have against misuse.
A user not having to understand government does not mean obfuscating the workings of the system.
The system should actively educate people about how their democracy works and where power and accountability lie. Put transparency at the point of use.
Be as vigilant against creating concentrations of power as you are in creating efficiency or avoiding bad user experiences.
Understand that collecting data to personalise or means test a service comes at a cost to a users time and privacy.
Sometimes the user need is ‘because democracy’.
Notes:
To repeat the intro, definitely not all my own work, but they are my words, so where this is wrong it is my fault.
Lots of government services require their users to report when things in their life or an organisation change.
This places a lot of responsibility on the user - they need a good mental model of the service to know what to report, when they should do it and how. It also generates a need for lots of secondary transactions and services: update this, report that, change this, re-apply for that.
The ‘digital assistant’ approach to designing public services could start to make things simpler and reduce the number of ‘Report an X to Y’ style government transactions.
After all, if services understand your past circumstances, why can’t they use those circumstances to ask you the right questions?
Here are 7 possible* patterns (there are probably many more).
## 1) Recurring change
Some circumstances need updating on a regular basis (things like monthly childcare costs). The ‘recurring change’ pattern notifies users (via push alerts or sms) that they need to provide some information. The service should be smart enough to know the optimal number of days to ask this before any deadlines.
A pictogram of a ‘recurring change’ user interface
## 2) Future confirm
If a user reports a temporary change of state, for example they are going on holiday or taking their car off the road, the service should be able ask the user if that state has passed.
A pictogram of a ‘future confirm’ user interface
## 3) Date determined confirm
Similar to ‘future confirm’ there are some circumstances that the service should be able to determine from information it already holds, for example if it knows the user has a child of a certain age.
A pictogram of a ‘date determined confirm’ user interface
## 4) Recurring confirm
A ‘dead man’s handle’ style confirmation, so the user has to actively confirm: “does your cafe still have 12 tables on the pavement outside your business?”.
A pictogram of a ‘random ignore to confirm’ user interface
## 5) Recurring ignore-to-confirm
As above, but inaction is taken as confirmation.
A pictogram of a ‘random ignore to confirm’ user interface
## 6) Random change
Ask a user to submit new information on a subject at random intervals to help keep their data up-to-date.
A pictogram of a ‘random change’ user interface
## 7) Cascading updates
Sometimes the service will be able to determine if a change in a particular circumstance is likely to have caused a change in a related circumstance. For example, registering for a particular license or moving premises may ask the user to confirm information relevant to a related tax.
A pictogram of a ‘cascading updates’ user interface
- No, they have not been tested, but then you can’t build what you can’t think of in the first place.
This has been sitting in my Google Docs since May, so I figured I’d just publish it here.
“This Place Is Ours” is the working name for an app that helps people club together to protect their local pub, skate-park, village hall or park by adding it to the Register of Assets of Community Value.
The Localism Act 2011 and associated regulations require local authorities to maintain a list of assets of community value.
An asset of community value is a building or bit of land that “furthers the social wellbeing or social interests of the local community”
Assets get added to the register when they are either nominated by a constituted local group (like a parish council) or when 21 people (who must be registered voters in the same borough) club together.
Once added to the register, the asset is subject to stricter planning regulations (eg a supermarket wanting to convert a pub will need full planning permission) and the local group may apply for a 6 month period to attempt to buy the asset if it is proposed to be sold.
Anecdotally, this power seems to get used reactively, when there is a known threat to a building, rather than proactively.
“This Place Is Ours” will make it easy for 21 people (who do not need to know each other) to prepare a valid application and submit it to a local authority.
To test the concept, an alpha project will aim to get the majority of pubs (and maybe another category eg adventure playgrounds) in the London Borough of Lambeth added to the register, and understand what works for users.
Currently there are only 10 assets listed on the register in Lambeth. Lambeth has more community assets than this.
So far, I’ve got as far as doing a trial with 1 pub, using Google Forms and promoting on Facebook.
More interesting would be using the using the Foursquare ‘check-in’ pattern for this sort of thing - building groups of people and civic campaigns around real world things.
Empathy, augmented - public services as digital assistants
sketches of a digital assistant for an imaginary public service
Google Now is probably the best known example of the so called ‘intelligent digital assistants’*. It suggests relevant information based on your location, your calendar and your emails. So, for example, it might automatically track a parcel based on a confirmation email from Amazon, or nudge you with the quickest route home based on your location.
Google Now is (for now) confined to day-to-day admin, and using it feels very obviously like having a machine help you out (and I’d guess a machine that runs off simple ‘given these events have happened, then do this thing’ rather than any artificial intelligence cleverness).
In addition to Google Now, there are examples of personal assistants that combine contextual notifications with a conversational, instant messenger style interface. So you get pushed some relevant information or asked to complete a task, but you can also ask a question or add a comment.
Native and Vida are apps that help you book complex travel arrangements and diagnose food allergies respectively. There is a good write-up here of how they work.
Compared to Google Now, these seem much less obviously like a machine talking**. Instead, you are having a conversation with a person (and it almost certainly is a real person most of the time), but there are these automatic nudges and nuggets of information that make the conversation richer.
What is really nice with these examples is that the differences of dealing with a real person and dealing with the purely digital parts of the service are abstracted away. There is a single interface onto a complex domain with computers doing the things computers are good at (joining together disparate data sets / reacting to changes in context in real-time), and humans doing the things humans are good at (empathy, and understanding complex edge cases).
So, what is the relevance for public services? Well, for most public services, probably very little. You don’t need an intelligent assistant to buy a fishing licence or book a driving test.
Where it is potentially revolutionary is in the delivery of complex services that require interaction over a long period of time and with many edge-cases - services where everybody is an edge-case and everything is always changing. Things like benefits, caring, health, special educational needs or mediation. Things that are complex and demand empathy.
What could a public-service-as-digital-assistant look like?
Smart to-do lists that make it clear exactly what the next steps a user needs to do to navigate the system. Very much like cards in Google Now, items/cards get added to the list based on a user’s context. Completing one task may trigger other tasks. For example, if asked to confirm how many children are in their household, and the number has changed since they were first asked, new cards might appear asking them to enter the details of the children. New cards can be added automatically by the system, at a face-to-face meeting with a government advisor, or when a user is on the phone to a call centre; there is one interface regardless of the channel.
A dynamic overview of a user’s situation right now. What this looks like will depend on the service, but should also change based on a user’s exact context. For example, the overview when an advisor is beginning to understand the caring needs of a family member may be very different once help has been put in place. Broadly though, these should communicate where a user is right now, how they are progressing through the system and what to expect next. The same view that is visible to a user should be visible to the government advisors who are helping them.
Augmented conversations that, rather than remove human interaction from a service, instead augment it. So if a nurse mentions details of a medicine a user is going to be asked to try, then the contraindications are automatically presented. Or if a special education advisor mentions a school, then the travel time and school performance are linked too. Or if a user notes down 5 jobs they have applied for, the pay ranges and locations are automatically summarised for the user and government advisor to comment on.
(The closest to this currently happening in the public sector is the work the Universal Credit Digital Service team are doing with to-do lists.)
Personally, I think these patterns provide an opportunity to design services that genuinely understand and react to a citizen’s needs, that seamlessly blend the online and the offline, the human and the automated into a single empathetic service.
I guess we’ll only find out by building some.
* I’m not counting Siri here, which is really more of a voice interface with a personality
** I’ve not used either of these directly, I’m just going on descriptions and screen grabs
This is the 3rd and final part of an essay about design and possibilities.
The first part - You can’t build what you can’t think of in the first place - was about the process of design being too linear, taking inspiration from evolution and the concept of hyper-volumes of ‘potential products’; the second part - Tools for exploring the margins - listed some approaches for thinking harder about the things that are possible in product design.
This final part is about power and about the obligations you now have if you make digital services in the 21st century.
Sampson diagram from The New Anatomy of Britain (1972) - largest bubbles are Civil Service, Industry, Parliament ans Conservatives
The image above is from The New Anatomy of Britain by writer and journalist Anthony Sampson. He wrote a series of books on the subject on political power in Britain, published approximately every 10 years from 1962. Each included a diagram of what he considered the current state of play. The one above is from 1971, the one below is from the final book in the series Who Runs This Place, published just before his death in 2004:
Sampson diagram from Who Runs This Place (2004) - largest bubbles are Media, Prime Minister and The Rich
I’ve always loved these diagrams (back at OpenTech in 2009 myself and Rob McKinnon used them to map civic tech projects).
Just like the biomorphs or ‘History of the World’ from part 1 of this essay, Sampson’s drawings are attempts to help us think about a subject that is inherently multi-dimensional. They are a tool for thinking about a problem - in this case how power is distributed and, taken together over the years, how it can change.
What might Sampson have drawn today, in 2015?
Well, politics is about the distribution of power in society, and in the early 21st century digital products are exerting influence on how power is distributed among us.
Redrawn 11 years later it seems clear to me a Sampson diagram would have large bubbles for the big digital services.
Politics in the 21st century will, in part, be about control over the digital services we now rely on, and which hold an ever growing concentration of our personal and household data, from how often we move (fitbit, jawbone), where to (Google Play Services), what we tell people (WhatsApp, Facebook) and to how often we burn our toast (Nest).
The same tight orbit that digital product design seems to be stuck in at the functional level (again, see part 1 of this essay) also exists at the organisational level: in the design of the organisations that run them.
The meme: ‘The only way to solve a given problem is to create a private company, provide a free service to users and mine their data’ is strong, but is also the equivalent in genetics of ‘the only animal that could possible exist is a hyena’.
And frankly, that’s getting a bit scary. As everything from household appliances to the most basic transport infrastructure gain an IP address and become fonts of data, at the same time as the democratic organisations of the last century seem unable to keep up, it is only going to get more so.
Software is politics now.
This was a subject that Vitalik Buterin, founder of the Etherium (a distributed, auditable computer) talked about at Nesta’s FutureFest event back in March.
There is a PDF of his slides here, but to try and summarise: the core utilities of the 19th and 20th centuries (roads, water transport, electricity system) were eventually run or regulated by governments, but the core utilities of the digital age (identity, communications, payment, sharing) are currently run by the first private company that happens to make its way to a near-monopoly. Etherium, a distributed auditable computer, is an alternative to unaccountable monopolies.
Just like water was in 19th Century London, where the adhoc organisations, with little accountability when things went wrong, were replaced with the first the Metropolitan Board of Works, and then a wider municipal democracy in the form of London County Council.
The story of the industrial revolution too often reads like that of entrepreneurs taking personal risk to weave the future against the odds. Now, granted that is a history, but not the interesting one in my opinion.
The interesting history is the one of the building of institutions that had the concept of accountability to the public baked into them - not an evolution of one thing to another, but active choice of a more accountable method of providing a service the public rely on.
Whether something like Etherium, which binds services to behave in a certain way via immutable code, is the right answer, or whether we need organisations that account for themselves in more traditional ways - membership, voting, but built for and of the digital age, are not the important things.
The first thing is recognising that the accountability mechanisms for a digital service are just another set of axes in product space - another thing that should be thought about and chosen.
Finding alternative models to run something like Uber, Google Now or Homekit that are viable is going to be hard (much as municipal democracy had a spluttering start and there were many failed attempts at finding a viable models for co-ops before the Rochdale Pioneers ended up with one that worked), but that’s no reason not to try.
The second, I think, is recognising the risk of designing services that are superficially the height of simplicity, but can never be understood. To steal a phase from Matt Jones’ brilliant talk on design fiction and the understanding of systems - “magic is a power relationship”.
If a user can never understand how something works, where is the opportunity for recourse? To pick an obvious example: what does it mean when you can’t view source on an ever more powerful Google Now.
To address this problem, I think the accountability model for a service needs to be an intrinsic part of the design of that service. Accountability needs to be embraced as part of the service design rather than abstracted away.
This creates some interesting design constraints: it means there is a delicate balance between designing something that people can use without having to understand how it works without totally obfuscating the underlying workings of the service.
The third is that the private sector does not have a monopoly on good digital product design, but equally more accountable digital products should not just be clones with a democratic overhead.
New technologies bring new possibilities for accountability. So design patterns like accountability at the point of use become relevant in a way they never would in a commercial context.
The final thing though, is recognising that if you build or design digital products in 2015 you have a new responsibility.
You are not just building the best, simplest, user experience, or the most elegant code. You need to be as vigilant against creating concentrations of power as you are in creating efficiency.
The image of power flowing from one part of a Sampson diagram should be ever-present in your head.
The reason? If you accept the argument that software is politics, you are by definition also designing a power structure, and that is an important responsibility.
Or to put it another way, sometimes the user need is ‘because democracy’.
Written between March and September 2015 - Brixton, Broadstairs and West Norwood.