Marginal changes and macro impacts

Fifteen years ago, I predicted the demise of the supermarket. Four years ago, I noted that that hadn’t turned out to be the greatest of predictions:

I wasn’t quite daft enough to think that everybody was going to do their food shopping online.  My argument was slightly more subtle, though it has so far proved no less wrong:  it was that enough of the cash-rich, time-poor customers from whom supermarkets make most of their profits would defect to make the job of serving the lower-margin customers insufficiently profitable to sustain the overheads of large sheds with even larger car parks wrapped round them.  As it turned out, Tesco roughly trebled both its turnover and its profits between 1999 and 2007, which demonstrates fairly clearly that they know more about running whelk stalls than I do.

Now Andrew Curry has written two interesting posts about the collapse of HMV and some of its wider implications, in which he points out that:

In passing, some of the lazier commentary on HMV points out that younger people don’t tend to buy hard-format entertainment content. But hard format sales of CDs, DVDs and Blu-ray still account for three-quarters of the market. And if you’re in a declining sector, the decline is going to be slower if your remaining market is an older demographic that has more money than younger consumers do.

Now clearly there are two kinds of internet disruption in the music industry and only one for groceries: the physical object can be bought from an online retailer (which, as Andrew notes, was the basis of the Channel Islands tax scam); or the object can be dispensed with altogether and and the product downloaded in pure electronic form. In the interest of scientific rigour in the writing of this post, I have just downloaded a CD (or rather, of course, not a CD, though I am not sure what other word to use instead) from Amazon. It took 91 seconds. That’s not a level of instant gratification which Ocado is ever going to be able to beat.

But even with that important distinction between sellers of inherently physical goods and sellers of potentially virtual goods, the basic gearing effect remains interesting. HMV, Andrew points out

still sells 38% of all hard-format music and 27% of all DVD and Blu-Ray discs. These are market shares that are conventionally regarded as being monopoly levels. So there’s something dysfunctional about an economy in which a retailer with this sort of share can’t make sustainable profits.

In effect, it seems that there is a level or rate of reduction of demand which retailers struggle to adapt to, even if on the face of it there is still plenty of business to be done. And while established retailers might struggle, as incumbent players generally do, to adapt to radically changing circumstances, there’s not much sign of upstart competitors filling the space either.

It’s hardly surprising that you don’t have to lose anything close to all your customers for your business to become unviable. What may be more surprising is how small the drop needs to be before the issue becomes critical.

Public services don’t tend to have quite those pressures and quite those metrics. The pursuit of channel shift is core to the government’s strategy, not a threat to be avoided. And yet. How many potential HMVs are there scattered around the public sector? And how many of those are managing to become John Lewises instead?



Design integrity

Designing complex systems is complicated. Sometimes there are interactions between the pieces which don’t quite join up, customer journeys which turn out to have structural obstacles. Some of that can be obscured or avoided through good interface design, but sometimes the underlying ugliness just pokes through.

This screen marks the successful end of a transaction. This screen irritates me immensely. This screen is an inadvertent illustration of last week’s post – swimming ducks and ungainly paddling cannot easily be separated.

Transport for London ticket machine showing "Your Oyster card has now been updated. Thankyou for using oyster. Next time why not top up online?"

There is nothing particularly wrong with this screen as a screen, but it makes an offer which it cannot fulfil. The transaction which the screen acknowledges and suggests would be done better online – the renewal of a bus pass – is in fact one which can’t be done online at all. So there are two pretty obvious questions. The important one is why it can’t be done online, the intriguing one is why it suggests it can.

The superficial answer to the first question is that bus passes can’t be bought online because buses are not equipped to load a season ticket on to an oyster card. There’s no terribly obvious reason why that should be so. There would be a data storage requirement to hold details of pending tickets, but a fairly minor one, and the functionality otherwise needed is already contained in the ticket machines on buses. But though it could work that way, it doesn’t. I have no way of knowing for sure, but I suspect that there is a gap in the system architecture somewhere, that the basic design wasn’t fully thought through, and that now it feels too hard to fix that gap, pretty minor though it ought to be.

So it can’t be done. Why then add insult to injury by patronisingly suggesting that it can? Pretty obviously because that’s a completely generic end of transaction screen, which lacks the basic logic to use the transaction which has actually been undertaken to drive what gets put on the screen. Maybe the people who do ticket machine design have no connection with the people who do web design. Maybe a train-focused organisation (and ticket machine) has forgotten that they run rather a lot of buses too.

Trivial though all this is, I think it it tells us something important. There are smarter ways than TfL so far seems to have found for glossing over a hole in the service. But no amount of gloss can get round the fact that the hole exists. In the end, the only way of fixing the hole is to fix it. Design is not a veneer – and just as important, veneer is not a design.


Reading and doing

The team have released a new version of their home page. It’s now quite a long way from the simplicity of where they started – but kudos to them for respecting the discipline of user focused design.

Most read on GOV.UK listBut looking at the page, one small thing struck me (well, one thing other than the other thing which struck me). There’s an interesting list of links under the heading of “Most read on”, and at the top of the list is “Job centre plus job match“. Fair enough and not surprising to find it there – job search has long been one of government’s most popular online services.

So it’s not the fact that it was there which gave me pause. It’s that it came under a heading about reading. At one level, of course, it’s trivially true that the primary action (still) for web pages is reading. Having found suitable job vacancies, you assimilate information about them by reading the words on the screen. Describing it that way, though, skates over the most important part of what has happened.'Directgov jobs and skills search - Job Homepage'

The jobsearch page is not fundamentally about reading, it is about doing. It rewards you for putting search terms in by providing search results back. It doesn’t reward you with anything much if you just lean back and read it. So while I am happy to believe that the job search page is the one which is most viewed, I don’t believe at all that it is the most read.

I also don’t think that’s how people think about it. “I read a job search” would be a much stranger thing to say than “I did a job search” (though admittedly “I searched for a job” is much more natural than either). Doing words are powerful: how we frame the doing matters.

That might be hair splitting of the highest order. But as I have argued before about, words have power. GDS is taking huge leaps forward and are doing so with grace and panache. Part of what they are doing is breaking further away from a reading-based paradigm of web design. That paradigm is a strong one though, and its tentacles will keep trying to pull them back.

So I am looking forward to seeing a “Most done on” list on their homepage. I hope it will be soon.

It’s bad to talk

I phoned a bank today. It was a call I didn’t need to make and which created no value for me. The bank may or may not care about that. It was a call which the bank didn’t need to receive and which cost them money they didn’t need to spend to deal with. The bank really ought to care about that.

The reason for making the call was that the process worked as it should have done from the bank’s point of view. The reason for making the call was that the bank didn’t seem to have realised that working for them wasn’t the same as working for me.

Start with needs* as the GDS design principles have it. The asterisk leads to a snarky footnote, *user needs not government needs. But it isn’t just government which needs to pay attention to that.

This is a story about opening an online bank account. The online bit was fine, but wasn’t enough actually to open the account, because banks legally and prudentially like to know who their customers are, and on the internet nobody knows you’re a dog. So they want to see some pieces of paper. The recommended approach is to take the relevant documents to a branch of the bank where they copy and return them, then send the copies on to their processing centre. Perfectly straightforward, so that’s what I did.

A week passed without hearing from them. Perhaps they are just a bit slow. Two weeks passed without hearing from them. Maybe something has gone wrong. Towards the end of the third week, a letter arrives: they are still waiting for the documents. That’s both irritating and worrying. Irritating to have to repeat the process, worrying that they seem to have lost documents – albeit copies – the whole point of which is that they are sufficient to support a claim to an identity. To say nothing of the fact that losing things is really not what you want a bank to be doing.

And so the phone call. After a little toing and froing I talked to somebody doing his best to be helpful and largely succeeding. There is, it turns out, no record of an account opening application under the reference number I have given, which seems more than a little odd since I have a letter in my hand with that very number on it. But one possible explanation it seems, is that the account has in fact been opened. Would I like him to check? It’ll take a moment, because that’s on a different system. Of course. And it turns out that indeed, the account was opened three days ago, or two days before I got the letter telling me that they were still waiting for the documents.

So from their point of view, everything has been fine: the right things happened in the right sequence through to a successful outcome – at least until I spoiled it by making a pointless phone call.

From my point of view, it’s not fine at all. I spent two weeks not knowing what was going on, followed by two days thinking that the process had failed altogether. A lot of that is down to the long period between their having the documents and their having the knowledge that they had the documents. The letter chasing me for the documents was dated twelve days after they had already been given them (and it took them a further six days to get that letter to me).

Amazon realised a long time ago that the most common question they had to deal with was, where’s my stuff? Overwhelmingly, they deal with that not by being able to tell you, but by making sure you never need to ask. Banks, it seems, still need to learn this lesson.

The more significant lesson though is the one I started with. Process efficiency is in the eye of the beholder. If as a service provider you don’t take the trouble to identify and address the needs of users, the best you can hope for it to meet your own needs as a provider.  And that best is not nearly good enough.

This is not really a story about opening an online bank account. This is a story about how service integration is still rooted in the base as much as in the superstructure. There is still much to do.

Game over

Some evenings on my way home from work, I play a small private game of chicken. Now I have played it for the last time.

If there are too many people at the bus stop – and no, I don’t know how many makes too many – that’s a sign that the gap between buses is longer than it should be. That means that when the bus does come, it will be more crowded than it should be. So far, so straightforward. But it also may well mean that there is a second bus not far behind the first. That’s reasonably likely partly because the second bus may be on time even if the first one isn’t, but partly because it is the iron law of buses that once one is delayed and fills up, the one behind will start to catch up.

So as the first bus arrives, I have a choice. I can get on the first bus with complete certainty, but less comfort, or I can wait for a possible second which may or may not be close behind and may or may not be relatively empty. Most people don’t make that choice: with rare exceptions, everybody else crowds on to the first bus. On good days, the second bus comes round the corner when the first bus has barely moved away from the stop. On bad days, there is a long wait for a bus which is eventually just as crowded as the first.

It’s a gamble, but I have learned to be canny with the odds, and mostly I win the game. Mostly.

The game only exists because of an information deficit. TfL knows where all its buses are, but the rest of us can only guess, except at the fairly small proportion of stops which have dot matrix indicators. The value of making that information more widely available has been self-evident for years. Just over two years ago, a team at Young Rewired State calling themselves TfHell won the Public Strategist award for “the service which most obviously ought to exist, for which there is no good reason that it does not exist and which now just needs to happen”.

But alas it did not just happen – until now.

I played my game for the last time this evening. Or rather today the game played me: the first bus was so full that it went straight past without stopping, so I didn’t get to choose. But within a couple of minutes of finding a seat on the second bus I found that the world had changed.

The effects should be interesting. Right now, it may not do much more than give me an edge in my decision making. If it were to get widely used, though, it should in theory produce an extraordinary feedback loop. If enough people wait for the second bus, knowing it is only just behind, rather than scrambling to get on the first, buses will get more regular and less crowded. A whole of host of individuals doing nothing other than making self-interested decisions based on a small but critical piece of information they don’t currently have, would improve the efficiency of the bus network.

The seminal work of observations of commuter behaviour, Notes from the Overground, is sadly very long out of print. Its author distinguished goats who seize and act on information to make their journey easier from sheep who do not. It’s a reasonable bet that even with better information, most people will still crowd on to the first bus (that is, after all, what happens now with tube trains, with the information available to everyone). That means that buses will continue to bunch up. But I will know that there is a second one, and that’s the one I will be on.

The web page says firmly that this is a test service – and indeed as a I write it is not working. Perhaps these last few hours have been all that we are to get for a while, and I will have to keep playing the game a while longer. In any case, there needs to be less to it than there is:  web pages designed for full screen browsers are a great way of consuming many kinds of content, but waiting times at bus stops is not one of them.

The first version of this has come much later than it should have done.  Let’s hope the second is not very far behind.

Update – 7 September

I have since discovered that there are three versions:

Desktop – with interactive maps

Text/Accessible – with static maps

Mobile – which is pure text with no maps.

Slightly confusingly, the desktop version has links to what it calls a text version, but is actually the sub-domain called ‘accessible’. The version which is actually text-based is not linked from either of the other two. Slightly more confusingly, although all three versions allow you to record a set of ‘My Stops’ for quick reference, the cookies which hold the data are held separately in the three sub-domains and so are invisible to each other (at least I think that is what’s going on; mine keep disappearing, even though I have given them an exemption from my standard cookie setting of eviscerate).

There is a page on the TfL site which explains a bit about the project – and it’s worth saying that this appears to be an unintentionally public beta at the moment, which deserves some latitude.

Independently of TfL, James Darling has produced a super-minimalist version which shows only times in dot-matrix effect red, and for hardier souls, Adrian Short has a command line version.

Picture by photosteve101, licensed under creative commons



On track to nowhere

There has been a lot of work in recent years on ways of improving the process of public consultation. It’s not something about which I have any great expertise or direct involvement, but I am conscious of great efforts to produce consultation material in forms which are not just useful and accessible themselves, but which make the process of responding straightforward and flexible, allowing respondents to express their views in ways which make sense to them. But it seems there is still quite a way to go.

There have been proposals to redevelop Battersea Power Station since Margaret Thatcher was prime minister. Every few years, a new developer announces an exciting new master plan with a Big Idea and says that development will start the following year. Then the money runs out, the site is sold, and the cycle repeats.

The Big Idea of the current lot is to get Battersea Power Station connected to the underground. That isn’t quite as daft and hubristic as it might sound. The Northern Line has a loop at Kennington which allows trains on the Charing Cross branch to head back north without reversing and TfL increasingly operates the Northern Line with all Charing Cross branch trains terminating at Kennington and only City Branch trains (which can’t use the loop) continuing to Morden. So adding a stub on to the loop to take the line on to Battersea doesn’t change the basic topology of the system and, at least by the standards of tube construction, it’s relatively cheap.

This is not the place to consider whether any of that is at all a good idea – London Reconnections is the authoritative site if that’s your bag. I am more interested here in how people find about and respond to the proposals, though admittedly with a less lofty purpose of being bemused about the priority of building a fourth tube station ten minutes walk from my house when I have a choice of three already.

If you were dimly aware of the proposal and wanted to find out more, what might you do? Well, the TfL home page would be an obvious place to start, and helpfully it has a big box at the top inviting us to see the Tube Update Plan. Clicking that gets us to line-specific options, and clicking that gets us to… no mention of this whatsoever. No matter, perhaps it’s not yet part of the plan. Looking round the site a bit, there is a huge list of projects and schemes, but none of them is this one. But if you are really persistent, you might eventually find a press notice announcing a public consultation on a proposed Northern Line extension.

Having found the equivalent of the locked filing cabinet in a disused lavatory, it all gets a bit odder. If you go to the press notice page now, you will find a link to the consultation site operated by the Battersea power station developer. But that link wasn’t there when the press notice was issued. The press notice talks about 40,000 leaflets being distributed to local residents and gives details of exhibitions of the plans at various places in the affected area, but says nothing at all, either in the body of the notice or in the notes to editors about the existence of a consultation website (and the version on the consultation site still doesn’t have so much as a link). That in turn meant that coverage of the story focused on the leaflets, with no mention of online information even in online coverage.

And so eventually to the consultation website. The first thing we might want to know is what is actually being proposed. And there is a link on the home page to the documents area:

See our Downloads section for a complete list of informative documents and maps about the Northern Line Extension.

Splendid. What might we find there?  Well, this:

Which of those might I possibly want to read? I have no idea. And this, remember, is a consultation aimed at residents, not at engineering professionals.

But there is also the consultation leaflet, the one being distributed to those 40,000 local residents. It is, perhaps not surprisingly by now, a pdf, which precisely reproduces the paper version, right down to the instructions to moisten the gummed area when returning a response. And while the response form consists of check boxes and a few short free-entry text boxes, the pdf hasn’t been set up to allow it to be completed with anything more high tech than a biro. There is, to be fair, a version of the questionnaire on the site, but only of the questionnaire. Would it really have been so hard to make six more A4 sides of pdf into proper web pages?

The central question in this consultation is

We propose to proceed with Route Option 2 (Kennington-Battersea Power Station via south Nine Elms). What do you think of this option?

This is apparently the result of an earlier round of consultation a year ago. There is no indication of what the other options are from which I am being invited to endorse a choice. But the leaflet does say:

Maps of all four route options as well as more detail on the current proposals are on our website:

After some poking around, I have found the four route options, not directly on the site, but as another pdf download (though not on the download page), this time of last year’s consultation paper. I can find nothing at all giving any more details of the current proposal.

I have laboured all of this probably more than the example is worth to make a point. None of this is impossibly difficult to find and use, but none of it is as straightforward as it could be and should be. If there is a serious desire to collect and respond to views, there are better ways of doing it than this.

And now, since I am one of the lucky 40,000, I will take my pen and the nice glossy printed version of the consultation and do my best to answer the really key question.

Nine Elms? Battersea?




Customer on a journey

Paul Clarke had been on a journey. Well two, actually.

The first was to get a bit of routine business done with government, updating the photograph on his driving licence. He has written a blow by blow account of how that didn’t work the way it should have done. Follow the link and read it now.

The second was to give not only that account, but a powerful message which should be heard and acted on by designers of public services – designers of any services – anywhere and everywhere. It is that it is the journey which matters, the whole package  which adds up to the complete experience of getting done what needs to be done. We need better web design. We need better form design. We need contact centres which can be human and effective. But we don’t need any of those things in isolation. They add up to a service, and it is the service we need to get right.

Cloud burst

As services improve, merge and become virtual, they disappear into the cloud. Which is fine when it’s fine, but sometimes the cloud bursts. We are, we hope, creating better ways of getting things done, but we are also unintentionally and perhaps unavoidably creating new ways for systems and services to fail. Those potential failures need as much thought as any other aspect of the service, but all too often don’t get it.

To make the problem trickier, some of those new points of failure may not be part of what is conventionally thought of as the service concerned at all. The service can fail even if the service provider performs faultlessly. One example of that comes from Emma Mulqueeny, who blogged earlier today about having to risk a parking ticket because she had forgotten the phone which was the only means of paying. Another comes from my last minute attempt to complete my tax return before the deadline today. I was uncharacteristically ahead of schedule this year, all ready to go as long ago as Saturday. But then I hit a snag. There was a piece of information I wanted to check. The easiest way would have been to log on quickly to the office payroll system. But it turned out that somebody had decided that this would be a good weekend to take the system down for maintenance. Suddenly the promise of access to my data vanishes just at the point when I might need it.

Those two examples suggest that there is a need for buffers and for leeway.

A buffer can take many forms. For my tax return, it is paper records which had appeared to be redundant. For Emma’s parking it could be a phone in the car park directly linked to the call centre. For my personal data, it is a first layer of backup which mitigates some risks and keeps everything easily accessible and a second layer of offsite backup which mitigates much more catastrophic risks but with less immediately convenient access to the data.

Leeway can also be found in many forms.  It’s a term I first came across not far short of ten years ago in David Weinberger’s proto-blog, and I still think it’s enormously important and relevant. Human operated systems always have slack in them. That slack can very easily be seen as waste, but it isn’t, it’s lubrication (except, of course, some of it is waste, so it’s essential to be able to tell the difference). Computer operated systems usually have the slack designed out of them, partly because that’s the easier way of developing them and partly because of the perceived benefits of consistency and conformity. The result is a rigid, and therefore brittle, system.

The London congestion charge is a great example of a system which started without leeway and gradually has had it added, with the result that the pain of compliance has been dramatically reduced. Originally, the system was completely unforgiving of any lapse of memory. Failure to pay on the required day triggered a penalty charge.  Some time later that was changed to allow payment to be made on the following day as well, at a slightly higher cost. From this month, it has been possible to sign up  – again at a higher price – and have payments triggered by the act of driving into the charging zone.

The need for buffers and leeway will not go away. Sometimes – perhaps often – the best place for them is outside what is perceived as and designed as the core system. Service designers need to reflect on that, and to do so in the context of the system experienced by the user, not the more narrowly conceived system offered by the provider. As Emma observed,

Whilst it is true that savings can be made and that consumers are becoming used to expecting there to be a digital option for pretty much everything – it is a mistake to cut out humanity completely. It is the kind of counter-productive behaviour that makes people very cross and frustrated, normally in times of deep stress or just general state of worry such as we find ourselves in today.

How soon is it right to ride the trend?

Truncating the axes is the oldest trick in the book, so the story this chart is telling is not quite as dramatic as the initial visual impression, but that story is still striking and important.

The proportion of internet usage from mobile devices is tiny, less than 3%.  That’s almost certainly an understatement, since the chart measures operating systems, not connection types, so includes mobile devices rather than devices which happen to be mobile, but the absolute numbers are still pretty much insignificant.

The trend is, quite obviously, another matter. The share of internet usage from mobile operating systems has gone up thirteen fold in the last two years, and the slope of the line is robustly upwards. For all we know, the line might hit some natural ceiling in the next few weeks and never break the 3% barrier. That seems remarkably unlikely, though: if absolutely nothing else were happening, the simple pattern of device renewal which characterises phones as opposed to computers, bakes in substantial future growth to come – and in any case there is no reason to suppose that nothing else is happening.

I thought the chart – which I have lifted from a wider analysis of operating system usage data by Ed Bott – was striking enough to be worth a post in its own right. But then it made me think of an exchange on twitter earlier today:

@Directgov: The Guardian’s Consumer App of the Week features the #jobcentreplus #app from @directgov #iPhone #iPodTouch #iPad

@Marthalanefox: @Directgov how many ppl going into a jcp have an iphone?

@Pubstrat: @Marthalanefox @Directgov Some certainly do – and the app is also available on android. It’s one more way in to accessing the info online.

@DavidCotterill: @pubstrat @marthalanefox @directgov a few months back the figures were 80k job searches a month from iPhone. Small, but growing…

We seem to be back with the question of whether government should be dabbling with minority interest technology at all. I wrote quite a lot about that in two posts last August – Apps for Elephants and More on Apps for Elephants – and I am not going to go over that ground again. But I think this graph and the challenge implied by Martha Lane Fox’s tweet encapsulate one aspect of that debate very neatly. In essence, the question is whether we should pay more attention to the fact that the absolute figure is still very low, or to the fact that it is growing so rapidly?

There is no inherently right answer to that, no simple rule which automatically determines the right answer. I think government should be very cautious about spending time and money developing at the purely experimental end of the scale. But that’s not where the mobile internet is any more, and it seems pretty clear that we have got to the stage where we should pay serious attention to its rate of growth.

And in the meantime, if you are on the move and looking for a job, you can get the Jobcentre Plus app for iphone/pod/pad or in the android market.