A while ago, I came across an interesting little gadget called a Poken. Pokens are a way to mash up real and virtual social interactions: a theme we’ve have been exploring as part of research projects such as EventConnect and peg.me, and one branded by Tim O’Reilly last year as “Web 2.0 plus World”. Pokens are small, keyring sized devices which contain an RF chip which supports very short-range wireless interactions when devices come into contact. This offline “handshaking” is backed with an online service which links up your social network with that of the people you “poked” in the real world. The necessary data is transferred when You dock your Poken with your computer via its in-built USB port.
At first, I was quite sceptical about the Poken. In the end, this kind of functionality will probably end up on your mobile once NFC chipsets start appearing in handsets as standard. Why have a special device for “poking” people? But I’m starting to think my scepsis was misplaced. The technical fundamentals behind many Web 2.0 applications including Flickr, Twitter, digg, or even the social networking platforms themselves were never particularly ground-breaking and yet they have proved incredibly popular. Uptake is more about ease of access than chipsets or APIs (although these turn out to be important to longevity and user retention). So I think that the Poken guys are onto something. They also have a key advantage over other Web 2.0 players in that they are selling something tangible and, in so doing, may actually have a workable business model from the get-go (providing they stay focussed on where their revenue is supposed to come from). They still have the usual issues around reaching the tipping point, but I think they will achieve this and also go on to succeed in the longer term.
An article entitled “Googling the future“ appeared in this week’s economist discussing the usage of aggregated search data to support macroeconmic forecasting. Citing Google chief economist Hal Vairan, an 18% improvement in the accuracy of motor vehicle sales forecasting was achieved by taking into account the second derivative of searches containing semantically relevant terms (for example, vehicle brand names). The principle is not new. Bill Tancer has discussed similar approaches at length in his book Click which came out earlier this year. According to the theorists, analyses of search activity, raw traffic and clickstream data can help predict real-world outcomes. The approach can be applied not just to economic indicators but a host of other fields such as epidemiology and pyschology.
Behind this application of web data analytics are a couple of research and statistical principles. The use of web data offers statisticians some distinct advantages over traditional methods. First, the percieved anonymity of user interactions on the web can help expose true intentions, unbiased by the social effects at play in taking part in a questionnaire or other traditional research tool (a problem known in the trade as Cognitive Dissonance). Second, massive amounts of data are being generated constantly, in a near real-time, ever changing stream. This means that significant trends in behaviour show up more or less in immediately - if you can only find the right way to frame your question (of course, this is trickier than it sounds). But if, on average, you can predict outcomes more correctly and faster than everyone else, you have an opportunity to exploit your superior insights in a market - something Tancer calls data arbitrage.
These techniques are already being used where arbitrage can, of itself, generate a direct return. Tancer is chief analyst at Hitwise, a competitive research company specialised in providing exactly these kind of insights on a consultative basis. Unsurprisingly, financial firms feature prominently on their customer roster. But the techniques are also relevant to marketers looking for insights which can generate a competitve edge. Razorfish’s competitive intelligence offering operates under the moniker “Global Solutions” and advises clients on exactly this kind of issue.
The competitive intelligence story is compelling in part because of its potential at both at the micro- and macroeconomic scale. But it is not the only reason for marketers to engage in close observation of consumer behaviour beyond the borders of their own properties. Our French sister agency, Duke, recently came up with a campaign for the French railway company SNCF (executed on April 1st this year). The idea was triggered by popular demand and here is what happened.
A week after getting the G1, I had to give it up. And then I went skiing for the weekend in Stubai with the agency, so didn’t get the chance until now to add some more about the first Android handset on sale in Germany.
In the meantime, I have seen some more Android-related news: that there is a second Android phone on its way, and that you can now get Exchange push mail to your Android using third party software downloadable to your phone via the market.
The push mail feature is good in principle although I personally find push-mail overated. Mostly because all that immediate but mostly valueless email processing ends up killing your battery and preventing you from doing something immediately valuable: namely taking and making calls. In practice, I tend to suppress warnings for incoming mail to prevent my phone beeping and buzzing non-stop, meaning that I only actually review email on the go periodically - at which point it might as well be fetched on demand rather than instantly (particularly if it means an extra couple of days battery life). Nonetheless, the mail issue is that of getting at your office mails simply and quickly, and these days that means using ActiveSync rather than IMAP. I tried the Touchdown app on the Android but, well, it’s rubbish. Moxier’s new application may be better, but I don’t know because the phone is gone now and the Emtrace Beta lacks many essential features like address book lookup. Google’s own licensing of MSFT’s ActiveSync patents appears to be more about allowing customers to sync Gmail across other devices (in particular, the iPhone) than about allowing Android users to sync direct with their companies’ exchange servers like on other handsets (in particular, the iPhone). In the meantime, Google is confining their “push” efforts to promoting their own mail services. So right now, email support is not what it should be on the Android, although this is likely to improve over time.
I also had a look at the IDE for Android application development. It took a while to get my Eclipse back up to date (it’s been a while) but once I’d got the basics sorted, I was able to get the Android SDK installed and the eclipse plug-in working pretty quickly. I had a play around but didn’t have enough time to build an app, though it became pretty clear that the barriers to entry for Android application developers are considerably lower than on the iPhone. It makes a big difference at the start being able to use a familiar IDE, code in a familiar language and use familiar API’s. Chrizz from work spent some time last year learning how to develop on the iPhone stack (his demo “Weather Map” app is available in the appstore to download) and although he has come to like them the likes of XCode, ObjectiveC and Cocoa seem initially obtuse.
So the overall impression after a week of using the G1 is that Android has the potential to be huge. The right decisions have been made to seed a genuine paradigm shift in the way mobile technology has been brought to market over the last 10 years. Until the iPhone, those last 10 years have often been about unfulfilled potential. The iPhone needs competition though and I think that Android platform is sufficiently different to stand a chance of providing it.
One of the social media guys at work, Johannes Kleske, just dropped a T-Mobile G1 off at my desk (otherwise known as the first Android phone). I have to give it back next Thursday, but thought I would try it out for a week. I’ve mentioned Android elsewhere but this is the first Android-enabled phone I’ve actually played with.
The first thing you notice is that it is not nearly as ugly as many have commented on. It’s definitely not as beautiful as an iPhone but then again neither is anything else, and many others have tried a lot harder and failed completely. The G1 is certainly quite chunky, but there is something solidly satisfying about its build, something I would love to say about my HTC Touch Diamond but can’t, because it sucks. As, by the way, does the crass attempt to pimp Windows Mobile with the ridiculous “TouchFlo” interface which is the equivalent of having to endure a mildly interactive flash-intro every time you want to make a phone call. Although this isn’t an iPhone killer you get the feeling it isn’t meant to be, at least not yet.
Like the hardware of the G1, the overall UX feels business-like, clean, fast and actually quite appealing. The screen is big, bright and contrasty and, having battled with “soft” keyboards for a few months, the fold-out “hard” version actually feels like a relief. Somewhat of a surprise is the little trackball for navigating around the menus which is zippy and, well, works. The screen-based UI is not multitouch-enabled but at least feels like it was designed with fingers in mind and doesn’t assume you want to waste your time fiddling with a plastic pointer which you know you’re going to lose any second. Pretty quickly, you end up using the hardware navigation aids, and they are really rather good. The menuing and navigation also feel like the the clearest and simplest I’ve used since the Nokia I had in London about 10 years ago (we’ll see if that impression lasts). Beyond the simplicity though there are also some nice touches, the transitional animations between apps are smooth, the semi-transparent canvasses are pretty and, shockingly, the back-button does what it should and works accross apps and window states. This is such a simple thing that you wonder why it strikes you, and then you realise that it’s because it actually exceeds your expectations - and that, in a nutshell is the G1.
I had a quick go with the “Market” application (the equivalent of Apple’s appstore), downloaded the barcode-reading app, half-expected the kind of mobile gag you are bombarded with in TV ad-breaks but no, it actually worked. Having grabbed and scanned one of Elke’s books from her desk opposite me (”Erfolg im Web, worauf es ankommt”) I could actually have bought another copy then and there if I’d wanted to. Admittedly though, this didn’t work with the bottle of Becks at the agency Friday-evening gathering.
So my overall first Android impression is really very positive because it feels like people have spent time getting things to actually work well. That might sound underwhelming, but in a world dominated by the iPhone, it’s the price of entry and one that noone else seems willing to pay. And hey, it’s only the G1: there is still some work to do to compete head-to-head with Apple, but the Android paradigm has some fundamental advantages over the iPhone’s which may, over time, slowly erode the gap. I just don’t see anyone else even on that trajectory right now.
I would buy this phone, and in fact, I think I will (let’s see if I’m still saying that on Thursday next week).
On the one hand a further tactic to push Silverlight plugin and on the other something quite exciting. Photosynth is one of the most interesting technologies to have emerged from MSFT labs. The zoomy bit (Seadragon) was incorporated into Silverlight a while back, but until now, the stichy bit hasn’t been. Hopefully, the 3D dotty bit will come allong soon.
Soon after that, a Silverlight client will undoubtedly integrate the whole lot with Virtual Earth and mash in huge volumes of public imagery (which should keep a few large datacentres warm for while). I hereby predict that this will happen by the end of the year, and I’m looking forward to seeing it when it arrives.
Scenario: System has populated user’s profile using explicit or implicit information gathering methods. User is requesting a page which contains a component you want to populate with content relevant to their profile. You want to highlight the most relevant content to the user.
Solution:
Personalisation can be used to help the user find what they want quickly and easily. Products can be selected from a wide product range. Relevant news can be filtered from a vast content stream.
Examples:
Forces
Data: Define the user profile data attributes required, define content profile data attributes required
Options: Define possible attribute values (preferred values)
Presentation: Define possible template outputs
Rules: Define logic/rules linking profile data attribute values to content metadata attribute values
Limitations:
The content selector assumes a fixed template. You need to nest content selectors inside templates selectors or vice-versa if you want to mix the two patterns. You may have situations where your rules contains some constants in them, for example in the form of “if-then-else” or “case” blocks. The most powerful rules do away with this kind of wiring, and allow values on each side to be compared directly (albeit with some decoration or using a general mathematical operator). For example, “get me the latest offers who’s minimum target age definition is less than the age of the current user”.
I just spent the week at Wincor World, the annual trade show of Wincor Nixdorf. This is the now the fourth conference where I have demonstrated applications built by Razorfish on Microsoft’s Surface platform. The reaction was almost universally positive, from board-level representatives of Wincor’s impressive customer portfolio across to the children of some of Wincor’s employees who looked around the show at the end of the final day.
Here is a video of Otto playing the piano on Surface at seven months. He really is very advanced for his age
Swap contacts with a business partner at a trade fair Send a friend the details of a cool shoe you just found in your favourite high-street store Add details of a local fashion show event to my calendar and invite friends. Maybe details of the event are displayed on flyers and/or the back of my till receipt Show and share your photos stored on flickr
Detail
For the sake of the demo, we created some Surface “Identity Tags” and printed them onto our own business cards/ conference badges. We created a facebook event and invited FB contacts to RSVP. We also created an EventConnect application allowing FB users to grant the necessary permissions and track contacts created at the event in real life. At the event itself people can do the following:
Swap contacts
A group of people can gather around the surface and each put their Surface Connect card face down on the table. Instantly, each person sees their contacts from their FB networks appear. They can then start trading contacts just by passing the virtual cards around on the surface. Every time this happens, the relevant introductions are being made behind the scenes on facebook. Say I pass you the card from my deck with by brother Tim’s details on. Tim would then receive a message saying that I would like to introduce him to you. My offline interaction with you created an online interation between you and Tim.
Swap photos
You can also see any photos you have uploaded to FB (in the EventConnect album created when you add the app to your profile). You can share the photos around with other people who are logged in, they then get the photos in their albums on Facebook.
Additional use cases…
There is an “event card” permanently on the table which shows all those who have signed up for the Facebook event. As a user you can therefore initiate a Facebook connection with anyone else who is at the event. You can also save “official” event photography published to FB by the event organiser to your profile. Finally, you can invite your own buddies to the event.
I was thinking today about the 700bn promised by the US government to help avert further meltdown of the financial sector. It reminded me of the carnage in the web world following the dot-com crash. According to wikipedia, the fall-out of these events wiped out 5 trillion of market value. Nobody knows how much the credit crunch will end up costing. In march this year Goldman guessed at 1.2 trillion. But then Goldman cashed in their chips today in order to take advantage of the US government’s generosity as lender of last resort, so they’ve probably since decided that they underestimated a little. In the DotCom meltdown there was no lender of last resort, you went bust, wiping out any money you had invested in your firm and all your shareholder’s money with it. But of course, this isn’t just about the initial losses. Killing the financial system has the unpleasant side-effect of making it difficult for other parts of the economy to carry on functioning. After the DotComs went bust, money could be diverted into safer havens like, say, into property…
But this post isn’t directly about the banks or about real-estate, it’s really about another form of negative equity. Bailing out the banks has another unpleasent side-effect, namely that of ploughing tax-payer’s money into city bonuses. The government rescue plan ends up looking like some kind of regressive taxation or reverse robin-hood income re-distribution. It’s really an extension of what George Bush started with his tax-cutting programme a few years ago. The average annual income of the top 1% of earners in the USA recently passed the 1m mark. Among those 1%, there are undoubtedly many who will welcome the intervention of the US Treasury, not only on the grounds of its being the right decision for the country’s economy, but also because that 700bn is effectively heading their way (ok, I’m exaggerating a bit, but it isn’t that far off). This top 1% also account for roundabout 20% of the overall household income (at least this was the score a couple of years ago). I read some more detailed data on Friday which was forwarded to me by a colleague (he forwarded it because it showed that digital media consumption in the top 20% income bracket now surpasses traditional channels like TV). But the point which is relevant to this rant is that according to the data, there is segment in the US which consists of peope with average earnings over 850k a year and liquid assets were in excess of 3.5m. And this segment contains around about 1m people. In other words, this lot are worth 850bn a year, at least.
Now, if I was an american, I would be thinking, “Hey, I’m a fan of free enterprise and I hate regulation and beurocracy as much as the next man, so guys, lets just do us a deal here. We’ll use our collective spending power of more than 8 trillion a year to bail out the system in the short term but you’ve made the best of this system in the past, so really, you should be paying over the odds for keeping it alive. You pay us back at 10% a year for the next 10 ten years and we’ll call it quits”. In actual fact, all they need to do is reverse the cuts that George handed out to 1m+ earners a few years ago and they’re done. Actually, this has quite a nice symmetry to it.
We have been doing a lot of work in the last 6 months to improve QS processes in the agency. One of the things we started on last year was using the Selenium test framework to automate functional and regression testing. We use it to complement automated unit test suites by batching end-to-end tests via a web browser.
The thing is that even with browser-driven tests, you can’t check the accuracy of the visual layouts as represented in the full range of OS/Browser combinations commonly used. We have been working on an app which allows you to do just that. The app was developed by one of ND’s lead developer’s Martin Preis and it can help at every stage of the development lifecyle: from template build (using it’s own lightweight templating engine) through to live regression testing (via integration with Selenium). Here is a white paper explaining how it works.
Entries (RSS)