Quantity or Quality? Measuring Enterprise 2.0

[Crossposted from the Headshift blog]

One of the most common barriers to adoption of social software in enterprise settings is the perception that using social media isn’t “real work”. Instead the term “social” in “social media” is seen as a synonym for frivolous and time-wasting.

To an extent, there are aspects of social media that reinforce this stereotype - Facebook is often cited as the poster-child for this problem. Give your staff access to Facebook, runs the argument, and they’ll spend all their time poking and throwing sheep at each other.

This is actually a debate that’s as old as management itself - I’ve been around long enough to remember the same debates around rolling out email, and I’ve no doubt that phones were seen as a terrible distraction and waste of employee productivity.

At the heart of the issue, it’s a generic management problem and comes down to whether you subscribe to the Theory X or Theory Y view of your staff.

But when applied to Enterprise 2.0, this issue is accentuated by the intangible nature of a lot of social media. How do you apply a value to contributing to a wiki or a forum?

A recent post by the Harvard Business School’s Andrew McAfee looks at this problem, and he’s come up with an intriguing take. His point is that this kind of activity is too broad to reduce to a single metric - and that attempting to do so can cause unexpected side-effects. Measure contributions by volume, and it becomes easy to game the system with quantity overriding quality.

And there are other factors in play, too. Analysis that I’ve seen done of the commenting patterns on a large UK network of blogs show a pronounced long-tail effect in action - active commenters are VERY active, leading to a pronounced power law curve when comparing activity across the population as a whole.

Professor McAfee proposes a multi-dimensional rating, that combines a number of activities. Authoring blog posts, editing wiki pages and contributing to discussions in forums would all build towards an overall rating - and measuring a number of activities allows some rather neat visualisation techniques.

Although this is still fundamentally a volumetric approach, it should also be possible to add in a rating factor - rather like eBay feedback, there are a number of techniques for rating contributions for quality.

I’m sure this model would work to an extent, although it would also likely suffer from eBay’s flaws - negative feedback is disproportionately weighted compared to positive. That’s particularly the case if you’re the proud owner of a flawless feedback record - a single negative rating hurts that in a very visual way. And eBay have in fact moved away from a straight-forward “buyer-rates-seller-rates-buyer” model to something which is slanted more in the buyer’s favour.

And there’s also the issue about how likely you are to negatively rate the contributions of people that you work closely with - frank and honest feedback is a hallmark of some cultures, but I do wonder if there would be a lot of “pulled blows”?

Perhaps the question is whether measuring the effectiveness of Enterprise 2.0 tools at the individual level is actually the right place at all? On the one hand, social tools are predicated on the idea that the total is greater than the sum of the parts, as network effects come into play. So perhaps a truer measure of effectiveness would actually be found by looking at a more aggregated level?

Social contracts

Spotted on Hugh McLeods’s blog - “how does a software company make money, if all software is free?” - an observation about the difference between closed-source aka Microsoft and open-source:

It took me a while to figure this out, but what applies to Open Source, also applies to Microsoft.

When you buy a Microsoft product, you’re not just getting ones and zeros. There’s also a form of “social contract” implicit in the commercial transaction. You gave them money, this entitles you to certain expectations.

A few weeks ago, I met a young developer who worked in an IT department of a large insurance company. I asked him what kind of software did he use. Answer: About 75% Microsoft, 25% Open Source. I asked him why did he not use more Open Source? I thought IT people loved Open Source?

“If something goes wrong with Microsoft, I can phone Microsoft up and have it fixed. With Open Source, I have to rely on the community.”

And the community, as much as we may love it, is unpredictable. It might care about your problem and want to fix it, then again, it may not. Anyone who has ever witnessed something online go “viral”, good or bad, will know what I’m talking about.

Which is only true for a given subsection of the Microsoft user base. If you don’t have access to that level of support - and most organisations below a certain size don’t - then you’re thrown back on the exactly same type of community resources regardless of whether you’re using open or closed source. The difference being that the open-source model provides the visibility of the source code, and the potential for fixed that this presents.

Mobile green screen

In case there was any doubt why the iPhone gets all the attention, compare and contrast these screenshots from Twitter clients for iPhone and Windows Mobile Live Personal Client 2007 (or whatever it’s called these days):

First, PocketTweets (for iPhone)

Then iTweet (also for iPhone)

And finally, Twmobile. Guess which platform?

After six months of fighting with my t-Mobile MDA Vario II, I dumped it in favour of an iPhone, flogged the Vario on ebay and haven’t looked back.

Tragedy of the wifi commons

One of the gimmicks marketing ploys that National Express East Coast employed when they took over the East Coast rail service from GNER was to extend the onboard wifi service that GNER had introduced. Prior to the NXEC takeover, wifi onboard was free to first-class passengers, and a minimum of £4.95 an hour for those travelling in cattle standard class.

In the GNER days, the service wasn’t bad - the connection speed was generally slightly faster than you’d get using a 3G connection, and reasonably reliable. The technology behind the service was innovative, too, using a combination of GSM and 3G links to maintain a pretty robust connection in fairly hostile circumstances.

The downside if you were travelling in standard class was that you had to pay out for the service. When someone else was going to pick up the tab, I generally paid up - after all, £4.95 isn’t *that* much if you’re going to get something productive done as a result. In first class it was free, which meant that my Sunday night trips were usually productive - or at least less boring - ones.

Then along came NXEC, and threw open the service free for all. In the process, they took something which was pretty useful, and turned it into something which is actually worse than useless. Despite the increased usage, they haven’t increased either the backhaul bandwidth or the onboard infrastructure - which means that most of the time the throughput has dropped to single-figures of kbps, and ping times have increased to the point where the average is generally several seconds or more. And that’s if you can actually get onto the network in the first place - the DHCP scopes are often exhausted, and the authentication gateway regularly buckles under the strain and drops your connection, meaning you have to re-authenticate.

All of which means that travellers would actually be better off without the service at all, because of the time that gets wasted trying to get the connection working.

Technorati Tags: nationalexpress, nxec, trains

Porcine cosmetics, or something new?

When does “cool new features” transmute into “unwelcome feature creep”? Two of the online services that I use the most are the subject of rumours about new functionality - Flickr is apparently about to launch video support, and Skitch is to add Twitter features.

Personally, I’m pretty ambivalent about Flickr and video. I’ve never really suffered from the urge to take moving pictures, probably as a result of being subjected to too many dire and wobbly home videos of holidays over the years. And it’s not like there’s any shortage of online video sites in the first place - why build another YouTube?

But Skitch and Twitter, on the other hand, does seem to add value. I already use it at work as a quick way of sharing image edits, or annotations to a screen shot - the fact that the end results sit outside the firewall and are readily-linkable make it really flexible.

Now add Twitter to that mix, and the enhancement (I think) is actually to Twitter rather than Skitch. Twitter is beginning to take on the shape of something much more than just a cross between IM and a stream of consciousness - I can’t quite put my finger on what it is now, but it’s clear that there’s something beyond simply novelty value which has kept it popular. The subject of another post, when I get around to marshaling my thoughts.

Installing and fettling Subversion

I’ve finally got around to figuring out how to install and configure a Subversion repository on one of my Debian boxes - these notes are more for my benefit than anything else, but through the power of Google they may come in handy for someone else.

Installing Subversion with HTTPS support from scratch:

Installing Subversion on Debian Etch Complete

Creating the initial repository

svnadmin create –fs-type fsfs /var/svn/myproject

Doing the initial import

svn import -m “New import” myproj https://mydomain.com/myproj

This lifts the contents of the myproj directory and imports it into the repository

Working around HTTPS certificate problems in a Subversion client

If you’re using a GUI client to access a svn repository through an https connection, this can cause the client to barf - if it’s an invalid certificate, you may not get the option to trust it.

The workaround (on OS X, at least) is to cache the certificate from the command line so that the gui client will then accept it.

If you fire off

svn list https://host.example.com/repos/project

at the command line, you’ll get the option to reject or accept the certificate. Permanently accept it, and this will be cached and used by the gui client next time you access the repository.

Technorati Tags: debian, linux, subversion

Faster, deeper, sleepier

My first Mac was a G4 Powerbook, and one of the principal “wow” factors was the sleep function - close the lid or hit the power button, and the machine would sleep more-or-less instantaneously. Waking up was just as fast, which was always good for impressing the Windows fanbois as they went off for a coffee or three while their Windows machines restarted.

When I upgraded to a Macbook, I found that the sleep function took a lot longer - not so much longer that it wasn’t still impressive to Windows users, but long enough to be irritating to someone used to a Powerbook. That’s because the Macbooks sleep AND hibernate, which takes a while as the contents of memory are written out to disk.

SmartSleep is a free-as-in-beer utility that installs as a preference pane in System Preferences, and allows you to change the default behaviour of the sleep function so that it just sleeps. This is dramatically quicker - but the clever part is that the utility will switch to sleep and hibernate if the battery level drops below a definable threshold.

All of which will save me, oh, minutes per day….

After two days with an iPhone

When the iPhone initially went on sale, I was fairly ambivalent them and didn’t immediately rush out and queue outside the Apple store at midnight. A couple of things made me change my mind - firstly O2 dropped their tariffs to something approaching reasonable value for money (or at least parity with my existing TMobile contract); and I had a good play with iPhones belonging to Aral and Pedro while I was at LIFT08.

My original intention was to jailbreak the phone and run it on my existing TMobile contract, but there were two reasons why I didn’t go down that route - firstly, jailbreaking the 1.1.3 firmware was too complicated and geeky even for me (although the latest version of ZiPhone changes all that); and secondly TMobile doesn’t have an EDGE service and the iPhone doesn’t (yet) have 3G. I didn’t fancy being stuck back at GPRS speeds after getting used to 3G.

So, now that I have a shiney new device up and running on O2, time for some initial thoughts. These are thoughts from the point of view of two points of comparison - a Nokia N73 on the 3 network that I use for (or was at least given by) work, and a TMobile MDA Vario II which is mine, and is a rebadged HTC Hermes device.

The first thing to point out is that I hate the Vario with a passion that I’ve never felt about an inanimate object before. It was an ill-considered purchase made in a hurry, and it sums up everything that’s gone wrong in trying to shoehorn the Windows OS onto a mobile platform. It’s just far, far too difficult to use - the menu-driven paradigm of Windows has been “adapted” to a touch interface, so it’s virtually impossible to drive one-handed. It’s slow - you rapidly get used to watching the spinning pizza of death as it tries to catch up with the human using it. And it’s really, really flakey - little things like the phone dialler, for instance. As you dial a number, it’s scanning the contacts database at the same time to bring up the details of the person you’re dialling. That would be quite a neat trick, were it not for the fact that with 1,000+ contacts the whole process is so slow that there’s a perceptible pause between hitting the key and it registering on the screen. It sounds trivial, but is incredibly annoying - if the phone isn’t very good at basic things LIKE MAKING A PHONE CALL, what use is it as a phone?

I just don’t care about the Nokia. I find it tricky to use - the interface is unintuitive (to me at least), and after using a Blackberry for several years, I found the numeric keypad and T9 extremely difficult to work with. As a phone, it’s rammed with features - but I don’t play games on it, the music player doesn’t come close to replacing an iPod and the web browser is let down by the fact that the screen is 2.5cm wide. All in all, it’s just “meh”.

So what about the iPhone?

Physically, it’s on the large side - as Aral pointed out while I was playing with his, it’s small enough to fit into a pocket, but it’s large enough that you have to take it out and put it on the table when you sit down - so everyone can see it and go “ahh” and “ooh”. But that’s balanced by the fact that there are no external protruberances or buttons to catch on anything - it’s the same smooth form factor that you’d expect from an iPod or a Mac.

The interface is also what you’d expect - extremely simple, stripped down and entirely touch-driven. It’s covered in affordances - clues about how the software operates and what’s expected of the user. For example,

Initially I was put off the touchscreen keyboard after playing with an iPod Touch - the keys seemed extremely small, and the lack of textural feedback meant that there was no way of telling short of looking whether your fingertip had actually “pressed” a key. But to my surprise, I’ve now got to the stage where I can type as quickly as I could on the physical keyboard of the Vario. The predictive text seems to work seamlessly, and without getting in the way - on the Vario, using the suggested word means breaking your typing flow and hitting the return key, whereas the iPhone assumes that you want the correct and inserts it as soon as a space or punctuation character is entered. Ignoring the suggested word is slightly counter-intuitive, because you have to touch the suggestion bubble - but quickly becomes second nature. While you may not want to write novels, banging out emails or quick Twitter updates is very straightforward.

Overall, I think this is a device that has dramatically raised the bar between what went before and what’s to come in the future. There’s simply no other device out there that can compete from a user interface point-of-view. Comparing the Nokia to the iPhone is like comparing a green-screen terminal to a GUI, and it simply makes Windows Mobile devices look ugly, slow, old-fashioned and buggy - they’ve been designed with an “it’ll do” attitude that is the polar opposite of the Apple “don’t release until it’s perfect” philosophy.

The interesting question is what the iPhone is going to do for mobile devices in the longer term - is it the shape of things to come, or is it just an early-stage disruption that will spur development off into other areas?

Technorati Tags: geek, hacking, iphone, gadgets, puppy

Off to Geneva

I’m off to Geneva for LIFT08. See you there…

Technorati Tags: blogging, geek

On Microsoft + Yahoo:

Microsoft’s web technologies are as unrelated to Yahoo!’s as the Dark and Light sides of the Force. Before you even begin your Comp. Sci. degree you’ve already made a decision to join one camp or the other. There’s no love lost between the two sides, and very few developers jump from one camp to the other mid-career.

Perl, PHP and Ruby developers carry PowerBooks with startup stickers on them, ride a bicycle and wear a tee and jeans. They are too skinny. They are more likely to have an iPod earbud in their ear than a phone. Microsoft developers wear chinos and a business shirt or collared tee, carry a black generic laptop identical to their coworkers. They are a little overweight, but only because they have a good wife at home who loves to cook. They have a full schedule of meetings and tasks always with them in their Exchange-connected phone, which they carry in a leather holster on their belt, with a blinking-blue Bluetooth headset always jammed in their ear. They think the Zune is “kinda cool” but like a quiet working environment.