Posted by: karisyd | May 9, 2017

Do algorithms make better decisions?

Recently, we have heard a lot about algorithms, machine learning and artificial “intelligence” (AI), and the promise these technologies hold for improving decision-making. The argument for the power of algorithmic decision-making rests on two assertions: Algorithms are said to be able to digest vastly greater amounts of data than humans, and thus make more informed decisions. And algorithms are said to be unbiased, and thus make objective decisions. But how realistic are these assertions?

Are algorithms better informed?

First, let us look at the claims about data. Recent research has indeed demonstrated the efficacy of algorithms that implement an advanced form of machine learning, so-called deep learning, for making decisions that rely heavily on pattern recognition. For example, algorithms have been shown to find cancerous cells in vast amounts of images from CT scans faster and with greater precision than human diagnosticians. Microsoft sales agents use deep learning algorithms to decide which lead to contact next, and self-driving cars rely on similar technology for navigating in traffic.

However, in order to judge what such algorithms can (or can’t) do for business decision-making it is important to gain some understanding of how they work. Unlike traditional algorithms where the decision logic is implemented as explicit if-then rules, self-learning algorithms have to be trained with existing data, from which these algorithms learn to infer relevant patterns when later presented with new data of the same kind. There are no rules, the algorithm learns by adjusting a complex, layered network of “neurons” to respond to patterns.

What then are the implications? First, such algorithms only work where the problem domain is well-understood and training data is available. Second, they require a stable environment where future patterns are similar to past ones.

It is easy to see however that many business decisions are not like this, in particular not those that matter for the future of a business. Once we realise that the future is rarely an extrapolation of the past, but actively created, we can see that algorithms that lock us into the past are not appropriate when it comes to forward-looking decision-making.

Consider hiring decisions: algorithms will have to be trained with data on which past hires were successful. A learning algorithms would then allow identifying candidates with the same traits as those successful previously. Yet, often hiring more of the same is not what will be best for the company going forward. It will also compromise diversity and be detrimental whenever the organization has to respond to change or wants to venture into new areas. Also, let’s not forget that such training data relies heavily on hindsight and by its nature is incomplete, because it does not include data on those candidates that were not hired.

Are algorithms unbiased?

Machine-learning algorithms are as unbiased as the data with which they were trained. The above example shows that if simply trained with past hiring data, the algorithm would merely perpetuate past biases. Of course, we could ‘clean up’ the training data to remove biases. But who would we entrust this task to? Whoever gets to decide on the training data will embed their biases in the algorithm.

And crucially, no-one knows, not even the creators of these algorithms, how exactly these algorithms reach their decisions. They resemble black boxes that cannot explain themselves and answer the all-important ‘why?’ question in justifying decisions. What we are left with is: “the computer says no!” Entrusting decisions to such algorithms would mean that we transfer accountability for decisions to those in charge of training them, effectively outsourcing our ethics.

So, where does this leave us?

To be clear, machine-learning works for operational pattern-recognition problems, in particular those involving high volumes of unstructured data. But these algorithms require conditions of ‘business as usual’. Ironically, because of the grounding in past data, this supposedly disruptive technology cannot cope well with disruptive change.

As for decision-making, most situations that matter do not present as clearly delineated options to be weighed up. What is needed rather is human judgement and expertise, and for decision-makers to commit to a particular course of action, guided by a clear purpose and a shared story of what we want the future to look like, and to motivate and convince others to follow, rather than mechanistically entrusting algorithms to make decisions based on the past.

And regarding bias let’s remember that every decision, by definition, involves enacting preferences, valuing some criteria over others. A decision is always biased in some sense; we might not hire on the basis of gender or race, but we might value some personal traits, degree programs or education institutions more highly than others. Rather than black-boxing decisions in an entity that cannot be held accountable we should seek to have an open and transparent conversation about which distinctions are in play in making decisions.

This post appeared first on the University of Sydney MBA Blog.

Posted by: karisyd | July 25, 2016

Pokemon GO – impressions from my self-experiment

A couple of days ago I ended my short but intense excursion into the world of Pokemon GO. Here I am sharing some of my experiences which I want to use to challenge some of the myths about the game. For example, is Pokemon GO really an augmented reality phenomenon? Or will it improve exercise of its mostly younger fan base? What I will not do is provide an extensive game review. I am neither a game critic, nor a credible gamer (and the review would come rather late since much has been said already).

Let me start by saying that I am on holidays, so I had time to get into the game. And time is what is needed to play Pokemon GO seriously. As far as my own achievements go I made it to Level 14. That’s not bad but not good either. My main issue with the game is: I have more or less already done everything there is to do in the game, yet I have not reached a competitive level: I have collected lots of Pokemon (55 different ones, didn’t catch’em all of course), evolved them, trained them, did power ups, competed in gyms and held a few briefly, collected medals and heaps of stuff at Pokestops.

My most advanced Pokemon is a Vaporeon with CP1160 – again not bad but not good enough either. In order to be competitive in the game anything below Level 20 is not good enough. But here is the catch – it gets progressively and exponentially more “expensive” to climb up the level ladder. It cost me about 5 days of gaming and 85,000XP (experience points) to get to level 14. It requires a total of 210,000XP to get to level 20, 710,000XP to get to Level 25 (where many competitive players in busy locations sit) and a staggering 3,000,000XP to get to Level 32 (the highest anyone has made it yet).

In other words, the time commitment to be competitive in the game is insane. And it is mostly about time, there is no skill involved, and it gets very very repetitive. As I said, I have done (pretty much) everything there is to do in the game – from here on I would have to do the same things over and over again, so I called it quits. Have a read of this article by Dominic Knight in the Sydney Morning Herald who nicely puts in words this frustration.

At the same time Pokemon GO is strangely addictive. It sucks you in and is quite fun to play (if it wasn’t for the frequent bugs and server crashes – seriously, is this the biggest beta test ever?). Walking around, finding Pokemon, collecting stuff, evolving Pokemon can be very absorptive. It’s a very sticky game, for better and for worse. But is it augmented reality? Will it have lasting exercise benefits? Is it a collaborative game?

Augmented reality (AR)?

The idea behind augmented reality is to “enhance one’s current perception of reality”, usually by overlaying certain imagery or information over a view of the “real world”. We have all seen the pictures of Pokemon where they seemingly sit in the real world. But here is the catch – this is merely a gimmick in the game, no one in their right mind will use it. It sucks too much battery and the little buggers will move around wildly on the screen. So no serious gamer uses the ‘AR’ feature but the more conventional game screen.

Of course, Pokemon, Pokestops and Gyms are overlaid onto a Google Maps mashup of the real environment – but does that qualify as augmenting reality? I would say, yes, at first at least. It was quite exciting setting out trying to find Pokemon or Pokestops in my environment. But this sense of hunting Pokemon in the “real world” wears off quite quickly, at least it did for me.

What happened instead is that I frequently lost track of my real surroundings. It also didn’t matter that Pokestops showed photos of real landmarks – I swivelled them and moved on, collecting stuff and pushing on hunting more Pokemon to keep those XP and Stardust clocks ticking.

Experientially, I would not describe Pokemon GO as augmented reality. Much like any other well made game it sucks you into its own world. What is new is the way in which the player controls the game – by walking around in the world. So it is more an innovation in game control than a form of augmented reality in my view.

But if Pokemon GO is taken as the flagship concept of AR, it will only muddy the waters and distract from what AR can actually, seriously do.

A great exercise App?

Sure, I was out and about walking  – in my case Townsville in Northern Queensland. But here is the thing. Catching Pokemon requires you to hold still and flick a ball, so most people will stop when they encounter a Pokemon. What is more people walk very slowly while playing (the game is slow and requires lots of tapping and flicking for each pokemon, you also don’t want to run into people). As a result players walking around will invariably be outpaced by everyone else, parents with toddlers, people on crutches, bugs crawling about.

And if you want to catch a lot of Pokemon, you go where there is lots of people, lots of Pokestops and thus lots of Pokemon. But then you won’t need (or want) to walk much. This explains the groups of young people sitting around in parks just harvesting Pokemon as they appear, attracted by the ‘lure module’ someone installed on the nearby Pokestop.

In other words, walking faster than snail pace is counter productive to serious game play. In my view it is more likely people will get seriously sun burned playing the game than walking off those extra calories.

Collaborative game play?

Pokemon GO has all the makings of a great collaborative game. But it is not at the moment. In fact playing can be a very solitairy affair, even though the game leads to mass congregations of people in certain locations. Every player plays for their own profile. Sure, it helps to team up with other people to attack and hold a gym, but that is about it.

There are no in game leaderboards, communication, sharing or swapping of Pokemon or items within the App. Other players do not appear on the in-app map. While some of those features might come in future releases it feels quite incomplete at the moment.

Encounters with other players happen. They are easily indentified because others show the same characteristic behaviour – slow, stooped movement, green glow from their smartphone and a lot of upward swiping. For me those encounters felt quite strange as they are usually very brief, short exchanges about what stuff can be found in a location, or questions about which team I was with or if I was attacking a gym nearby, before each one disappears back into their own collecting frenzy.

At the moment Pokemon GO is a remarkable collective phenomenon but it is not yet a collaborative one.

On a final note

Pokemon GO is certainly fun and addictive. But as with any addiction, the fun part wears off rather quickly. What remains is the compulsive-repetitive behaviour characteristic of any addiction as players chase the next level-up. Plus there is the loss of control and the lack of  sense of one’s surroundings.

Players completely absorbed in the game are rendered into stimulus-response zombies, controlled by and responding to whatever the game engine decides comes their way. I certainly observed myself falling into this pattern. It does not require much fantasy to see that this offers a powerful form of crowd control. Whoever is in charge to decide where the next rare Pokemon will show up has the power to congregate large crowds of people in certain places. This has all the makings of a dark (not-so-science-fiction) movie plot.

So, while I thoroughly enjoyed my short time with Pokemon GO, the game leaves much to desire. But it also is an interesting experiment and a platform on which others will no doubt improve with future ideas and new forms of gameplay.

Posted by: karisyd | June 26, 2016

Principles of the Sharing Economy

This is an edit of a video talk I gave last year for a course on Sustainable Development at The University of Sydney Business School – see video at the end of this article

As the name suggests the “sharing” economy is all about sharing. The basic idea is that, as a user of these services, I don’t have to own everything, I can share things with other people in the economy. I can participate and benefit from other people sharing their assets, products, service or their time with me.

There are many different examples, businesses that claim to be part of the sharing economy. It is worthwhile taking a closer look at the phenomenon and see what we might truthfully call “sharing” (and what not).

For example, are there services that we call sharing (or that claim to be all about sharing), but that are not actually about sharing at all? I am going to argue that there are actually two different notions of the sharing economy.

Some Examples

An Australian company called TuShare set itself the task to reduce waste and landfill (note that the company no longer exists, but the idea is still a good example). The idea was that, if we don’t need a certain item anymore, say a bike, a camera, a piece of clothing – we can make it available to a community of people registered on the platform who would then be able to receive this item and share their own items. This is well and truly a sharing idea.

There are other, similar examples, such as MamaBake, run by parents who join forces to cook together; who cook large batches of food, and then share this in the community with other people, who will then in turn, cook on a different weekday for the community.

On the other hand, we are all aware of examples like Uber and Airbnb, which are often discussed as the front-runners in the sharing economy. The idea here is that as users we share with other people “assets” that we own, like our houses or our cars, because they might sit idle, and might not be used most of the time. If we have spare time we can act as drivers and earn some money “on the side”.

Further examples include TaskRabbit or Airtasker. The idea here is that, if people have spare time, they can take on certain tasks that other people put up on the platform, which they carry out to earn some money on the side. So the asset here, more broadly speaking, is time being shared – that people make available on the platform.

Clearly those examples have certain commonalities, but they also have some very important differences. Let’s look at two main principles that underpin the sharing economy.

Principle 1: Sharing of things and services

The first one, quite obviously, is sharing. The sharing of things, the sharing of what we might call assets, more broadly, such as cars, rooms, time. Using various services people share these assets among each other. But if we look at definitions of the sharing economy and where it came from, where it started out – its roots – it’s an idea that is based on sharing more broadly.

Principle 2: Shared ownership, decision-making, collaboration

The second principle is based on a peer-to-peer, or collaboration aspect, revolving around shared decision making, shared ways of deciding on the rules by which this particular part of the economy is operating, fairness, the greater good, sustainability, reducing waste and alternative ways of organising and doing business. If we look at these principles, we could ask the question (by way of example):

Is Uber actually part of the sharing economy?

First, there is certainly the sharing aspect: people sharing their cars with other people who they drive around. On the other hand though, the peer-to-peer aspect, the collaboration aspect is lacking because Uber is owned and organised centrally. Uber is a commercial entity with a clear profit goal. The actual transactions on the platform are not organised in a peer-to-peer way.

What is also quite noticeable is that, if we look at the conversation around Uber, it’s not so much about sharing itself, but about the convenience the service offers its customers. This individual aspect of Uber is very important, to the point where this is almost defining its narrative now. It’s not about peer-to-peer sharing, about collaboration between drivers at all. In fact, Uber doesn’t quite encourage drivers to talk to each other. Those drivers are individuals, they compete. They provide a service for other individuals.

So the collaborative, the collective, the actual sharing aspect is somewhat lost in the Uber narrative. In the end it is much more about the exploitation of underutilized assets by a central company that incidentally organizes this as a form of sharing among individuals. This however has a very different quality to the second principle I outlined earlier.

The concierge economy

If we take a closer look at this narrative we can see that there’s a lot of other – often smart phone app-based – businesses that have emerged in the market, who claim to be the next Uber, or the Uber of something. Services with which we can outsource picking up parcels from the post office, our washing, or other daily chores to people who have spare time and will take on these tasks. Sharing in these models is very much reduced to: people with spare time “sharing” this time with people who don’t have enough time. These apps bring those two together.

But this is a far cry from the principles of the sharing economy outlined earlier, to the extent that this is model has been called the “concierge economy”. A recent article in the Guardian, for example, makes a good point: this is no longer about sharing, it is about exploitation.

In these examples the sustainability aspect is lost. If we looked at this in a more cynical way, we might say that some of these services are exploiting unemployment, people who have spare time but no access to the primary employment market. But the work arrangement that these apps create can be quite precarious.

What is more, those apps black-box, in a sense, the kind of employment they create. People using these apps, they lodge their tasks, but often they don’t get to see the people who are doing their washing, doing their window cleaning, at the other end of the app. I have heard the term “exploitation economy” being used for this kind of service.

The issue is that these services water down the sharing idea. The moniker “sharing” does not befit these latter cases in my view. Call it “concierge economy” or “on-demand” economy, but Uber hardly embodies the idea or ethics of sharing.

Posted by: karisyd | January 24, 2016

Microsoft and Yammer – a misunderstanding?

A couple of days ago news broke that Microsoft is laying off its entire team of Yammer (now Office365) Customer Success Managers (CSMs). To me this is the strongest indicator yet that Microsoft has missed the point of what Yammer embodies and that the company has not come to grips with the opportunity that the acquisition of Yammer presented. However, I am not surprised. It was always an uphill battle for Microsoft to understand the nature and potential of Yammer, because its very idea was so very different to the business that Microsoft engages in.

Yammer the startup had a vision. It was to make the world of work more transparent and connected, to break open the rigid structures in corporations and to let information travel freely for the good of more collaboration, innovation and responsiveness. Yammer the platform was the conduit, the trojan-horse so to speak, to achieve such an ambitious social change agenda.

In 2012 I had the fortune to visit Yammer the company at its old headquarters in San Francisco to carry out a set of interviews. It became clear very quickly that the company very much embodied and lived this ideal and was drawing on its own experience in driving the development of Yammer the platform and the change in its customer organisations: “we want our customers to become more like Yammer the company” was a frequently heard statement. To become a place in which work happens in the open, where problems do not linger, help is offered and acknowledged, and people are motivated by being part of something they understand and believe in. Yammer was to be the platform that enabled customer organisations to pursue this vision. And the CSMs were the people who worked with those organisations in guiding them in this process. The crucial role of the CSM was stressed often and for good reason.

So what is Yammer (the platform)? A simple enough question. But for a company that starts out on the implementation process not an easy one to answer. Yes, it is social software, it is an ESN, but what to do with it? Things are what they are for. But what is Yammer for? Many things, different things. The point is, it is an infrastructure for making change happen – its uses and affordances are specific to a context, they have to be discovered through experimentation over time – they might not initially be clear. As a consequence the adoption process is not straight-forward, because Yammer is not a tool for a particular task, it doesn’t plug a hole or address an immediate problem for the most part. I have written and talked about this crucial difference between tool and infrastructure previously.

This is now where the role of the CSM comes in – CSMs understand this process, how it can be guided, how adoption and diffusion can be grown, the success stories be shared, users be encouraged to persist even as adoption might not proceed smoothly and linearly (as illustrated in the SNEP model). In other words, what CSMs do is crucial to the success of Yammer and its vision of openness with all the associated benefits. CSMs help organisations find a place for Yammer, what we call “place-making“.

But Microsoft at heart is a software company, it builds tools and sells products. A look at its licencing models should be enough to understand the way in which the company understands the world. If you are in this business, the focus is on products and its features, on selling licenses. The vision is very different naturally. To literally see what Yammer is (a platform for change, not a tool for a job) is difficult, if the corporate ontology doesn’t have a place for it.

So, Yammer became an add-on to other products. This is not to say that Microsoft lacks commitment to Yammer, but to say that Yammer the product for Microsoft is very different to what it was for Yammer the company. Take the statement by Microsoft Office Division Senior Director Jared Spataro who in 2013 confirm the commitment of Microsoft to Yammer:

“Yammer is our big bet for enterprise social, and we’re committed to making it the underlying social layer for all of our products. It will power the social experiences in SharePoint, Office 365, Dynamics and more. Yammer’s unique adoption model appeals directly to end users and makes it easy to start enjoying the benefits of social immediately.”

Note two things: 1) Yammer is a product, a social layer for other products (not a vision for change), and more importantly, 2) “users enjoy the benefits of social immediately”. But far from it. Granted, you can start exchanging messages straight away once you have a login. But the true and deep implementation and adoption of ESN is non-trivial, needs work and commitment. The true benefits will only emerge over time. In many cases it was the CSMs who did the hard work with organisations in making it happen.

But in a product world, where business is selling licences, in installing products, what role does a CSM play? After-sales services at best, a hidden cost at worst, dispensable the moment the company takes to cust-cutting.

We are only at the beginning of the evolution of social technologies and the changes that these platforms have to offer to organisations in rethinking management models, in finding new ways of engagement internally and externally. But it is clear to me that it requires different approaches to managing technology, to integrating technology into businesses. CSMs were Microsoft’s champions, experts, true assets and bearers of knowledge of how to innovate this part of technology management.

Their role could have been elevated, the company could have learned a lot from their expertise. Letting them go is a curious business decision. I am sure there were good business reasons that made much sense for Microsoft in letting them go – that is if you are in the tool, not in the infrastructure business.

The people in question will find good employment elsewhere, they have much needed skills to drive transformative change in organisations. What it will do to Yammer and its success is a different question that only time will answer.

When it comes to Digital Disruption, one of the most vexing and important questions is:

Why do incumbent businesses have such a hard time dealing with digital disruption even when it unfolds right in front of them?

Drawing on my work and experience in this field I have distilled a number of important factors into a framework, which I name the VIRUS model. The acronym emerged conveniently from the process of isolating these factors, but carries a deeper meaning: It captures the ways in which the disruptive product or service is able to emerge slowly, steadily and unrecognised – when symptoms are first noticed by the wider market, it is often too late, and full-blown disease strikes.

VIRUS stands for: Visibility, Information, Risk, Utility, and Speed. Each of the factors are explained below.

VISIBILITYCan’t fight what you can’t see.
Despite what the name suggests ‘disruption’ doesn’t happen suddenly. The disruptive technology, product or service usually has been around for a while before it unfolds its disruptive potential. Why then do we frequently (dis)miss it? Because the disruption typically doesn’t make sense initially; incumbents literally can’t see the disruptive potential in emerging ideas. This is because disruptive innovation is revolutionary, not just evolutionary, it is path-breaking – it challenges the background on which the industry is currently understood. Therefore it appears as irrelevant, as a niche or fringe product initially. Yet, the disruption happens when it brings about a tectonic shift in understanding of what counts as a valid product, which can catapult the disruptor from the fringe to the core and the incumbents to the margins in a very short period of time.

Think of mp3 and CDs, or the iPhone and Blackberry/Nokia – initially dismissed as fringe products they have redefined the very idea of what counts as music media or a mobile phone – a fact that appears self-evident in hindsight but not while unfolding. Neither the first generation of mp3 players, nor the initial iPhone were a great success, yet both have disrupted entire industries, relegating previous incumbents to the fringes. The problem is to know before the fact which of the many (sometimes competing) emerging ideas will have that effect.

INFORMATIONInformation rules!
Software is eating the world, according to Marc Andreessen. Digital Disruptors change the nature of competition in many industries from a dominance of physical assets (hardware) to a business dominated by software and digital information and data. Digital Disruptors are ‘Information First’ businesses; they change the rules of competition by becoming very good at working with data, collecting and exploiting information to add value to the industry. They turn physical into digital industries. Because of the very different nature in their business model, these emerging ideas are easily misunderstood or dismissed initially.

Both mp3 and the iPhone are good examples of this, mp3 has turned a formerly physical into a digital product. The iPhone has redefined the mobile space from a hardware to a software dominated one. Further examples are Uber, Airbnb, Yelp or Tripadvisor all of which redefine business not by owning the physical assets in their respective industries, but by redirecting customer allocation and value creation streams by exploiting information and data in innovative ways.

RISK:  Risk adversity is the greatest risk.
Incumbent businesses become hamstrung by their own success. In stable markets, asset exploitation, efficiency and compliance through process optimisation and risk management through rigorous budgeting processes all make sense and underpin success. However, when markets are disrupted those traits become the greatest risk. When faced with a disruption those structures make it hard to innovate and change, all the while the existing business acts as a powerful disincentive to necessary self-disruption. First, there is the fear of self-cannibalising what is still a profitable business in favour of a new way of doing business that is not yet proven to work. Second, internal incentive structures are built on the old way of doing business, the risk of which is that people will not be inclined to get involved with something that doesn’t add to their KPIs, leading to the “not involved here” phenomenon. Finally, budget processes are based on rigorous cost-benefit analysis; yet benefits are foundamentally unknowable when it comes to disruptive change (as I have argued previously for the NBN example). Risk adversity becomes an inhibitor of the capacity to innovate internally.

Take Kodak for example: Kodak had all the technology and patents to be a leader in digital photography, but could not pull it off for the above reasons.

UTILITY: Different, not just better!
Clayton Christensen in his work on disruptive innovation has argued that new products or services initially start out as inferior to the incumbent product, which makes them appear harmless in the short term, but that they eat away at and slowly emerge as a powerful and disruptive alternative to incumbent products. So, initially the product’s utility appears inferior, but later it’s not. My point is that the change that happens is not just one of linear improvement, but a subtle, yet radical change in the understanding of what counts as utility in the market in the first place. Disruptors are not delivering an initially inferior, then better solution – in essence, they do something different and thereby, over time, redefine the rules of the market. Once this tectonic shift in what counts as utility happens, their product appears as vastly superior – but only on this new understanding.

Take mp3 again – initially it appeared inferior in terms of sound quality to the CD (it still is by the way!). But our understanding of music consumption has changed fundamentally. When the original iPod was released many people asked “what do I need 1000 songs in my pocket for?”. Today we take mobile music consumption for granted, with streaming of anything anywhere a given reality – this marks a tectonic shift in what counts as the actual product!

SPEED: Late but slow…
This last one is the accumulation or outcome of the previous factors. Once the digital disruption is widely recognised within an industry the disruptor tends to have a strong head start on the incumbent players. And because of the inertia of existing business, the shift in perception of understanding, and the ways in which internal structures tend to hamstring the incumbents, reacting to disruption becomes an uphill battle. Remember: disruptors not only came earlier to what is now a different market environment, they are also quicker in execution…

Well, all of this then raises the next big question – what can the incumbent do?

I will return to this question in one of my next posts.

My thanks goes to all colleagues in the Digital Disruption Research Group, and in particular Ben Gilchriest at Capgemini, all of which have inspired and contributed to these thoughts through joint work and discussions.

Posted by: karisyd | March 5, 2015

ESN communities – built from top to bottom

Or: How influence in ESN changes over time

In a recently published paper we report on a study we carried out with data from the ESN network at Deloitte Australia. We investigate the ways in which users derive influence from a) their position in the company hierarchy and b) their activity in the ESN. Importantly, we measure how these forms of influence (formal and informal) change over time as the ESN community forms and matures.

What we did

We measured a) if users in higher positions in the hierarchy derive more influence from their position and b) if more active users (measured by number of messages posted) derive more influence than less active users. Influence is measured as the average number of replies a user elicits for each message they post. The assumption is that it is a sign of influence when users are able to get more responses to their messages from the community.

We then split the 110,000 message in our data set into three time periods with equal number of messages in order to capture any changes in the above measurements – to see if influence in the network changes as the community emerges, grows and matures.

Please note that due to the large scale and quantitative nature of the analysis, we only utilise structural data (meta data), e.g. who responds to whom, but did not take into account the actual message content. We would also like to note that, while only based on one case study, the Deloitte case is ideally suited to study influence, for three reasons: 1) as a professional service firm, hierarchy is an important mechanism for work allocation, so that the case allows studying the role of formal influence, 2) much of the work at Deloitte is knowledge work, which means the ESN plays an active part in information search and knowledge work, which allows studying the effects of user contributions (informal influence), 3) the Deloitte ESN is highly successful and widely adopted in the organisation, which makes it a good candidate to study ESN community emergence.

What we found

Formal influence is only present early:

  • The formal influence that users derive from their position in the organisational hierarchy is present only in the early stages of ESN emergence. It disappears in later periods.
  • We also find that ESN afford users to move into positions of informal influence by way of their contributions to the network – more active users are more influential. This influence is strongest in the early stage, meaning that early adopters move into influential positions initially. While still present in later stages, the influence diminishes.

The community becomes more egalitarian over time:

  • Both forms of influence, formal and informal, diminish or disappear over time, which means that the ESN produces more egalitarian and inclusive communication structures as the community matures.
  • While quite imbalanced early, all hierarchical levels become more equally involved with the ESN over time, and communication between hierarchical levels intensifies (see figure 1).
Figure 1: Average number of messages per user on each level of hierarchy

esn community1

What it means

With this study we are able to find support for the widely held assumptions that a) ESN will allow active users to gain influence regardless of their position in the formal hierarchy, and that b) ESN lead to more egalitarian and democratic structures of participation. Figure 2 summarises our findings.

Given that the Deloitte case presents a highly successful ESN we can also make conclusions regarding the way in which successful ESN proliferate: they resemble existing influence structures early, but build out unique structures over time. And ESN appear to be a vehicle to get the organisation talking across hierarchy levels.

On a final note, the case shows that the senior leadership of the company (e.g. partners) were very active early in the life of the community, which we see as an important success factor, because it legitimises the use of the ESN (leading by example).

Figure 2: Summary of ESN community development in three stages

esn community2


Riemer K, Stieglitz S and Meske C 2015 ‘From Top to Bottom: Investigating the changing role of hierarchy and influence in Enterprise Social Networks’, Business & Information Systems Engineering, DOI 10.1007/s12599-015-0375-3. (if you would like to receive the full paper, please email me:


Posted by: karisyd | February 28, 2015

What drives value in ESN?

In a recent survey of 193 Yammer users my colleague Matti Mäntymäki and I investigated which forms of usage drive perceived individual value in ESN. The study was based on prior case studies I had carried out with various other colleagues in which we identified five major use cases of ESN: the ways in which Yammer was mainly used within the workplace.

These five use cases are: 1) Discussions and informal talk (people discuss various matters of interest), 2) Status updates and event notifications (people tell each other what they are working on and about events in their work environment), 3) Idea and input generation (people post links to interesting information and brainstorm new ideas), 4) Problem-solving (people crowdsource solutions to pressing work problems), and 5) Social feedback (people share success stories and praise each other).

While these studies have shown that all of these use cases contribute to the shared value for the organisation, these cases were all based on the analysis of Yammer messages – what those people do that actively post to the Yammer space.

But what about the silent majority of passive users (note: don’t call them lurkers – they’re not)? So in this study we wanted to know, what makes Yammer valuable more generally for all users?

We went through a multi-stage process to rigorously develop and operationalise constructs and an ensuing questionnaire instrument, which we distributed to Yammer users across three organisations.

The results

Our results (see figure below) show that exchanging and obtaining ideas and information is by far the most important source of value for ESN users. Users see ESN as a space for sharing information with other users, for discussing ideas and for reading and obtaining information that feed into the users’ work processes. This speaks to the usefulness of ESN in the context of knowledge work and for knowledge exchange.

Moreover, users also perceive problem-solving and updates & events as valuable, yet to a lesser extent. The other two constructs were non-significant.

But does that mean that informal discussions and the social aspects of ESN communication are useless and should be avoided? Of course not, it only means that when asked directly, users find value mostly in the information sharing aspects of ESN. The social aspects however are necessary because they underpin the ESN community, without which there wouldn’t be a useful space for information sharing to begin with.

esn survey


Riemer K and Mantymaki M 2014 ‘Information, Ideas and Input: The Value of Enterprise Social Networks’, Proceedings of the 25th Australasian Conference of Information Systems ACIS 2014, Auckland, New Zealand, 10th December 2014 – Download full paper

Posted by: karisyd | September 8, 2014

Activity vs hierarchy – influence in ESN

In a recent study we looked into the ways in which users gain influence in Enterprise Social Networks: Is it through their position in the organisational hierarchy (formal influence) or by way of being an active contributor to the network (informal influence)?

We measured influence through the ability to garner responses to one’s messages posted into the network. In other words, does my ability to convince others in the network to reply to my messages depend on what my formal role is or does it depend on whether or not I am an active contributor myself.

Study approach

For this study we had access to the ESN data of Deloitte Australia, who were also able to provide data on the formal role and hierarchy position of about 65% of their users. All data was anonymous; users where only identified through an ID.

We operationalised three separate dependent variables: reply received yes/no, number of replies received, and time lag until first reply received.

We then hypothesized that 1) the higher one’s position, the more elikely they are to receive a reply, the more likely they are to receive more replies, and faster. 2) The same hypotheses were formulated for activity: The more active someone is in the network, the more likely they are to receive more replies, and faster.

After statistical testing, our results confirm both forms of influence. Interestingly however, the informal influence of communication activity is the much stronger effect. 

Hierarchy in ESN matters – but not much

On average, a user’s hierarchical level has a statistically significant influence, but this influence is so small that it does not to matter in day-to-day practice. For example, a user who outranks another user by one level in the hierarchy is able to elicit on average 1% more responses.

Interestingly, messages from users in higher positions elicit slightly slower responses on average. We reason that people might need more time to formulate an adequate answer or take more care in editing replies to users in higher positions. The reason might be the perceived social distance between sender and receiver.

Being active matters most

The strongest finding from our study is that a user’s communication behaviour has a much stronger effect on response behaviour than their hierarchical position. Our findings thus confirm the long-standing argument put forward by proponents of ESN that the social networks that emerge on ESN platforms can lead to a re-balancing of influence in organisations – away from formal hierarchy toward recognising user contributions.

In other words, people who have something to contribute will be recognised by the organisational community and be able to derive influence from their standing in the community, even if they do not sit in high-level hierarchical positions.

At the same time, our findings show that formal hierarchy does not lose its influence entirely, as both formal and informal hierarchy show up in our data. We also need to point out that the results operate on averages, it does not mean that single individuals will not derive significant ESN influence from their position on the org chart.

Yet, being the first of its kind, our study demonstrates that prolific knowledge workers might benefit from their contributions to and investment into the ESN, because they are able to draw on the network for contributions, not having to rely merely on information flows along the organisational hierarchy. We are conducting further research to unpack these effects with more detailed analyses. I will post more results in due course.


Stieglitz S, Riemer K and Meske C 2014 ‘Hierarchy or
Activity? The Role of Formal and Informal Influence in Eliciting
Responses From Enterprise Social Networks’, Proceedings of the 22nd European Conference on Information Systems ECIS, Tel Aviv, Israel, 11th June 2014. Download the paper.

Posted by: karisyd | August 29, 2014

NBN cost-benefit analysis fails the imagination test

We now have our long-overdue cost-benefit analysis for the NBN.

The current government and many in the business community have long asked how an infrastructure project such as the NBN could possibly have been launched without a thorough cost-benefit analysis – common sense, really.

The analysis, commissioned by Communications Minister Malcolm Turnbull, has found the cheaper mixed-technology model (MTM) proposed by the government emerges as the cost-effective winner. 

The former government’s Fibre-to-the-Premise (FTTP) model was found to deliver essentially the same benefits at much higher cost and later, and has subsequently been decried as the result of technological romanticism and reckless spending. Again, common sense, many people would say.

And so we move on and miss a crucial opportunity to unlock the next wave of technological innovation, all because of a fundamental lack of understanding of the nature of infrastructure technologies, and because of common sense.

Drop your tools

Unfortunately, common sense is not always a good adviser. The organisational theorist Karl Weick famously called on the management community to “Drop your tools!” in order to effectively deal with unforeseen situations.

My point is that our established management tools, such as business plans and cost-benefit analyses, might well work for decisions about everyday technology that provide incremental change, but they fail when applied to potentially game-changing infrastructure technology.

According to Weick the problem is that these management tools are “tools of rationality” that “presume that the world is stable, knowable, and predictable.” But the world is not stable and the role of an infrastructure technology such as FTTP is precisely to unlock innovation and bring about change.

It is inherently impossible to undertake a cost-benefit analysis in any credible way that will do justice to such technologies. The problem is that, while we are well able to extrapolate the cost of building the NBN, the benefits it will unlock are fundamentally unknowable and unpredictable.

Think about it. The current analysis purports to envision cost and benefits until the year 2040. Now think back 25 years to 1989, to what was the pre-world wide web era, would we have been able to predict in any way how the internet and all the services and products that were built on top of it would have enabled the world we live in today? Just 8 years back, could anyone have been able to predict how the iPhone’s touch screen platform would unlock the current spur in technological innovation? If the answer to both is no, we have the answer to the question of whether the current analysis can in any credible way predict the benefits of FTTP.

One might object that the above technologies were not built by governments but by the business community. Correct, but irrelevant. The issue is that FTTP would not be built by the business community, because of an inherent “chicken and egg” problem: Users will only demand FTTP if there are services using it, but it will only be economical to build such services once a large enough installed base of FTTP exists.

We need Imagination, not Rationality

The problem in making decisions about game-changing technology is that our economic tools are tools of rationality, not imagination. They predict the future on the basis of our past. They are thus always biased towards the more conservative solution.

This is evident in the way the promises of the NBN are typically discussed in both the report and the wider debate: mostly in terms of higher download speeds with associated benefits for industries such as media and entertainment. Both MTM and FTTP will provide higher download speeds. But the promise of FTTP lies in the way it is able to provide each user with the same high upload as download speeds. While the report outlines quite well the much higher upload speed, it does not explicitly discriminate its analysis in this way. This is not surprising, because we have no way of knowing what users and small businesses might possibly do with such upload speeds. And so it is not part of the conversation. But that is precisely the point – we have no way of knowing.

Yet, we might begin to imagine far reaching innovations in health and education, new ways of organising work and changes to the ways we live, with flow-on changes to the make-up of our suburbs, and maybe even ways to relieve our congested city centres. Yet, not surprisingly, the current analysis comes to the conclusion that the benefits for health and education “will probably be extremely limited” as these sectors do not require much higher download speeds to deliver their services. Again, common sense, but missing the point.

There was much wrong with Labor’s NBN, such as the political nature of picking electorates for roll-out, blunders in setting up NBN Co and the failure to sell its ideas convincingly, but it wasn’t the lack of a cost-benefit analysis or the choice of technology.

We need to drop our tools and start imagining, rather than extrapolating. True innovation has always been a product of romanticism, not rationality.


Posted by: karisyd | May 18, 2014

Why work gamification is a bad idea

This is a blog post that I wanted to write for quite a while now to express my uneasiness with the idea of gamification in the workplace. I will outline why I think work gamification won’t work beyond the short-term and why it is an ethically and economically questionable approach.

What is gamification? Why is it used?

Gamification has been a topic of interest for quite a while now. Gamification refers to “the use of game thinking and game mechanics in non-game contexts to engage users in solving problems”. Typical game mechanics are scores that are earned for mastering certain activities, badges, so-called level-ups and finally a scoreboard, often to rank players.

Gamification rests on the observation that most people like playing games, and that people become truly immersed and engaged and won’t mind spending time doing so. This has led to the belief that it is part of human nature, that people are intrinsicly motivated to play games and that this motivation can be drawn on to engage people in other contexts as well.

‘The opportunities for gamification are everywhere, and everyone is a gamer – it’s part of human nature. By applying these tactics, you can employ subtle psychological responses that will keep your customers paying and engaging.” Fast Company 5th March 2014

Gamification is widely used in many different contexts. For example,

  • In education so-called serious games are used to immerse the learner in a situational experience to get an educational point across. For example, the beergame lets players experuence first-hand the effects of systems dynamics.
  • In advertising game mechanics are used to entice people to share marketing messages to create viral messaging effects.
  • In health policy gamification is experimented with to break people’s habits to combat smoking or obesity.
  • In scientific research, gamification is used to engage large crowds to solve large-scale problems.

Why gamification in the workplace?

The reason for taking gamification into the workplace is simple: Many people in modern workplaces are said to be disengaged (studies show numbers as high as 70%), and gamification has proven to be successful in engaging people in contexts such as the ones listed above. Disengagement negatively impacts on business productivity. More concretely, people are often unwilling to carry out certain tasks, such as data entry into business software, which results in a lack of data quality.

Against this background, gamification looks like an appealing solution. If we were able to turn work that has proven to be disengaging into a game we could thus appeal to people’s intrinsic game-play motivation. The result would be to re-engage people and to get them to carry out those menial tasks that now feel less like boring work and more like a game. Through earning points, badges and level-ups and through the appearance in public scoreboards people would gain recognition and would thus be motivated and more engaged in their work again. This is the argument in simple terms.

What is the problem then?

On the one hand, we have disengaged people; on the other hand, we have a method that engages people. This looks like a perfect match, doesn’t it?

It does, but it isn’t!

Here are four points why I think it is a bad idea. I will elaborate each of these below:

  1. Gamification is a short-term concept.
  2. Gamification addressed the symptoms of a broken system, but does nothing to fix it.
  3. Gamification is disrespectful of employees.
  4. Gamification only looks good on a simplistic understanding of human activity.

Gamification is a short-term concept

Gamification is a short-term concept. First, it is in the nature of most games that they have some end point. Second, most games lose their appeal after a while (it is only fun for a while to earn that next badge). Finally, the usage of scoreboards exhausts itself fairly quickly. Once the first positions on the board are taken, what is there to compete for for the rest of the game population?

This last point raises another issue. Games are typically competitive; individuals compete for recognition. While gamification works well in certain short-term or one-off activities such as campaign management, in the early stages of community building or even in graduate recruiting, the question arises if the short-term and competitive nature is congruent with building sustainable work systems.

Gamification only treats the symptoms of a broken system

The main problem with using gamification to engage people at work is that it only treats the symptoms of a system that is broken at a deeper level. Shouldn’t we rather ask why so many people are disengaged at work?

Very often, people are disengaged because they can’t see the point of what they are doing, the purpose of their tasks in the greater whole of the work system. Yet, if people can’t see the purpose of their work, then this is the actual problem, disengagement is merely the consequence.

Unfortunately, this is a common disease in many work systems that are designed to follow strict efficiency criteria. The wide-spread and often top-down application of business process re-engineering methods has led to the creation of structured business processes with atomistic tasks that come with clear rules for execution. Yet, such work systems, designed for efficiency, come with unfortunate by-products. Besides an in-built lack of flexibility, the most notable one is that people are degraded to task bearers, to cogs in a well-oiled machine. In such a system it is not important, not possible, and indeed often not wanted, that the individual task bearer gains too much overview of the entire process.

The flip-side however is that people become disenfranchised from the overall enterprise of the organisation. This is where the actual problem lies! If people can no longer see, understand and indeed buy into the greater purpose of the organisation, but are instead treated as labour-for-hire, disengagement is the logical result. I am of course painting a certain picture here, but all too often this is the reality in organisations and the reason for the high levels of disengaged employees. But when people become disenfranchised from work, no gamification will ultimately save the day.

This raises the question of why to use gamification at all, why not fix the underlying problem? First of all, the idea of gamification as a tool “to fix people” is well compatible with the ideology and values that created the structured work system in the first place. Second, fixing the system will be much harder, requires a rethinking of management approaches at a deeper level and requires genuine leadership, all of which is much harder to achieve than the quick-fix promise of engagement through gamification.

Gamification is disrespectful of employees

In applying gamification in the workplace there is a real danger that, even though it might achieve some intended benefits in the short term, it will actually worsen the situation in the long term.

Gamification, despite its intricate features and the game-like environment it creates, is a blunt instrument. Its aim is to change people’s behaviour through making them play games. But does this not amount to trickery? – “You don’t want to do this task, so I make you play a game in the course of which you will do it anyway?” Will people not see through this? And for people who are already disengaged and disenfranchised, will this not make them even more cynical about how they are treated by the organisation? Rather than being empowered and treated seriously as collaborators in a greater enterprise they are now turned into game players. In my view this raises important ethical questions regarding how gamification and the game design aims to exert agency over people through turning them into gamers.

Gamification rests on a narrow understanding of human nature

Gamification is often motivated using a philosophical argument, where gaming is located as an inherently human trait. Playing games is said to be part of human nature which apparently justifies its use in influencing people. After all, gamification only treats people as what they are anyway, game-players at heart. It must be ok then.

But what it actually assumes is that people have certain inherent traits (such as a propensity of gaming) and are bearers of certain behaviours (such as engaging in smoking or refusing to work). Two things are wrong with this view: First it treats people as individuals, when they are in fact social beings at base. Second, it adopts a narrow view of behaviour, as something belonging to the individual. While this view is all too prevalent in everyday thinking (and many academic disciplines), it has long been challenged.

It is well known that we are social beings, who strive for purpose in life. We assume multiple identities in different contexts. We are who we are through becoming part of different social practices. In short we are what we do and we draw our purpose in life from our practical involvement in various activities, at work and otherwise. And we spend much of our awake time at work!

When people care about and are invested in their work, when they draw a sense of purpose and identity form their work, when they understand themselves as part of a greater whole, gamification is not needed. Rather than trying to change people’s behaviour when they are reluctant to engage in tasks, the point of which is lost on them, organisations should ensure that people to draw a genuine sense of identity from their work. People are truly engaged when work is meaningful to them, when they can see the purpose of what they do in contributing to the overall purpose of the organisation.

Yet, this raises two more important points:

First, in many organisations this original purpose (e.g. to become a world-class full service airline) has already been lost and been replaced by a narrow and all too generic quest for profit or (worse even) cost reduction.

Second, while effective in increasing productivity and efficiency in stable environments, the top-down, structured work system (not least through disenfranchising people) poses a considerable risk when the organisation needs to change and become more responsive. Such responsiveness can only be truly achieved systemically, through the self-organising ability of the organisation. But this requires empowered employees who care about and understand the joint enterprise and can engage in the joint sense-making to reorient the organisation in the face of external change. Work gamification and the idea of the responsive organisation don’t gel.

Older Posts »