UC Insights Logo
UCI Bannner

Other insights into UC>>

The Economics of Interoperability


by Russell Bennett, UC Insights

October, 2011

Leave a comment

Introduction

Nowadays, it seems that one of the most wished-for UC features is that of seamless interoperability.  In many ways, it is a sign that the UC industry is starting ‘cross the chasm’ when customers have moved on from asking about the feature set or the ROI and now want to know if a UC system will work with other UC systems.  Having spent the last 5 years of my career working on UC interoperability in one way or another, I thought that I should be able to provide some insights on this topic.

[Note that I am not going to get into a laborious explanation of why standards don’t guarantee interoperability, or why achieving interoperability is hard work – but if that interests you, please see this article.]

In several recent papers, I have referred to Metcalfe’s Law as an explanation for various issues surrounding UC.  Several years ago, in a conversation with a UC vendor (who shall remain nameless) an interesting interpretation of Metcalfe’s Law came up:

“A small network that connects to a large network gains disproportionately from the transaction, relative to the large network.”

At the time I dismissed this as unimportant to the conversation at hand.  However, now that I have thought about it more carefully, I realize that this interpretation has some quite profound implications for interoperability that have little to do with technology and more to do with economics and game theory.

An Explanation of Metcalfe’s Law

Robert Metcalfe formulated this notion to describe the so-called ‘network effect’, i.e. that a communications network becomes more valuable when more people (or devices) start to use it.  This was expressed mathematically as:

Network value = n * (n - 1) / 2

[Where ‘n’ is the number of users or devices on the network.]

The network ‘value’ is the sum of all the potential connections that could be created between each of the ‘n’ users in the network.  However, if ‘n’ is a very large number (say 10 million), this is not to say that each of us has 10m contacts; but it becomes more likely that all of our current and future contacts are included in the 10m.  This is where network theory enhances Metcalfe’s Law.

Metcalfe’s Law and UC

Clearly, Metcalfe’s Law and network theory have accurately predicted the development of the internet, and the various services that can be found on the internet, including email and instant messaging as well as social network services such as Facebook and Twitter.  The adoption of UC is also subject to Metcalfe’s Law and network theory, although some UC modalities are affected differently than others.

UC voice communications can take advantage of an entirely different network (i.e. the PSTN) to significantly enhance the UC network effect by leveraging the PSTN network effect.  Gaining access to the PSTN was a key determinant for UC gaining any kind of adoption.  An interesting side note is that this proves a key element of Metcalfe’s Law in that small networks do gain disproportionately by attaching to large networks.

On the other hand, the more esoteric UC modalities such as data collaboration and application sharing (and, to some extent, video conferencing) are subject to the more negative aspects of network effect because few industry standards exist and therefore they tend to be developed in a proprietary fashion.  Even between vendors that offer similar features, these services are not interoperable and therefore customers’ utility of these features is limited.   This will impact the establishment of these modalities as mainstream forms of communication.

Why ‘Metcalfe’ Prevents Asymmetric Interoperability

Expecting UC vendors to spontaneously implement mutual interoperability is to ignore the competitive environment.   There are many competing standards and each vendor’s unique value proposition is based on their proprietary implementation of the standards that they support.  Therefore, inter-vendor interoperability is necessarily a ‘lowest common denominator’ implementation, which negates the benefits of each vendor’s innovations.   Even if there was a single standard that all vendors could agree on, I believe that this would not solve the interoperability problem because, absent any regulatory imperative, there is little incentive for competing firms to interoperate.

The table below illustrates Metcalfe’s Law by showing the network value of a series of networks, ranging from size 1 to 1,000.  The value of a network of size 1 is 0 (self-evidently), and the value of a network of size 100 is 4,950.  The inner (lower, rightmost) matrix of numbers is the combined network value of the two networks on the X and Y axes.

Metcalfe Network Value Table

Network Size, Network Value and Combined Network Value

Note that a network of size 1,000 gets a tiny value lift from interoperating with a network of size 1 (i.e. from 499,500 to 500,500 or an increase of 0.2%), whereas the network of size 1 grows from a value of 0 to a value of 500,500.  However, the engineering work of establishing interoperability between those two networks is same as that of two networks of size 1000 (which yields a 300% increase in network value for each network).

The result of Metcalfe’s Law is that, when considering where to expend scarce resources, a network will focus on the most economically advantageous partners: i.e. everyone will want to interoperate with larger networks.  When applied strictly, this rule of thumb ensures that no-one will interoperate.  However, as with any other commercial or social transaction, common sense normally prevails and partnerships of approximately equal mutual value are negotiated.   Barring any extraneous factors (e.g. a small network being able to offer something of value other than network size) the probability of highly asymmetric interoperability agreements being concluded is normally low.

Interoperability as a Competitive Weapon

Network effect is more than just a cost/benefit analysis tool; it is also a competitive weapon.  The largest network has nothing to gain by interoperating; their best strategy is to drive organic growth while waiting for smaller networks to ‘wither on the vine’.  However, the mid-sized networks can gain market leadership by interoperating with each other; so the risk of this happening has to be weighed by the market leaders.  In order for a market leader to establish an insurmountable lead, they would have to create interoperability with a smaller network (ideally the number 2 network) such that the combined value of all remaining competing networks would not be a threat.

Network effect as a competitive weapon is not a newly recognized issue: a speech by a US Department of Justice official made in 1999 articulated the issue more succinctly than I have done:

“In a network industry, a likely form of such anticompetitive conduct would be for a firm seeking to obtain dominance to degrade its rivals’ access to its network. By denying compatibility, a larger firm will have less to lose by decreasing compatibility than rival firms; the value of the rivals’ networks will decrease more than the value of the larger firm’s network, leaving the larger firm in a better relative position and increasing the likelihood that customers will switch to it. Also, by working to deny rivals or entrants access to its network, a larger firm will deny its rivals the benefits of network effects and raise a barrier to entry.”

Social Network Effect

However, basic network effect only considers network size, whereas in UC, like other forms of social networking, there is a more powerful factor involved, i.e. that of existing human and commercial relationships.  To illustrate this, let’s examine personal social networking, e.g. instant messaging services such as AOL IM, video chat services such as Apple’s Facetime and, of course, Facebook.  Although joining a consumer social network is normally free, there isn’t much value in joining a network of which your friends are not members.  For example, I have no incentive to join Renren.com, the largest social network in China, since I know very few people in China and I cannot read ‘Chinese simplified’.   So with social networking, as with UC, the buying/joining decision is significantly influenced by who else is on that network, almost regardless of overall network size.

In a ‘greenfield’ situation, the social group can collaborate and make a rational determination on which network to join, or which UC technology to purchase.  However, if some members of the group have already made an investment, then the others have little choice but to follow.  Given the dominance of Facebook in consumer social networking, there is no incentive (in the English speaking world) for anyone to join any other network, since it is a reasonable assumption that anyone with whom they might want to be ‘friends’ is already on Facebook.

The graph below illustrates this issue.  The Y axis is the probability of a network joining event; the X axis represents network size in percentage terms of a given addressable market.

Metcalfe Probability Graph

The probability of joining a network, p(Join), with no prior relationships and all other things being equal (e.g. features, price, etc.) has a linear relationship to the size of the network.  However, if you wish to be able to communicate with one other person on that network, then the probability is now the joint probability of both you and the other person (or company) being members of the same network, as shown by p(1 contact).  As can be seen, the probability is affected by every additional contact with whom you wish to communicate.  Once a network has more than 50% of the market, then the probability that you will join that network increases rapidly.  Clearly, this is a virtuous cycle, as the dominant network starts to enroll all new joiners as well as taking members from other networks.  It is this situation that took Facebook to 750m members.

So What Are The Options?

Clearly mobile telephony, email and other synchronous and asynchronous forms of communication have rendered the PSTN obsolescent and there are already indications that the PSTN will cease to exist as a mandated regulatory entity in the foreseeable future.  If UC is to take over as the default means of enterprise communications, then inter-domain interoperability must be established.  The question is what market dynamic will overcome the self-interest described above to make this happen?

Regulation is one option – but the glacial pace of regulation has, so far, been unable to keep up with the pace of UC innovation.  I have written in two other papers ('UC Cloud Service Implications' and 'Microsoft Buys Skype') about the challenges of regulating UC, especially UC in the cloud and this is, by no means, as simple as regulating the PSTN.

Customer pressure is another driving factor, especially where existing investments in technology are required to interoperate with new investments: UC desktop video interoperating with traditional video-conference suites is a key example.  However, in the recently released 2011 Gartner Magic Quadrant for UC, Bern Elliot provided an interesting insight regarding single-vendor UC suites starting to prevail:

“Previously, most vendors offered a UC portfolio — that is, functionality based on a range of elements drawn from their broad array of communications capabilities; in these portfolios, the interoperability of all components was not assured. The emerging generation of UC suites will be an improvement over the UC portfolios, in that interoperability within the suite should be seamless, the administration will be centralized and the deployment will be easier. Although the suites vary in functionality and maturity, they represent an attractive trend, and Gartner expects UC suites to become the dominant product paradigm during the next two years.”

If inter-vendor or inter-domain interoperability becomes a key factor in the buying decision, then this would change the vendor self-interest equation.  Certainly, if two or more mid-sized vendors were able to offer a joint federation feature that started to gain market traction, then this would get the attention of the larger vendors.  However, at the time of writing, only the two largest vendors (Cisco and Microsoft) are able to support inter-domain federation, but not with each other.

Probably the most realistic driver for inter-domain interoperability would be for a market to be established to enable UC vendors to be paid incremental revenues for establishing interoperability.  In other words, they would get to invoice customers for inter-domain communications just as the telephone networks do now.  Either the “Bill and Keep” or “Calling Party Pays” models would work, although the former is simpler to implement.  Having said that, this would ‘open up a can of worms’ on the funding of the IP networks (which are largely owned by telephony companies) and would create numerous issues about Net Neutrality and Quality of Service.   This is a complex subject that merits a paper in its own right: my best attempt so far is in a section of  "UC Endgame Part 1".

Conclusion

There are no easy answers to the interoperability conundrum.  Certainly, free market dynamics create no incentive for UC interoperability, absent a way for UC vendors to be paid for interoperating.  Quite apart from anything else, there is too much money at stake.  Various analysts value the current (2011-12) UC market between $4-15Bn.  In 10 years’ time that will easily grow ten-fold.

‘Big Government’ advocates (who mostly reside outside the US) will try to regulate for this.  However, I expect that, as always, those regulations will have unintended consequences.   Within the US, some entrepreneur will try to set up a business that facilitates UC interoperability.   However, if that business were to be successful, the UC vendors must be enrolled in the scheme, as it is too easy for them to thwart unilateral interoperability.

We, in the UC industry, live in interesting times…


If you liked this article, please comment, share and/or rate it below.   If you didn't like it, please comment!

Comments powered by Disqus+