Value and utility

One of the greatest aspects of blogs is that you have an instant archive of your writings and theory development for later access. Thus when I checked my search engine and couldn’t find my previous writing on value and utility, I panicked. Then I realised that instead of actually posting these thoughts, I had emailed them. Relief! I could access them. But the problem with email is that it is highly unrelaible as a sustainable source of information. Therefore, I have decided to copy a recent conversation I had on the distinction between value and utility in networked environments as a means of ensuring I do not lose these thought processes again.

Feel free to add your thoughts to the discussion in the extended entry; my own have been further clarified since creating this post about a month ago, but I’d appreciate any input from interested parties!

I think my issues with the readings on the so-called “value” of open source are not just about the difference between social values and commercial values, but within the context of commercial values, I’m concerned about the confusion of utility, and price. Often cited in Benkler and other works on open source is the ubiquitous Metcalfe’s law, which states that the value of a network is approximately the square of the number of users. However, there is no value in a one user connecting with no-one on a network – you must have 2 users for it to *be* a network connection – so your calculation *should* be the number of users (n) multiplied by the number of users minus the original user (n-1), over 2 (the number of users required for any pair connection in a network) … n(n-1)/2. This is all known, so it’s not anything new. But to demonstrate, if we assume that there are four users in a network, then the “value” of the network is 4x(4-1)/2 = (4×3)/2 = 6. So, if you had four users (John, Jane, Simon, Sally), you could have the following connections:
Simon-Sally …. 6.

So the idea behind the internet and open source has been that the value of a network increases exponentially with the number of users. In this case, network value is based on number of connections, regardless of how much any user contributes to that network. Okay pretty straight forward so far.

Now Andrew Odlyzko (I mentioned him last night as the guy who originally said that content wasn’t king) has questioned the validity of Metcalfe’s law on the basis of the contribution of users to a network. His calculation is a tad more complex (n by the log of n), but rather than try to explain it, I’ll summarise it that the value of connections between users in a network does not remain consistent for all users as some connections remain disused, or not maximised. So what he says is that the value can be calculated as the number of potential connections, but growing at a rate slower than the quadratic equation implied by Metcalfe’s law.

David Reed argues (as Reed’s law) that the utility of a network grows exponentially with the number of users in a network because he argues that the number of sub-groups that can be formed in a network are not accommodated by Metcalfe’s law. So back to our group, he argues that the number of potential connections actually is:

….. 11.

Reed has found the formula of (2 to the power of n)-n-1=utility actually applies… so with our four people, (2 to the power of 4)-4-1 = 16-4-1 = 11. But where Reed is interesting is that he’s talking about the total number of connections between users as the maximum potential utility of the network, rather than the value of the network. Of course this still doesn’t accommodate for the concerns that Odlyzko raises, but because he differentiates between utility and value, he’s actually giving the ceiling potential value of a network, rather than the actual value. And even if we consider Odlyzko’s formula, we’re still only talking about the ceiling, mean or likely number of connections.

Here’s the interesting bit. (Or at least I think it’s interesting.)

In none of the works I’ve read on open source has there been a clear effort at differentiation between value as the number of connections and value as the number/quality/price of outputs from a networked environment. Leaving aside the intangible benefits of social interaction and community building (psychic value, for want of a better descriptor), I’ve often come across the rather fatal flaw in literature of equating the number of connections in a network as being indicative of the price or quality of goods being derived from a networked environment. Ie: if you have 4 people in a network, then the price you could charge, or the quality of goods being derived from the network is proportionate to the number of connections – 11 – compared with that of a linear production system, which effectively would be proportionate with the number 2 (end-to-end production processes). Trouble is that this isn’t actually about the value of outputs (still leaving aside all the social considerations). It’s about utility. It may well be true that you can derive proportionate utility from a networked environment which cannot be derived from standard production systems, but that doesn’t mean that the number of users in the system will have any bearing at all on either the quality or price of outputs.

And of course none of this has taken into account the cost of implementing, sustaining and maintaining the networks themselves, not the opportunity cost of failure to connect with a networked environment.

Still with me?

Okay so in summary, I think what’s lacking in current literature on open source networks and the rhetoric of value to be derived therefrom is a recognition of different categories of value, and some trends on variations of value that can be derived. When employing open source environments to produce software/hardware/knowledge/application outputs, value could be considered in terms of the following categories (but not limited to just these categories – this is just a start):

* Social production value – the increased level of commitment to a mutual project, based on the psychological influence of community participation;
* Reputational development value – the influence of ambition on specific personalities and the development of leadership in a networked environment;
* Entrepreneurial value – simplified projects, with a small number of users, but attracting risk-takers in a networked environment;
* Raw connection value – the number of participants/productive connections in output production;
* Utility value – the maximum possible opportunities to be derived from integration of all stakeholders in a production process;
* Risk value – the functional aspects of an output that render the output less likely to be subject to litigation, the threat of outputs becoming obsolete, or production inefficiencies;
* Transparency value – the perceived trustworthiness of a networked system in comparison with a closed production process;
* Efficiency value – minimisation of duplicated effort;
* Time value – minimisation of time to delivery/execution in the production process;
* Functional value – maximum functionality of the outputs for use by a mass audience;
* Quality value – the maximum functionality of the outputs for use by custom or niche audiences;
* Repeat business value – the likelihood of an output to spawn further projects/sustained application development.

As I say this is just a start. In addition to this list of output-oriented values, there are also the intangible and decidedly political values of open source systems, not merely measured in terms of outputs, but in terms of outcomes.

* Community development – increased (or at least the amount of) participation and care for community members;
* Cultural development – the development of distinct languages, practices and ethics of open source communities;
* Relationship development – the psychological and social benefits of connectivity between community members;
* Socio-political development – the use of open source networks as a channel for public education and activism;
* Interactive value – a clumsy way of articulating the differences between interaction between community members in an open source system as compared with similar interactions in “real life”.

Yes there are fewer categories of “value” in the social division, but I think their value (to steal Odlyzko’s idea) is probably logarithmic rather than quadratic!

Be Sociable, Share!
This entry was posted in Uncategorized. Bookmark the permalink.