Before you invest time, content, relationships or money in any desktop social software play, be forewarned, this idea is already "old hat" and there already several apps out there that combine social networking, chat, and community features. Note that here I am placing emphasis on "desktop" -- my point is not to malign social networking in general, but rather to reveal the weaknesses of any business model that is focused around trying to make money from a desktop software tool for social networking. In contrast to desktop tools for social networking, Web-hosted social networking portals such as LinkedIn, Orkut, Ryze, Tribe and Friendster make a lot of sense and are proving to be highly viral (although not necessarily useful or commercially viable yet). This article presents detailed arguments that make a case for why desktop social networking software tools will not be able to survive in the long-term.
There are already several desktop social networking software products, and none of them have really caught on compared to Web-hosted social networking services like Orkut, LinkedIn, Friendster, etc. I predict that all such products -- at least if they base their business models on making money from their software -- are destined to either remain small and obscure or to fail as large businesses. Two leading examples of "desktop social networking software" are WiredReach and Huminity. While these products are nifty little toys, they do not compete with the functionality, popularity and range of hosted social networking services.
All desktop social networking tools have two things in common: they have been rapidly outpaced by web-based social networking sites, and they have no defensible business and revenue model nor an exit strategy --something we all concluded when we originally considered that as a launch application of our technology.
Why Desktop Social Networking Tools Are at a Disadvantage Compared to Web Hosted Social Networking Services
- It is simply too easy to make a desktop p2p social networking tool - any coder can do this in a short time -- but there is too little defensible advantage and almost no way to differentiate or sell anything there. There will definitely be numerous free, open-source tools of this nature in the next 2 or 3 years. By cobbling together a few existing open source libraries -- or just using libraries like JXTA from Sun, or .Net from Microsoft, such apps can very quickly be created. Any small group of college students can make such an app and give it away for free. In other words, social networking -- at least on it's own -- is just not "hard enough" a problem to build a business around. There is nothing defensible, nothing that is not easily replicated in a short time at low cost.
- You cannot access your social network or information when away from your computer because it is locked in a local application. Hosted, web-based social networking applications do not have this problem since the information is stored in a central, globally-accessible location.
- When your PC is turned off, your presence on the network ends. Nobody can see you. Nobody can interact with you. Hosted services do not have this weakness. Even if you are offline parties can visit your profile and send you messages that are cached for later delivery for example.
- Microsoft and its allies can easily kill emerging desktop social networking products by giving away similar functionality with a future version of Windows; and they will do this as soon as anyone proves that there at least $100 million to be made in the social-networking business OR that social-networking is a requirement in order to maintain competitive leverage against emerging services. Furthermore, there are already rumors of social networking products designed for Longhorn, also see: This article for more info.
- Social Networking and IM are converging. This convergence will favor existing IM network providers over upstarts. Prediction: We will see social networking features emerge in MSN and MSN Messenger in the next two years. In fact, Microsoft is already experimenting with next-generation "friendsware" such as ThreeDegrees. Likewise, AOL - with an estimated 90% of the IM market using AIM or ICQ, can also do this very easily, and in fact recent rumors have emerged that just such an offering is in the works. Likewise Yahoo is looking into offering social networking capabilities as well. If "desktop social networking" (versus "web-site-hosted social networking") takes off as a category it will most likely be via the next-generation versions of existing popular IM clients and will be dominated by existing IM network providers: AOL, Microsoft, and Yahoo.
- Desktop Client-Server social networking systems do not scale to large user-bases without significant distribution of servers. If on the other hand, servers or superpeers are required and also provided then the network can potentially scale but may not actually scale -- the uncertainty comes from the question of who will run these servers and why? Corporations generally will not install and run new servers for anything unless they have to, and not until they are thoroughly convinced of the need, cost-of-ownership, integration, security and reliability of such systems. End-users and techies may run such servers (as is the case with Jabber and Gnutella, for example) on a volunteer basis -- but they generally will only do this unless the value to them of doing so is highly compelling. Social networking applications have yet to demonstrate truly compelling "must-have" value for anyone -- and furthermore there are already so many available alternative social networking systems that "do the job" that there is no good case for running servers as opposed to just participating in a hosted system that is already provided (such as Orkut, LinkedIn, etc.). As a result, unless a social networking application really presents an unprecedented, unique, defensible, "must-have," "killer-app" capability, there is not enough compelling benefit to grow a distributed network of ad-hoc servers -- so far we haven't seen this happening.
- Truly decentralized P2P architectures are not appropriate for broad social networking. With no servers or superpeers in the network bandwidth is consumed rapidly (enterprises will ban such tools as they have in the past). The solution to avoid flooding the network is that the "search horizons" of all peers must be artificially constrained, thereby limiting the maximum effective range of every peer. In other words, a totally peer-to-peer implementation suffers from the trade-off of reducing the size and effectiveness of the social network. This is a paradox -- social networking is supposed to connect you to the universe of people you are not directly connected with -- but if such networks are artificially limited by search horizons then users will not be able to connect with all the parties in the network with whom they might want to interact. In p2p file-sharing networks this problem is solved to some degree by virtue of the fact that the most popular files are replicated around the network so that the probability of locating such files increases for every local search neighborhood. But in social networking systems there is no equivalent to this -- people are not replicated in this manner and furthermore the relative popularity of a person in the network does not directly indicate their potential usefulness to any given party. Decentralized peer-to-peer networks are appropriate for simple file-sharing, but they are terrible if what one wants is a comprehensive global search in which there is a guarantee that "if something or someone exists on the network you will find it" -- truly decentralized peer-to-peer networks cannot make such a guarantee unless the number of nodes in the network is quite small. Only by adding superpeers and servers can this be solved; but we have already seen that client-server networks are just as problematic.
- Desktop social networking apps cannot make enough money from ads. They may try to make money from running ads in their user-interfaces, but ad-supported software is still not a viable business model as numerous past cases have shown; the CPMs are far too low, advertisers don't trust the medium, and end-users find it annoying. Furthermore, most software-advertising networks are associated with "spyware" which further reduces the likelihood of end-user adoption.
- End-users hate to pay for client software and usually won't because they can often get the same thing for free from someone else.
- End-users hate to download new, untrusted, untested software; it's a very significant barrier to entry that instantly cuts down the potential user-base
- Many enterprises do not allow users to install software on their PC's -- this also cuts the potential user-base
- Desktop social software apps may not work through firewalls. This immediately cuts out most corporate users.
- Desktop social software apps may not work across all operating systems and devices. Unless they are implemented in Java or are built for each operating system (a big task).
- Social networking is a "nice to have" not a "must have" technology. Here's why: Social networking tools offer little to no value-add for interactions with people whom you already have relationships with and with whom it is far more efficient to simply send direct e-mail, use the phone, or IM. If you already know someone why bother locating them in a social networking tool and interacting with them via several intermediaries? So where is social networking useful: Social networking tools only add value for locating, establishing relationships with, and interacting with, people whom you do not already have relationships. But how often, and in what contexts, do we need to locate and interact with people whom we don't already know? The answer is that social networking is really only useful for "marketplace" activities such as business development, sales and marketing, dating and classified advertising and communities.
- Most of the people who want to do social networking are already doing it on LinkedIn, Orkut, Friendster, and Ryze. These services work easily through all Web browsers; there is no barrier to entry, no download required -- they are free -- and they have a head-start and are growing fast. There is no real benefit over such services to downloading a new desktop social software app.
- It is highly likely that services such as LinkedIn and Orkut will open up API's to enable third-party software developers to integrate applications with their networks. This will extend the reach of such services down the desktop -- instantly bringing their critical mass into the realm of desktop tools and effectively blocking market-entry by smaller upstarts with desktop-only social networking applications. It will also make it easy for anyone to turn their existing application into a "social network aware" application. This will be the death-knell for applications that just do social networking. All applications will do social networking; there will be no need for dedicated apps for this purpose.
- Open standards for social networking are likely to emerge in the near future. These standards will make personal identities and social networking profile content platform and application independent. Desktop applications in this space will have to either adopt these standards or risk being marginalized. But by adopting such standards they will lose control of their users who may then easily switch to the next "cool" application whenever they want to.
CONCLUSIONS
Desktop social networking software has no viable business model, is not defensible, and suffers from a number of technical disadvantages compared to Web-based social networking services. Companies in this space must find a way to generate revenues that does not depend on their desktop software. While desktop software can be a useful complement to a Web-hosted social networking service, it is not a substitute for such a service.
The best approach may be to combine a Web-hosted social networking service with an optional desktop social networking tool. Even if this combination is created however, it still begs the questions of "what is the business model for social networking?" and "is social networking just a fad?" Social networking as a category has reared its head in the past (if you were there in the "old days" you may remember the original Six Degrees service -- now a distant memory).
I believe that social networking systems are not nearly as useful as well-organized directory portals or decent search engines. So far, while they are fun to play with, I have never found that such services were more effective at locating and connecting to parties whom "I don't already know but want to reach" than simply using Google or Yahoo to locate them directly. Given that social networking as a category is already not that useful, to limit it further by restricting it to the desktop of a particular device only makes matters worse. While there may be some temporary "cool factor" to doing so, I doubt there is any long-term compelling benefit or business model there. Social networking will probably not go away (this time around), but it is likely to soon become either so ubiquitous that nobody notices it or is willing to pay for it, or so annoying that people simply get tired of it.
But this is not the end of the story... It may be possible to use social networking as part of something else that is actually useful and that actually does have a business model -- such as a b2b marketplace, a community-of-interest or community-of-practice portal, a dating and classified advertising service, or within the global scope of a very large collaborative Intranet (in which there are so many participants that parties cannot know everyone directly). However, in such cases social networking is relegated to the status of a "feature" rather than an "application" or a "business."
Another area where desktop social networking may have a defensible play is in highly vertical b2b areas such as those being pioneered by Spoke, and Plaxo to a lesser degree. Within specific industries or business networks highly verticalized, legacy integrated social networking tools might provide more utility compared to more generic, non-integrated Web-based social networking services.
Enterprises may prefer to host and run their own social networks for their intranets, extranets and b2b exchanges and trading networks. But before we get overly excited about the "enterprise social networking opportunity" it might be informative to look at the example of "expertise referral systems" -- a form of enterprise social networking that has been around for several years and has still not caught on significantly in corporations. For example, companies such as Ask.com and Tacit Knowledge Systems -- both of which provide systems for helping people locate and network with experts in large organizations. These technologies are far more powerful than the primitive social networking tools we have been discussing above in this article, yet they have not become breakout hits in the enterprise yet. Bottom line: Before "enterprise social networking" takes off as a category someone will have to demonstrate compelling ROI's for deploying social networks in the enterprise.
In summary then, I conclude that for consumers, and most likely for the majority of professionals, the most widely adopted social networking services will be Web-hosted portals rather than desktop tools. In the desktop category, social networking will most likely merge with IM and will be dominated by existing IM products rather than by dedicated desktop social networking products by new entrants. That said, nobody has yet demonstrated a compelling "must-have" business case or business model for any social networking system -- whether Web or desktop mediated -- and this must happen first or the entire social networking movement will end up as nothing more than another technology fad.
Thanks Stewart, I've fixed that error! Corrected now.
Posted by: Nova Spivack | February 27, 2004 at 04:54 PM
Oop. No HTML. Links are:
- Flickr: http://flickr.com
- Flickr's API: http://flickr.com/services
- WiredReach: http://wiredreach.com
Posted by: Stewart Butterfield | February 27, 2004 at 04:37 PM
Hi Nova,
I'd invite you actually check out Flickr (Ludicorp is the company, not a product): it is entirely web-based. (It is also not a social networking service, but the former is the slightly weirder error, since it does incorporate some socnet stuff as a feature of the application -- the app itself is for real-time media sharing/collaboration/creation/editing, etc.).
It also has open APIs that people are building cool applications of off (not all of which are public yet). Just getting started on it (Flickr development itself only started on Dec. 8 -- still a few releases before we get out of beta ;)
Could be that you were thinking of WiredReach?
I'd agree that there is not much of a business model for desktop social networking apps, but I don't see much of one for, say, Friendster either (LinkedIn, OTOH, could work out).
- Stewart (Ludicorp president)
Posted by: Stewart Butterfield | February 27, 2004 at 04:36 PM
That was a fantastic intelligence report.
Posted by: Denny | February 27, 2004 at 10:47 AM
Nova,
Just to add to your list, look at Spoke (and to a lesser extent Plaxo) as a hybrid desktop/web social networking tool.
Plaxo is only somewhat of a social networking tool, however Spoke is explicitedly a social networking tool but also have a number of client applications that interface with their systems, the client apps gather and collect information from the desktop and/or are plugins to existing tools for interfacing with Spoke's network (brower bar, menu bar in Outlook etc).
While I personally have not found Spoke of very high utility yet, I do think it has one of the more defendable business models in the social networking space - they sell enterprise versions of their software which collect information from systems enterprise wide, as well as leveraging their public network. Very reasonable and easy to understand business model (sell software to companies who can get real measurable benefits by enhancing performance of sales organizations).
Shannon
Posted by: Shannon Clark | February 27, 2004 at 09:48 AM