Blogs

RedMonk

Skip to content

Network Offering: If You Build It, I Will Buy It (And Some Other Folks Might, Too)

Over the past 12 months or so, I’ve been having numerous offline conversations and minor blog mentions of what I’ve variously referred to as a network offering, productized package management, and more. The idea is in actuality fairly simple, but does have a number of moving parts and is not terribly easily described. As mentioned the other day, I had an entry mostly complete on this subject but didn’t post it because it was still opaque and not terribly descriptive. Casting about for another approach, one that would allow me to explain it succinctly, I turned to my old friend the Q&A. What follows is a simple question and answer session for a product I’d dearly love for someone to build.

Q: Before we begin, any disclosures to make?
A: No. While I’m sure the idea has precursors, and is in fact a mere amalgamation of several existing businesses, the idea is mine. I’ve spoken of it to a variety of parties in an attempt to entice enough interest to get someone to build it, but the idea itself is not the confidential property of anyone but me.

Q: What’s the short version of the idea?
A: It’s basically an extension of the existing network offering model that Matt mentions here – the model increasingly popular within the open source world. Basically, the idea is to take these network offerings to the next level, by leveraging, well, network effects. As I put it in an email to a CEO not too long ago, “think Red Hat/MySQL network (platform bits) + AppExchange/Ubuntu Marketplace (commercial apps) + Portage / Apt (community apps) + Craigslist (job postings / resource brokering) + eBay (bids + rating system).” That’s the short version.

Q: Where’s the overlap there, what are the synergies?
A: Well, ask yourself whether or not it makes sense that I can install any one of a couple of thousand applications and then have no one to configure and set it up for me? Or that we have libraries with thousands of applications and job boards proliferating like rabbits and thousands of very talented developers seeking employment, but no connection between them? Or that we can rate sellers of pretty much every kind of material good there is on eBay, but very little ability to do so with developers or their (potential) customers? Those are the questions that I’ve been thinking more and more about.

Q: From a customer perspective, what problem does this solve?
A: As Pyra/Blogger/Odeo’s Evan Williams says, it makes sense to try and build – or in this case, get someone else to build – for people like ourselves. My day job, of course, is not necessarily to do that; it’s to understand the needs of a variety of constituencies from businesses like ours to the Fortune 50. But in this case, the idea grew out of a basic problem that we, and lots of other businesses I know of, have in spades: getting things done. Technical things.

Take something as simple as the Apache domain setup process I went through here. It’s a trivial technical task – literally 2 minutes – but one that I had to learn to do, which took a lot longer than 2 minutes. Now I don’t mind doing that, because it’s good for me as an analyst to keep a hands on relationship with all aspects of our infrastructure, but most small businesses wouldn’t even consider doing that.

So who do they turn to for help? The best case scenario is that they a.) have a talented internal IT staff or b.) an established relationship with an SI, and that either A or B has good Apache skills. But what if either of those is not true? Who do they turn to? Craigslist? Monster.com? Both seem like overkill for a two minute job. And they don’t necessarily provide any feedback, any guarantee of quality.

Or take a more obscure need, like backup / disaster recovery, firewall configuration or VOIP: who would you get to build such a solution? What applications would they use? What if you already have an application in mind – say, Asterisk for VOIP – how would you find someone capable in that application for a quick gig?

Basically, then, I’d describe the business problem as connecting those with technical problems with the right resources, leveraging network effects for efficiency and trust. How does a customer get from point A, they have a problem, to point B, they have a solution to said problem? By connecting the applications with the developers and implementers experienced with them, with maximum efficiency and trust.

Q: Is this just an SMB type offering then?
A: I think that’s the clearest market for it, but it certainly need not be that way. Even the largest IT shops will need to call on specialists in particular software packages; while it’s highly unlikely that JP Morgan or Wal-Mart is going to need to call on outside assistance for anything as trivial as Apache configuration, they might for something like DTrace. I had the opportunity to speak with a DTrace specialist at Sun’s Analyst Event who did brisk trade with large businesses, and can certainly envision those types of markets being served via this offering.

I think SMBs would be likely to leverage such an offering differently – and possibly more heavily – than their larger cousins, but I don’t think it’s out of the question that there’s market opportunity up and down the business size continuum.

Q: Ok, I think I’m starting to see the problem. What’s the technical approach to solve it, then?
A: Well, I tend to describe it as consisting of a couple of different component pieces: pieces:

  1. Application Network:
    An application repository, like those that Linux and Solaris offer today (and in all likelihood, based on those existing networks), that includes both a community maintained library of free and open source components, and a second, optional, library that houses commercial applications which may or may note be open source, and may or may not be free (as in beer). Applications should be easily installed via the available interface.

  2. Developer (or SI) Network:
    The equivalent of the application repository, but for individual developers, who register for free. The resources are matched directly to applications in the repository.

  3. Ratings Network:
    Borrowing from the likes of Amazon and eBay, the network incorporates and applies reviews for applications and ratings for individual developers and – importantly – their customers.

  4. Support Network:
    Support is stratified into commercial and non-commercial tiers and tied, again, to the application. For commercial support, the network allows for easy subscription to the various support options offered by the likes of Covalent, MySQL and more. For those applications that do not currently have commercial support options, options are a.) to list individual developers, b.) to link to support wikis, or c.) to post requests for support.

Q: How does this make money?
A: There are several potential revenue angles here. In my mind, there are two things that absolutely need to be free of charge: application and developer listings. Without sufficient critical mass in both areas, this type of offering would be DOA. Eliminating barriers to entry in those areas, therefore, is imperative, meaning that listing has to be free.

Beyond that, there would seem to me to be ample opportunities to benefit from the individual transactions, whether it would be on a commission or subscription basis. Likewise, commercial providers of resources or support could potentially be monetized at signup, unlike individuals.

Would I, as RedMonk, pay a reasonable subscription fee to get access to a repository of prequalified resources for all of the little problems that come my way? Certainly. The target market here is those willing to trade money – in reasonable denominations – for time.

Q: How does this differ from existing offerings, either job boards or freelance pages?
A: It remains to be seen, but I think in three important ways:

  1. The resources are tightly married to applications at an operating system level so that everything’s in one place. I can install an application and pick a resource to configure and manage it simply and easily, all in the same place. Think “apt-get install asterisk asterisk-developer-from-denver”

  2. The offering extends the developer – employer relationship by incorporating feedback and trust mechanisms
  3. The offering bundles in extras like commercial support

Q: Who’d be interested in this type of service?
A: From the exploratory conversations I’ve had, lots of people. But more specifically, I think there are potential wins all the way around.

  • Customers: get access to a service that potentially makes their lives easier

  • Developers & SIs: get access to a potentially significant new channel or market, and make the job of marketing themselves much simpler
  • ISVs: get access to a new market, and better connections to their implementer community
  • Platform Providers: get to deepend their relationship with customers, and simplify the support sales workflow

Q: Who loses here – who would be threatened?
A: Some have contended that such a marketplace would represent a clear and present danger to SIs, but I don’t agree. Amazon and eBay, after all, have not obliterated existing retail businesses. They impact them, certainly, but also can serve as an important new channel. I really don’t see any preordained “losers” here, at least generally speaking. Ok, maybe straight IT body shops, but I’ve never been a fan of those anyway.

Q: This seems like a fairly massive effort, on the face of it, and the difficulty with network offerings like this is that they’re heavily dependent on critical mass. If customers show up and can’t find developers for package x or package y, how likely would they be to come back?
A: That is indeed the problem with network offerings, and frankly there’s no perfect answer. My recommendation would be to start small, maybe focusing initially on, say, LAMP+Ruby and some of the more popular open source applications like WordPress and ensuring they’re well stocked with devs you’ve recruited ahead of time before the doors are wide open.

This is a problem, but a solvable problem as entities from eBay to Wikipedia have proven.

Q: Who could you see building this?
A: Lots of different entities, actually. The most logical choices would be operating system providers like Microsoft, Red Hat, Sun, SuSE, and Ubuntu, but it’d take some community building savvy that not all of the above providers have. Large job board communities might want to play here, but they’d need to establish ties back to technical communities. Likewise, large ISV/SIs like IBM would also be potential players. Dark horse candidates might be eBay themselves or significant ISV packagers like MySQL.

As I’ve told everyone I’ve described this to, however, I don’t really care who builds it – I just want to buy it.

Q: This could be a good idea, why give it away?
A: Several reasons. I was talking with a very well known technologist the other day who lamented the fact that he had more ideas that he could execute on, and that’s the basic problem here. This type of application would almost certainly have to be built on top of existing businesses, and I have no intention of joining them to do that – I’m happy where I am.

I’m also interested in seeing this get built sooner rather than later, because we’d be the first customer. So it’s in my best interests to farm this idea out in case someone has the resources to run with it – we don’t.

But more importantly, I think that if someone built it it could be very good for the developer community, and it’s my job to give back to that group as much as possible. If this gets built and helps them, I’m a happy camper.

And hey, if someone wants to consult with RedMonk to build it, our door is open. But if you build it, I’d be content with credit and attribution – and maybe some discounts on the service ;)

Q: Anything else you’d like to add?
A: Just that I’m very curious to see what you all think of the idea. I’ve had a lot of conversations on the subject – more positive than not – but would appreciate any feedback you might have. For the developers in the audience, if this existed, would you sign up?

Categories: Marketplaces, Network.

  • http://www.joncollins.net Jon Collins

    Hi Steve,

    I *think* I understand what you’re wanting, but I’m not absolutely sure – essentially you’d like a pre-configured online environment based on freely available components, that woudl enable you to set up a business? Am I right or have I missed the mark?

    Not that I’m offering, it sure does sound interesting though :)

    Cheers, Jon

  • http://www.redmonk.com/sogrady stephen o’grady

    Jon: close. most package management systems will take care of at least basic configuration tasks. the problem is rather in extending it. so, for example, installing Apache might be straightforward, but how do you set up other domains? subdomains? and so on.

    it’s basically taking the existing applications and customizing them so as to make them useful to less skilled technical end users.

    and as for the free, that’s not strictly necessary. i do expect – like many of the infrastructures i see these days – much of the infrastructure to be free & open source. but i also have no problem with plugging in commercial components at any point, if that’s what customers want.

  • http://www.joncollins.net Jon Collins

    OK makes sense ta!

  • rektide

    From a customer perspective, what problem does this solve?

    Its not a customer problem. Its a community problem. Customers just get to suffer for worshipping their false idols^H^H^H^Hsilos of convenience until someone builds a real software platform, a, as we both are fond of saying, ecosystem. In spite of XML, in spite of REST, in spite of WS-*, in spite of a dozen sweet arse languages & platforms, software has remained totally insular to itself with perhaps the exception of whatever holes the framework it was built around exposes. We need better paradigms, ones that have the potential to rule them all and capture real network externalities and interactivities.

    It probably seems like I’m talking about something else entirely. I just think the system of code development we have now is not condusive to the sort of endeavour you are describing, that there’s a strong re-paradigming required before such social software dynamics can emerge. The ecosystem needs to start at the lowest level, the code, not the social constructs we build around the code. And code itself? Its barely changed in the past 18 years.

    I think Spring is the WRONG direction. We need to bind configuration much closer to hte program. We just need our programs to be much more configurable, much more scriptable( and much more interscripting). Thus instead of learning yet another .conf, you can throw some bits on a live sandbox and watch what happens. See: Smalltalk. Execute on your sea of ideas, imminently.

    /rant