A Compliment for the Linux Adoption Curve!

My readers have been telling me that they’ve missed my blog entries where I look into some of the ideas around why Free and Open Source systems fail to gain traction. Today I’ll be conjuring that almost trope, the ‘Chasm’ adoption curve.

Basically the curve describes how any new technology must cross a barren dessert called the chasm from 10% to 20% of market share in order to gain enough share for the adoption to have enough momentum to go on and conquer the whole market.1 Often in the FreeDesktop ecosystem we see our adoption curve being really, painfully unable to push adoption past 10% of any slice of the market.

The problem we have, I think, is that we fail to create enough compatibility with our “compliments”; also known as all those requirements for tasks we want to do with our computers. I wish to use this to illustrate on one hand a rationale for why pushes into the mainstream fail and why I think peripheral hardware compliments should be a priority for all FreeDesktop programmers.

Every advocate knows that it’s easier to get people who are determined to adopt a Linux based FreeDesktop system than it is to get a windows expert to adopt one. For simple reasons, a self determined user will either make the sacrifices or make the investments to get his compliments compatible. This determination can take the form of either programming upstream new drivers, creating new applications or even just using toxic workarounds to fix an issue that causes things to not work on a fresh install.

Advocates will also often tell you about how successful they’ve been in getting their grandmothers using their FreeDesktop distribution. In fact many of us suspect that Ubuntu and similar distributions are very ready for typical technological Laggards, more than we are ready for early adopters. I think this has much to do with Laggard’s low investment in compliments and subsequent low exceptions about what computers can do for them.

This illustration (right) attempts to show the people to whom our software can be used as an acceptable replacement. In order to improve this, we’d need to either a) improve the attractiveness of the platform to encourage sacrifice of compliments or b) systematically increase compatibility of compliments.

The job of Unity in the new Ubuntu system is to improve attractiveness. An important attribute for sure. But many cycles has gone with a failure to improve compatibility with hardware compliments and this has shown that the gamble for Canonical is that they can improve the attractiveness to such a degree that the sustained investment into compatibility will come from the hardware vendors themselves.

I believe this is a mis-calculation. The hardware vendors will only invest in our ecosystem, when we are attractive compliments to their products. But they aren’t going to invest in their old discontinued products, but only into their new products. This leaves the old products without support and it just so happens that a great number of our main-stream users have made investments into hardware and are not willing to simply buy new hardware just yet. in conclusion, I think we can count on hardware makers providing us with drivers eventually; but for as long as they are not, we should be investing in all their old product lines and making sure they work with our desktop distributions.

This is why I believe it is important that Red Hat and Canonical stop playing around and put money directly into hardware peripheral device support. Printers, scanners, drawing tablets and even phone syncing. Everything that would improve our ability to attract new users over the chasm, by removing the things they would have to sacrifice in order to join us.

What do you think?

1 For a given sense of market, markets can be sliced and diced into different metrics and general purpose computers can be diced quite a fair few ways. For instance the programmers market is fairly healthy.

13 thoughts on “A Compliment for the Linux Adoption Curve!

  1. Excellent graph! I consider myself an early adopter that just likes being a little different. I’ve performed much Googling to get the features I miss from Windows. The rest of my family wouldn’t touch Ubuntu without my efforts. I should mention this is from a home and really hobbyist perspective. There isn’t a chance in the world I could make use of Ubuntu in my profession. Not necessarily due a lack of effort from Ubuntu or Linux, but my tools are very proprietary and very Windows embedded. It would be difficult to do my job with a Mac as well. Not ready for a career change just yet.

  2. John – All those bits of software you speak of are also compliments. Getting those to either be compatible or be made linux native is important. But of course it works better with software because software always updates and the investment continues past the sale of the product (generally speaking)

  3. One of the items I have noticed in “chasm” products is what gets to the later stages is rarely what people see in the early stages.. and early stage people are always thinking that what they wanted never occurred. [The old “where is my flying car?” cry]

    I personally think that the year of the Linux desktop has already occurred when more android items were shipped than windows and iphones. Was it gnome or KDE.. no but it is Linux and a ‘desktop’ in a way that late adopters and laggards will use.

    Computer/laptop desktops are always going to be a minority product compared to the 100 times more items that will have Linux in them but no one will know. The “desktop” will be there and by the time desktops move to it.. we will have our flying car.

  4. I just attended a training where they talked about the adoption curve and “the Chasm”.
    Do you know how Apple jumped the chasm after the second coming of Steve Jobs? Product placement. They plastered the colourful plastic iMac all over Hollywood, and people started to see them as an alternative, not as a niche any more.

    Hardware support is difficult to get when vendors don’t disclose the details of their software. IMHO it’s better to do what Apple did *with regards to the hardware* (not the software, what seems to be the trend today 😉 ), and release computers built with only hardware that works with free drivers and a lit-up circle of friends on the lid.

    This will not do anything for printers and phone sync (but Android removed the need of a “PC Suite”, at least for me), but cross the chasm and the vendors will come.


  5. This compare is really incomplete.

    Android with its growth in phones is expanding hardware support.

    splashtop usage by a lot of makers is also expanding hardware support.

    We also have editshare like groups going after video editing software.

    This is a team effort. Each bit joins and helps.

    The problem is hurding of the cats to produce something useful

  6. IMO the original diffusion of innovations concept (http://en.wikipedia.org/wiki/Diffusion_of_innovations) is even more fitting to the situation with ubuntu. From that wikipedia page “Rogers’ 5 factors” section lists some points for users deciding whether they adopt an “innovation” (ubuntu in this case): relative advantage, compatability (to what the user is used to), complexity/simplicity, trialability and observavility.

    What your original post talks about falls most directly to complexity/simplicity part in the sense that as long as “everything just works” ubuntu is pretty simple and goes to very complex when it doesn’t. Trialability and observability I think are fine, Live CDs are easy to test and Ubuntu is quite visible and has a clear identity.

    Now relative advantage is much more difficult to prove, mostly windows/osx/whatever came as factory installed works and there is not that much reason to change. Compatability (to user expectations) is something that Unity and related changes are not helping in the short term, because the environment becomes less familiar compared to everything users are used to, including previous versions of Ubuntu. On the long term who knows, maybe it will bring some relative advangates…

    I am not sure if I had a point, but I was familiar with the concept of diffusion of innovations and the graph stuck my eye 🙂

  7. Since we do currently not even have a satisfactory driver situation concerning fundamental hardware like graphics cards I fail to see why users should adopt Linux as an OS.

    Unfortunately the Linux community and the main contributors fail since 10 years in getting in a grown-up communication with Nvidia and ATI(AMD) to provide basic hardware support for users who do not run headless servers.

    So I do not think scanner support and phone sync will make a big difference…

  8. Sun Tzu – I seem to have had no problem running Ubuntu on almost every machine I’ve ever come across when it comes to graphics. There are issues, but not nearly as bad as you say.

  9. How did Microsoft conquer the desktop market ? Not by creating a better OS with more hardware support, but simply by contracting with IBM so that every PC sold would have Windows embedded (and paid for). Quite unfair and unethical given the lack of choice, but very effective…

    It seems Google recently contracted with phone vendors to do the same with Android.

    In an top-down ecosystem where big companies decide what OS you have to use, how can Ubuntu or other Linux distros compete ?

    I think the only way for GNU/Linux to gain momentum on the desktop would be for companies like Canonical to contract with PC manufacturers on a large scale and do the same as Microsoft did. Ship on a lot of PCs, mobile phones etc. If Canonical or other companies could do it without losing all ethics, the job would be done. Is it possible ? Desirable ? I don’t have the answer…

  10. @Jack – The problem is that the situation is very different, Android is 100% compatible and 100% of expectations because they enter into an empty market. Plus the attractiveness is massive and you can sweep along a whole bunch of manufacturers. Microsoft were starting with a clean market with virtually no competitors, it’s a very different situation to what Canonical must do with OEMs, which is to convince them to change from something existing to something new which is a risk.

  11. The problem with Linux phones is people don’t necessarily know they’re Linux phones. I wonder if universal branding would help get people to make the jump, or at least show interest in a Linux desktop. Showing Tux on all Linux devices/distros might raise awareness that there’s more than Windows and OSX available on the desktop. The realization wouldn’t happen overnight, but having a universal brand that people could identify with should push things along.

  12. what an interesting read.

    a technology really has a good slice of the market when most of the users don’t actually know (or even care) about what technology they are using.

  13. @ Sun Tzu:

    NVIDIA has a perfect relationship with Linux, their drivers are up-to-date and their QA is in many regards now superior to that of Free drivers. The only problem is their drivers are closed source. In terms of support for total number of NVIDIA cards, Linux is ahead of Windows Vista/7 – where more of the older cards are supported. In terms of newer cards, Windows and Linux are equal.

    Nouveau aims to fix the problem of closed source as the only option by offering reverse-engineered drivers, in very little time they’ve approached the performance of the proprietary NVIDIA drivers. Given another few years, performance will exceed the Windows NVIDIA drivers.

    ATI had a rocky relationship with Linux originally, with poor quality fglrx drivers and no intention of supporting any open source efforts. When AMD purchased ATI, this changed. Now ATI still supply fglrx, but also provide documentation and sponsor developers to work on Free drivers.

    Today, fglrx is still nowhere near the quality of NVIDIA’s proprietary drivers, but the same can be said with Windows, where XP had 10MB of files dedicated to patching Windows XP to work with ATI cards.

    Intel graphics is perfectly fine, both in 2D and 3D. Nothing to say much there.

    So, your opinion comes from outdated knowledge methinks.

Comments are closed.