Targeting Early Adopters - Say No to Windows

Posted 13 May 2002 at 15:12 UTC by garym Share This

Windows may be the right platform for applications pre-installed on end-user machines, but for revolutionary services, it's time to rethink our assumptions about which Internet population to target.

(originally published May 1, 2001 on osopinion)

As a contract consultant, I'm often asked to create new services to deploy over the Internet, and in many cases these services require some sort of end-user component. Whether the goal is browsing stock quotes or for the latest peer-to-peer gizmo, almost universally we're asked to create a version that only serves the Windows community.

Why Windows? The reason most cited is market share, with some strangely accurate figure that "93 percent of the Internet population runs Windows." While that 2-digit figure seems way too accurate for statistical credibility, I'm also wondering if it isn't based in innumeracy.

Quite apart from the plain, old-fashioned innumeracy of false-positives, what's lacking in this number is any qualification of who those people are or how the numbers were collected. For example, a large news portal site may find a very high percent of Windows browser agents, but this may be expected. A Windows audience may not be comfortable finding their news direct from multiple sources using RSS or Usenet. The population from such a source is as self-selecting as finding a high-degree of UNIX browsers at SlashDot or SegFault.

Less is More, More or Less

But let's imagine there's some grain of truth, and we'll round that to 90 percent (+/- 3 percent), leaving the rest of the population as split between Mac and Unix. This still seems incredibly low, but let's err on the side of the accusers. That leaves roughly 10 percent for Mac, UNIX and OS/2, and the other "fringe" operating systems.

Now, let's consider our initial question: Where is our target audience for some new and cutting-edge Internet service? This is where the unqualified statistic may be helping to drive so many otherwise-clever dot-coms to extinction.

In my own experience, Windows tends to be the first choice of first-time computer users, easily 90 percent of the time, probably more. These are people like my mother, who would never dream of adding software to her machine, or my younger brother, who is far too busy being a lawyer to want to fiddle with new technology.

It describes my neighbors who use e-mail and do a bit of online shopping, and it describes those office workers who hadn't thought about collaborative groupware until they walked in and found it waiting for them. It also describes only a small precious few developers or tinkerers I've known.

Finding the Bolder Buyers

So I wonder, if 90 percent of that 90 percent are a staid, conservative and timid user base, that leaves only 8 percent who are not. Those 8 percent are faced with a diversity of services deployed for the full 90 percent. So, it is virtually impossible to target them through one media channel.

If we are very, very lucky, we might reach 10 percent of them to tell them about our new cutting-edge service, which means we're down to 0.8 percent as potential customers. The other platforms (Linux, UNIX, Mac and even OS/2) on the other hand, are early-adopters and rebels almost by definition. Of our 10 percent population, we might get 80 percent bold enough to try our new kit.

Doing the Math

Having been marginalized by innumeracy for the past five years, these rebels are also hungry for any sort of service that improves their situation. Since the diversity of their service pool is a corresponding 10 percent that is split among all of them -- Mac users (OS/X aside) are unlikely to know or care about Linux services -- we are left with a miniscule set of communications channels which can reach extremely high community coverage. Those channels include advogadro, osOpinion, SlashDot and LWN.

This might leave us with a sales channel capable of reaching over 5 percent of the entire Internet early-adopter population, or 7 times the audience for the Windows-only service.

The Experiment

This is, of course, all musings and conjecture, and my figures are highly suspect.

Still, asking these questions may have some merit. We recently ran an experiment, which suggests taking a closer look at how we reach new audiences with new technology. In this experiment, we deployed a preview release for a new technology from one of our clients, a P2P application for downloading large files.

Our first trial placed this new technology before a general Internet audience on a large news site. The response was approximately zero.

The second trial was on a popular developers Weblog site, predicated on an assumption that many developers would be early adopters, and developers may be more likely to support other developers by trying pre-release technology.

While the response was noticeable, half came from unsupported Linux browsers. Our third test was a small leak (reply to a thread) to core hackers of a Linux distro. The news spread like a wildfire, out from their mailing list and into the general community. And our site stats resembled driving west across the Midwest plains and ending in Colorado.

News Travels

Now, don't read too much into our experiment. There are many, many variables at work, which could account for the differences.

Still, the speed with which the news traveled from a mailing list to a major community hub was significant (within hours) and, in contrast to the largely apathetic response of the other communities, the Linux response was positive, instant and gracious.

Right on Target

Do we really want to target Windows? I think we need further investigation. It may be the platform to choose for staid and conservative applications destined for pre-installation on end-user machines, but for very new and revolutionary services, we may want to rethink some of our fundamental assumptions about the Internet population.

We also need to be especially careful to ask the right questions about who it is we are hunting. That "93 percent of 1 billion" sounds like an irresistible market, but if I were really hungry, I think I'd rather catch my fish in a small, quiet, well-stocked lake of easy trout than take my chances chasing barracudas on the open ocean.

What businesses are these?, posted 14 May 2002 at 13:11 UTC by brlewis » (Journeyer)

The businesses you're dealing with are doing so well financially that they don't care if they lose 7% (their number) of customers right off the bat? Who are they? What's their secret to success?

For other businesses, where every customer counts, making a web interface that's usable on multiple platforms is so easy as to be a no-brainer.

Interesting Article - I did not look at it that way, posted 14 May 2002 at 15:34 UTC by shlomif » (Master)

Very good article. Naturally, Linux people tend to be more of the "power user" types than Windows folks do. So sometimes ignoring their share, however relatively small can be a mistake.

An issue is that creating software that works equally well on both Windows and Linux is not trivial. So some people pre-assume that Windows is a better choice because it is more common.

I'm not sure if it qualifies as innumeracy, in the John Allen Paulos meaning of the word. It's more of not taking all factors into consideration.

Slashdot visitors mostly run Windows..., posted 17 May 2002 at 16:39 UTC by tapir » (Journeyer)

Before Slashdot was bought by and before the IPO, Slashdot used to publish daily figures on how many visitors it got and what browsers and platforms they were using.

They killed this feature during the run-up for the IPO because the results were embarassing: Over 90% of visitors to Slashdot were running Windows... The ratio didn't look terribly different from most other sites.

Slashdot isn't primarily a site where Linux fans get together, it's more of a place that Windows users hang out to bitch and fantasize that Linux is going to solve their problems for them. However, it just doesn't happen because they work for Dilbert's boss, Linux is too hard to install, and they think Linux is hard to use because they struggle with a half-baked GUI for hours to do what you can do easily with one typed command.

/. browsers, posted 19 May 2002 at 21:28 UTC by garym » (Master)

again, tapir, be wary of mixing science with myth-information: Could the slashdot showing have been at all because, until just a few weeks ago, the only browser which actually worked was the one bundled with Windows, or could it even be that some significant population of ./ participants steal those moments while at work, and I know from my own experience, that the shops I visit where Windows is the defacto desktop does not at all mean it is the preferred desktop (or the platform of the innovators and early adopters), it only means that this shop is constrained by MsOutlook.

An old .sig from NASA read: Absense of evidence is not the same as evidence of absense. Be wary of reading too much into any unscientific statistic, and question all those who claim to be science.

Multiple OSs, posted 21 May 2002 at 17:30 UTC by julesh » (Master)

I agree with the last poster. Most linux users (a) work in an office with a Windows desktop and (b) multiboot linux on their home machine. I do almost all of my web browsing from windows, because it works more smoothly, despite the fact that for other non-work related tasks I use linux about 40% of the time.

I would definitely concur with the article. Supporting non-Windows OSs is important for any new company coming in to the market with a new product. In case you missed it, there was a link on slashdot recently to an article about management of innovation. The guy who wrote it was basically saying that the only way to start a totally new product is to start with a niche market where it will be better received, and gradually assail the mainstream (radically oversimplified, but catches the gist...). So, for some new software product, I would definitely look at attracting Linux, MacOS and *any other, even less commonly used* OSs I can easily target. Because the users of these OSs aren't used to getting full support like this, and will become your best advocates when you try to break into the mainstream.

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!

Share this page