Eyeballs and Downloads: Déjà Vu All Over Again

Analysis

Eyeballs and Downloads: Déjà Vu All Over Again

By: Robin Bloor, Chief Research Officer, Bloor Research
Published: 21st February 2005
Copyright Bloor Research © 2005

You may remember that at the height of the dot com boom, various dot com start-ups were claiming large volumes of web site visitors (eyeballs), and the figures they gave, which were probably accurate at least within reason, supported ridiculous stock prices – until it eventually became clear that, eyeballs or no eyeballs, these companies weren't selling much. Then, all of a sudden, the apparently forgotten link between revenues and commercial viability was re-discovered and sanity returned.

There is now another phenomenon of this kind emerging that maybe ought to be strangled at birth, or at least regulated, before it grows into a similar kind of beast. I am now being repeatedly informed of Open Source products which are clearly destined for success because they have achieved x million downloads.

Clearly, a downloader is no more a user than an eyeballer is a buyer. In the dot com era the eyeball statistics were at least likely to be genuine because they were being reported to the market (as a substitute for reporting real revenues) and there was also a tenuous link in that eyeballs really did mean business for some of the dot com wunderkind (Amazon, Ebay, Yahoo, Google, etc.).

The download statistics may also be a useful indicator in respect of some Open Source products. For example, Mozilla Firefox is now reporting having provided 25 million downloads – a quarter of a million per day. I doubt if that translates exactly into 25 million current users, but it is probably close to the mark. Some will try it and revert to IE (or their chosen alternative) for one reason or another. However some will also try it, use it and deploy it to multiple PCs from a single download. The download figure here almost certainly indicates a very high level of take-up but gives no accurate guide of the real level of usage.

However, I have also heard download figures that I simply don't believe - for example, an obscure Open Source database claiming millions of downloads. Open Source operations and commercial companies masquerading as Open Source operations are under no obligations to have their download figures vetted or even to state how they arrived at them. Some will certainly be motivated by whatever means are available to create the impression that take-up is massive - i.e. some will tell complete lies.

At the moment, SourceForge.net (a good source of reliable information in my experience) reports its registration of over 95,000 Open Source projects. Think what this means. In almost all areas of software development now there will be multiple Open Source projects with identical or strongly related goals. There are, for example, over 300 open source database projects. There are not going to be 300 successful Open Source database products. It won't happen.

In any event, Download figures should come with a health warning and should also be provided with some proof of reliability - and a disclaimer. Otherwise it will be eyeballs all over again.

Reader Comments

Posted: 21st February 2005 | By Christof Wittig :

Sir--
you might want to take into account that the pioneers of open source databases, Berkeley DB (Sleepycat) and MySQL, do not only have many downloads aka eyeballs, but are profitable companies.

I agree that there won't be 300 open source databases and the professional community should focus on those that are vendor backed.
We at db4objects have released the open source object database db4o for Java and .NET (www.db4o.com) and we've instantly seen 10,000 downloads per month. A recent survey has shown that virtually all of them are software engineers (i.e., people who make their living writing software) and 40% intend to use our product for a commercial project.

So it's time to replace ignorance or disbelief with trying to understand why users and customers like open source products: Because open source produces better software, is a more cost-effective way to produce and distribute products and provides more choice in the IT landscape.

--Christof Wittig
CEO db4objects

Posted: 22nd February 2005 | By Rick. :

I for one downloaded Mozilla Firefox twice (two locations) and then went to install it on fifteen system. I guess I kind of fly in your face on this issue don't I. How many others have done the same as me?

Sure, some people will download it and not use it. Others will do what I have done. Perhaps the number of downloads greatly underestimates the number of installs?

Posted: 22nd February 2005 | By Jean-Marc Liotier :

If you need Firefox usage data, there is a reliable source : your HTTP server user agents logs.

Posted: 22nd February 2005 | By anonymous :

Yep, usage data is a far more accurate barometer of browser share than DL numbers. However, even here it's tough to quantify because so many sites that track their usage religiously tend to be sites that early adopters use. Still, the trend that we've been seeing for over a year is steady erosion of IE's share; most of which is going to Firefox. My best guess from looking at a variety of these is that IE's share has probably dropped to around 88% and is still falling.

It's too bad that Google pulled that statistics page of theirs down. It was probably the most accurate barometer for actual browser usage that was out there.

The messages above were all contributed by IT-Director.com readers. Whilst we take care to remove any posts deemed inappropriate, we can take no responsibility for these comments. If you would like a comment removed please contact our editorial team.

We automatically stop accepting comments 180 days after a post is published. If you would like to know more about this subject, please contact us and we'll try to help.