No doubt, it’s sometimes difficult to measure Ubuntu’s business success. Canonical says Ubuntu now has about 12 million Linux users — up from an estimated 8 million users in 2007. That sounds impressive, but what does that really mean in terms of Ubuntu’s market penetration, Canonical’s march to profitability and Ubuntu’s progress as a Linux distribution? In my mind, there are additional data points Canonical could share to further show Ubuntu’s momentum. Here they are.
On the one hand, I want to give Canonical credit: 12 million estimated users is a big figure. And it provides a foundation upon which Canonical can promote additional services — Landscape, Ubuntu One, Ubuntu Music Store, etc. — to consumers and business users.
But just how well is Canonical doing? In March 2010, new Canonical CEO Jane Silber said the company had about 320 empl0yees and was on a path to profitability. But she conceded the company wasn’t yet profitable.
As a privately held software company, Canonical certainly doesn’t have to say much about its financial performance. But frankly, I think Canonical can share a bit more information without having to completely open its books.
A few ideas…
1. Time to Phone Home: In a LinuxPlanet article, Canonical VP Chris Kenyon conceded that the company had has no registration or “phone home” process for Ubuntu, so deployment and user numbers are always a guestimate. In my mind, it’s time for Canonical to more closely track Ubuntu. Find a way to maintain customer privacy but begin to have Ubuntu phone home. The data points will show Canonical key regional and international trends that could help future business development.
2. Change the Landscape: Canonical offers Landscape — a tool for remotely managing Ubuntu systems — in SaaS and on-premises configurations. For the SaaS version, Canonical should start disclosing just how many Ubuntu systems are being managed worldwide. If it’s too early to share those numbers — after all, Landscape is just getting started — then simply share year-over-year growth percentages for Landscape-managed systems.
The idea here is to push beyond the consumer desktop mindset. By sharing some Landscape management stats — even rough percentage growth rates — Canonical can more effectively position Ubuntu as a business solution for notebooks, desktops and servers.
3. Open Up Ubuntu One: I realize Canonical may not want to disclose how many people are paying for the storage and file sync service. But there are other ways to show growth… For instance, percentage growth (quarter-over-quarter) for the amount of storage Ubuntu One handles. And soon, some stats for the Ubuntu Music Store would be a solid step in the right direction, too.
Reality Check
Ultimately, Canonical faces many familiar challenges that we frequently hear about in the open source market. Other than Novell and Red Hat, it’s difficult to find publicly held companies that disclose in-depth information open source revenues.
But there are ways to show progress. Many open source companies will disclose how many “paying customers” they have. Others will disclose year-over-year revenue growth rates.
In the months ahead, I hope Canonical takes similar steps. The effort could help to silence Canonical’s critics while also giving customers and partners peace of mind as they consider more Ubuntu-oriented business efforts.
Read More About This Topic
Share This Post
Tags: Canonical Landscape
Interact: Add a Comment | Trackback Link | Permalink
Subscribe: RSS Feed
Don't miss Charlene O'Hanlon's weekly columns...
I think #1 on your list is obvious and should have been done from the start. It amazes me how the radical wing of the open source crowd are so terrified of something as harmless and useful as a head count.
I WILL NOT give Canonical credit for that 12 million estimate until they articulate how they come up with that number. If they can’t publicly explain how they get the estimate, then its a worthless PR stunt. To silence me on this particular issue they need to publish a methodology on how they create that userbase number.
I continue to find it very suspicious that how the number is generated is held as a closely guarded business secret and not something subject to the ideals or radical transparency that Matt Zimmerman recently claimed is Ubuntu is founded on in a talk he gave.
You know what else Canonical could do? Publish the full report from the Ubuntu Server survey. The survey is a wealth of information. For example of out the ~3000 respondents roughly 100 of them have reported they are using Landscape. Is that good or bad news? I’ll let you judge.
What else does the survey say about interest in Canonical’s products? Canonical won’t publish the full survey report. Again… is this living up to the ideals of radical transparency when it comes to interacting with the Ubuntu community? They asked the community to respond, they even published the questions on the wiki prior to the survey..but then the survey results are held privately and Canonical cherry picks the few questions the could spin positively and releases them.
-jef
Josh, Jef: Thanks for weighing in.
Jef: I know you’re suspicious of the 12 million estimate. I tend to give vendors the benefit of the doubt, so we published the 12 million figure but continue to push for more details about how Canonical is performing as a business. I will be curious to see how Canonical shares business info amid Matt Asay joining the company and Jane Silber’s move to the CEO slot.
-jp
How about counting every time an update package in the main repository has been downloaded from the update servers? Granted some people have several Ubuntu powered devices so that would push up the numbers, but some users never update either so that could compensate. It would be approximate but better than nothing. Or maybe that’s how Canonical came up with 12 million already.
@jef If you had to pluck a value totally out of the air for “number of Ubuntu installs”, what would it be?
@Clem We can’t count hits on update servers because we have numerous mirrors around the world which are not controlled by Canonical. We have no visibility on the number of hits to those mirrors.
There are other things Canonical could do though. Such as look at the logs for the time servers given ntp.ubuntu.com is the default. This is also flawed because some people will have multiple installs behind one IP, and others wont allow NTP through the firewall.
We can of course know how many Dell laptops, Litl’s (if they count), and Sharp PC-Z1′s shipped with Ubuntu pre-installed. It’s hard to know how many people install themselves.
Anecdotally I find it staggering how many people outside of the circle of Ubuntu community I’m in, that are already running Ubuntu. I often find myself not needing to advocate the use of Ubuntu to new people I meet because they already run it!
People even walk up to me in the street and talk to me in stores and tell me they run it, and that’s without any prompting, I just happen to be wearing Ubuntu branded clothing.
One guy in my local corner shop asked me what the bird on my t-shirt was (the hard heron one) because he’d seen it somewhere before. “Where have I seen that?” to which I replied “on your desktop?” as a lucky guess, and I was right.
On another occasion I was walking through London past a large trade show venue when someone shouted over to me “Hey! Ubuntu!”. I went and talked to him. He said that he ran all the lighting and projection systems for the Ski/Snow show in the venue. All the boxes running that setup were running Ubuntu.
Final anecdote, I was on a train on the way to an Ubuntu release party in London on the day of the Hardy relese when two guys got on and started talking about “upgrading to hardy”. I asked if they were heading out to the release party and they said “nah, we’re not involved in any of that, we’re just two guys who happen to use Ubuntu”.
So it’s my opinion from purely first hand anecdotal evidence that there are significantly more people out there running Ubuntu than you might think. Maybe even more than 12 million.
What do you folks make of the claims that Fedora has many more users than Ubuntu? The Tux Radar podcast mentioned figures that would show Fedora has twice as many users. In 2008, Red Hat claimed Fedora had ~2 million more users.
Alan:
I don’t pluck numbers out of thin air. I care deeply about getting an accurate metric for linux usage. I’m not going to make up numbers as part of some speculation excecise.
And clearly Canonical does not make up numbers out of thin air either. Kenyon is quoted recently as saying there is a methodology with these estimates. All I’m asking is that Canonical live up to the self-affirmed concepts of radical transparency that governance in the Ubuntu project is meant to follow and publish what that counting methodology is.
Are you going to defend the fact that Canonical execs have been issuing Ubuntu userbase estimates to the press (not to the community…but directly to the press) for 4+ years and have never explained how the count is established? If the methodology is valid, then there’s no downside in publishing it. If the methodology is not valid, then Canonical is doing a disservice to the entire Ubuntu community by publishing numbers based on an invalid methodology. It’s past time for the counting methodology to be published.
-jef
@jef Ok, rather than pluck a number out of the air, how about you have an educated guess? You can’t tell me you never estimate anything? Like how much pepper to put in your food or what proportion of fuel is left in your car? I find that hard to believe.
I can imagine multiple reasons why Canonical wouldn’t want to reveal how they get their userbase estimates. I suspect that in a spirit of openness if they told the userbase “We looked at how many unique IPs hit the time server” then people such as yourself would jump on them from a massive height claiming a privacy invasion.
They can put a module in ubiquity asking users to voluntarily register their name or at least give them a web address if they would like to register later.
Gee, I can’t imagine why anyone would ever be suspicious of data collection by any third party, and especially a for-profit company. Could it have something to do with decades of rampant privacy invasions, collection of personal data, and the mining, buying and selling of all that data? Without disclosure, or with limited disclosure and a whole lot of lying? Really, it is a mystery why anyone would be upset after decades of abuse at the merest suggestion of yet more of it. Silly privacy advocates, don’t we understand that if we have nothing to hide we have nothing to fear?
Measurement is integral to the management of organizations, projects, processes, and products. “What cannot be measured, cannot be managed” and “What cannot be managed, cannot be improved”.
Alan:
First of all, to be as clear as I can be on the matter. I don’t think counting ip addresses is an invasion of privacy as long as care is taken to not reveal the ip addresses as part of the data products. It is possible to build and publish analysis scripts to harvest ip addresses from server logs that do not leak ip addresses. This is a very common practise. Many sites uses packages like awstats to do exactly this sort of counting.
As long as Canonical publishes the scripts for me to review, I will publicly defend them against allegations against invasion of privacy as long as I can confirm for myself the analysis scripts don’t leak ip address information in their output.
Alan, trying to hang me in a noose of terminology semantics is going to get quite boring. I don’t care what you call it. You want to call it an educated guess, fine. The question remains, what is the methodology Canonical is using to come up with the numbers they have made public repeatedly in the press? What input is being used? What algorithm is being applied? I don’t know. You clearly don’t know either.
-jef
@Jef Why so defensive? I’m not trying to trip you up with linguistic games. Heaven knows I don’t have the time to do that.
I’m merely asking a simple question. I fail to see why a simple finger in the air estimate is beyond you.
Of course I don’t know what algorithm Canonical uses, I’m not party to that kind of information. I don’t have a problem discussing and speculating though as you seem to.
Jef Spaleta Said: “What input is being used? What algorithm is being applied? I don’t know. You clearly don’t know either.”
Now that’s unfair! Alan clearly stated that Canocial had an educated guess, didn’t he? You can’t tell me you never estimated anything, can you? Today, for example, I estimated how awesome I am. On a scale from 1-10, I thought: “OMG, 11!”. When asked, my friend replied with the same “OMG”, but adding “-11!”. So, how awesome am I? I don’t know, he clearly doesn’t either. But he taught me educated guesses are not scientific algorithms or anything, and must have an intrinsic relativistic factor.
So, you asked, and there you are: Canonical counts us like “how much pepper to put in your food, or what proportion of fuel is left in your car”. Finally, I find that comforting…
Alan:
Yes such an estimate is in fact beyond me.
-jef
What amazes me is all the articles and comments that came up when someone mentions a number about what could be Ubuntu’s user base. Unless you are a competitor or someone looking for professional services why should you even care. Has it not even been considered that the size of the user base does not equate to the best operating system? In the grand scheme of things 12 million is not that large of a number. I use Ubuntu and I don’t care how many people use Fedora, Mandriva, OpenSuse, or any of the others. Distributions with corporate support are the only ones who should care. The Linux community is too competitive when it comes to numbers. The service and support business model for corporate sponsored distributions are the only ones that count. Desktop user numbers are irrelevant.
It doesn’t matter in the least how many users there are.
If you like Ubuntu and contribute your time and a few dollars more dollars to the project than you consume snd you appreciate ubuntu. Canonical will continue and grow.
They may become a publicly traded company but ownership seems secondary to the stated goal.
The strictures of commerce are unavoidable but not entirely desirable.
There is something more to be gained by freely working together and perfecting the tools used to exchange ideas.
Cheers,