James Dixon’s Blog

James Dixon’s thoughts on commercial open source and open source business intelligence

The Analyst’s New Confusion

with 22 comments

Brian Prentice at Gartner has a new post titled Open Core: The Emporer’s New Clothes – click here to read it

I met Brian at OSBC recently and chatted with him for a while about several topics including patents and Microsoft’s open source strategy. While I like Brian, and we agree on many things, I (somewhat predictably) don’t agree with this latest post.

There are some important things that are missing from his piece.

For most proprietary software companies the biggest expenses are sales and marketing.

Let’s look at a proprietary competitor in Pentaho’s market – MicroStrategy (stock symbol MSTR). In their last quarterly report they state that

  • Revenue from Product Licenses $34.4m
  • Revenue from Support and Services $69.6m
  • Sales and Marketing (S&M) Expenses $31.5m
  • Research and Development (R&D) Expenses $11.4m
  • G&A Expenses $14.2m

MicroStrategy’s figures are fairly typical for an independent proprietary software vendor. You can see from these figures that R&D is the smallest expense, and S&M is the biggest. In fact the S&M expenses are more than all other expenses combined. You can see that revenue from licenses is about the same as the S&M budget. It takes almost as much money to convince people to buy the software as they get from the sales. The word ‘convince’ is important – the S&M teams have to convince companies that the software both does what they need, and is worth the cost. Each quarter the money made from selling licenses is poured into next quarter’s sales and marketing budget. It is therefore support and services where they make most of their money and all of their profit.

To re-iterate, because the S&M expenses are so high, MicroStrategy make no money selling software licenses, they make most of their money (and all of their profit) on support and services. From this perspective they sound very similar to the pure-play open source companies – who make no money selling software licenses, they make it all on support and services.

If you contrast the S&M strategy and budget of a proprietary software company with a commercial open source one you will see fundamental differences. The commercial open source company provides open source software. This software has to be be fully-functional and useful, otherwise their will be no adoption. The R&D costs of this software will be lower and, more importantly, development will be quicker than that of a proprietary company. But, as proved above, the R&D costs are the lowest costs anyway – lowering your lowest cost does not provide you with a significant benefit. The major difference is that the open source community edition of the software enables an efficient, cheap, in-bound marketing strategy instead of an expense, out-bound, enterprise marketing strategy. By switching to a primarily in-bound sales model a commercial open source company can drop its marketing budget dramatically. The reduction of the S&M budget allows the license fee to be eradicated. So the tasks and strategies of S&M in an open-core company is significantly different from those of a proprietary company.

These facts about the S&M activities of proprietary and open source companies seem to be largely ignored by many analysts and commentators:

  • Tarus Balog (CEO of The OpenNMS Group), says that open core vendors use open source as a marketing gimmick. As far as I know Tarus has no prior experience running a software company so I’ll forgive him his confusion. This is no gimmick, this is a fundamental part of the business model.
  • Matt Aslett (451) and Brian Prentice (Gartner) talk about code contributions and R&D practices but rarely mention S&M

I have worked my entire professional career in either proprietary software development or commercial open source software development. I have experience of the differences between these models first hand. I have only covered differences in the S&M domain. There are also numerous, significant differences in the development, QA, partnering, business development, support and service departments etc. Some of these I cover in part II of the Bees and the Trees

As I have said before, the software industry is in a transition from proprietary software, to some mix of proprietary and open source software. Is that eventual mix 0% proprietary and 100% open source? I don’t know. I do know this transition has been in progress for over a decade and has a long time (decades) to go. Maybe the open core model will only be relevant during this transitional period. But this transitional period is long, and many start-ups and existing proprietary companies have no viable way through this transition without the open core model.

Brian Prentice seems to be trying to find something to say about open source. Using invalid comparisons with ‘freemium’ and SMB strategies isn’t a good idea. Heavy-handed, one-sided statements, about an entire business model  – based on seemingly incomplete knowledge of that model, doesn’t seem good either. Companies like Gartner are struggling to show value in an increasingly-open world. Brian’s latest post, in my mind, doesn’t exactly help them do that.

Written by James

March 31, 2010 at 6:52 pm

22 Responses

Subscribe to comments with RSS.

  1. We met at OSBC for the the first time and in this post you resort to an ad hominem attack against me? How disappointing. To paraphrase Silent Bob in “Chasing Amy”, what you don’t know about me might just fill up the Grand Canyon.

    If we want to discuss business acumen, I should point out that your company had to resort to a fourth round of funding at $7 million, whereas I am certain my company paid more in taxes this year than you made in profit. (grin)

    You state that “commercial open source company provides open source software”. So by your definition Microsoft is an open source company. Google is an open source company. It’s so broad as to be meaningless.

    Brian was dead on, and the beauty of it is that more people are going to listen to him than ever listened to me … or you.

    Tarus Balog

    March 31, 2010 at 7:27 pm

    • Hi Tarus,

      It was nice to meet you at the Palace Hotel.

      I did not mean to offend you in any way.

      You have made many derogatory comments about our business model, and called companies like us ‘dishonest’. We also clearly disagree on lots of points, and I don’t think we will ever agree. So our interactions are destined to be mildly antagonistic.

      You understand the differences between the pure-play and open-core models (although I think you miss-represent them from time-to-time). I was trying to show that in open core model, the open source software is not a marketing gimmick, it is fundamental to the whole model. It is not a gimmick, not shareware, not bait-and-switch. None of these models work.

      Unless you have significant C-level experience in a proprietary software development company, you won’t know how different the open core model is from the proprietary one. From what I read on LinkedIn you don’t have that experience, so you are unlikely to understand the differences between these models. If LinkedIn is wrong, please correct me.

      I am not surprised that you don’t understand all of these differences. But I am somewhat shocked that Brian doesn’t.



      March 31, 2010 at 8:34 pm

  2. Hi James – Brian here.

    I think you missed the broader point of my blog post. The thing is that I don’t really care about the cost structures you face when running your software company. What I care about is whether your business model creates incremental value for your intended customers. Those are the people I spend most of my time speaking to. And those are the people who clearly understand the value that Gartner provides in an increasingly open-world. What they want from us is to put vendor claims to the test – and that’s what I did in my blog.

    Believe it or not James I actually do understand the dynamics around S&M costs. I spent a good part of my career in product management, marketing and sales roles at software vendors. Those managing software companies should be concerned about these costs. But the nature of those cost structures are far more complex and entrenched than advocates of open core models make out. And the jury is definitely out whether the open core approach will create a lasting impact. I personally think it won’t.

    But again James – so what? If you can lower your S&M costs you’ve increased your profit margin. Where’s the value to your clients? I think you should be spending more of your time blogging on that topic. What is about open core that makes companies like Pentaho uniquely valuable as a potential supplier. And there’s the rub – there isn’t. That doesn’t doesn’t detract from your company or product. It just means you’re like every other software company out there vying for people’s business. Therefore you should be treated by those potential customers no differently.

    Brian Prentice

    March 31, 2010 at 10:33 pm

  3. […] difference Matthew Aslett, April 1, 2010 @ 6:32 am ET Pentaho’s CTO James Dixon has been defending the open core licensing model against criticism from myself and Gartner’s Brian […]

  4. Hi James,

    As it happens we’ve just published a new report on sales and marketing for open source. We found that open source can enable significant savings in software sales and marketing, but it is often a case of spending differently rather than spending less. More details here: http://blogs.the451group.com/opensource/2010/04/01/sales-and-marketing-for-open-source-same-difference/


    April 1, 2010 at 12:13 pm

  5. James,

    Open source companies are really just a new way of doing more cost effective software start-ups. You are right the S&M costs are lower but as the company grows, so too will your S&M costs.

    Brian is bang on with his analysis, from a customer value perspective there is really no difference between an open source software company and a proprietary software company.

    Ian Skerrett

    April 1, 2010 at 12:43 pm

    • I agree. From the perspective of mainstream customers there should be no difference between proprietary companies and open source ones. Any differences will confuse them.


      April 1, 2010 at 2:28 pm

      • I think that is all Brian was trying to point out in his initial post. Enterprise customers should not expect anything different from open source companies? I am not sure where is the confusion?

        Ian Skerrett

        April 1, 2010 at 3:24 pm

  6. Why can’t the open-core approach be both a gimmick AND and a fundamental part of the business model?

    There’s no shame in using a gimmick. The shame is in having one’s fingers crossed behind one’s back when using the words “open source” as the opening words of one’s marketing tag line. Nobody is asking you to change your business model, only to be honest about what it is and is not.

    I’m also curious why your response failed to address Brian’s “super-size trigger” point.

    Jeff Gehlbach

    April 1, 2010 at 1:17 pm

    • We are using the same terminology that everyone else is. We are honest about it. This terminology battle has been going on for years without any kind of productive outcome.

      The super-size trigger applies to some extent to all open source companies – you are either a customer, or not, you either have support, or you don’t.


      April 1, 2010 at 2:27 pm

  7. Removing from the equation any personal consideration (as to maintain a peaceful discussion) I would say that the biggest problem is not actually related to S&M, but how the OSS model (or part of it) can introduce a difference in the traditional commercial software development structure.
    The traditional open core (with a split functionality between a community edition and an enterprise one) has more or less the same disadvantages of dual licensing, in the fact that the shared core tends to receive limited external contributions, when compared with pure OSS cores (we are talking about licenses, not business models). This means that for all the aspects related to development and research, having the core OSS is a limited advantage in reducing the R&D expenses, and can only provide an advantage in terms of market seeding – that is, it will be easier to disseminate information about the product, people will try it more and so on. In this sense, there is not much difference between open core and the traditional “express” offering by most proprietary software vendors.
    Our own research found that S&M expenses are not significantly different from traditional proprietary companies in the same revenue range (confirming what Matt found in his CAOS reports – not that Matt needs my confirmation of course :-)) so from the point of view of the large scale economic picture in the relations between producer and consumer there is little change between open core OSS companies and proprietary companies offering an express edition and an enterprise one (and in this sense, there is no positive/negative connotation in this: simply the fact that average economic parameters are similar).
    One of the reason I believe made Tarus so suspicious of open core companies may be related to the fact that many used the “open source” moniker to promote what can only be considered a glorified demo; this is also one of the aspect that seem to be described in Gartner piece.

    Carlo Daffara

    April 1, 2010 at 1:42 pm

    • So which open core companies only have a glorified demo in open source? I hear this from multiple sources – that the core code is limited/hobbled/crippled.

      I agree that this is a bad strategy because the adoption that fuels the growth won’t happen.

      I have asked numerous people to name these companies, but no-one has.

      So Carlo, who are they?


      April 1, 2010 at 2:37 pm

      • One example would be Groundwork Open Source. It is basically Nagios with a (basic) frontend. It misses every kind of server monitoring, only has “ping” as data source, only very basic discovery (that “basically” does not discover anything at all), no email notifications. Not to bash on GWOS, of course – just one example. I will be happy to name more, but I will only get more enemies 🙂

        Carlo Daffara

        April 29, 2010 at 9:22 am

  8. James,

    All software from open core vendors *are* demos / entry-level software. It’s inherent to the freemium business model. Which is fine, but remove all benefit of open source *for the customer*. For the vendor, of course, it’s great but that’s not so different from proprietary software vendor relying (and contributing) heavily to open source (ex: Day, ORCL, GOOG, etc.). Open source is neither a gimmick or distribution channel. Open core voids open source of its brand value. That’s a shame. Open source means open access to the source code. Open core model is based on proprietary source code of the software. No ideological issue: just call it what it is. And that’s not open source.

    Full answer is here: http://blogs.nuxeo.com/ebarroca/2010/04/business-open-source-my-take-on-the-open-core-debate.html



    Eric Barroca

    April 1, 2010 at 9:57 pm

    • Sorry Eric, but you are wrong. Try Pentaho. It is not a demos. We have thousands of installations world wide using the community edition. Our business model does not work that way.

      Tell me which open core community editions are demos. I’d love to know. I keep asking but no-one can answer.

      If you want a battle on the terminology that has always been used around this model, you came to the wrong place.


      April 1, 2010 at 10:02 pm

  9. Eric,

    Maybe I have this wrong, but doesn’t Nuxeo Connect include ‘Premium Tools’.

    As far as I am concerned you are the CEO of an open core company. Yet you are bashing your own business model.



    April 1, 2010 at 10:07 pm

    • Hi,

      Nuxeo Connect include SaaS. Not tools / code / additional features to apply on the open source software. Pretty much like RedHat for its OS or for JBoss. So, no, we’re not an open core company.

      And again: I don’t have any problem with the open core and the proprietary approach. I do believe there is value delivered through this. I’m just saying, it’s important to not blur the positioning and undermine the value of the ones that take a different route. Open core is not open source, just because open core is a business model (most known as freemium) while open source is not (it’s just a licensing / development model). And I don’t believe open core is the best way to deliver value from OSS to customers. But that’s just my opinion. 😉



      Eric Barroca

      April 1, 2010 at 10:19 pm

      • So you don’t have any source code that is not in open source? Even in your SaaS offering? All the source code for your ‘Premium Tools’ is in open source?



        April 1, 2010 at 10:32 pm

  10. […] buy software. James Dixon, CTO and Founder of Pentaho, has some excellent blogs on this. (“The Analyst’s New Confusion”, “Open Core Business Model Revisited”) One of the common ideas of open source companies is […]

  11. […] buy software. James Dixon, CTO and Founder of Pentaho, has some excellent blogs on this. (“The Analyst’s New Confusion”, “Open Core Business Model Revisited”) One of the common ideas of open source companies is […]

  12. […] they find, they can then upgrade to an enterprise subscription. James Dixon, CTO of Pentaho, hasargued for years that the cost of marketing and selling closed-source commercial software means that most vendors […]

  13. […] they find, they can then upgrade to an enterprise subscription. James Dixon, CTO of Pentaho, hasargued for years that the cost of marketing and selling closed-source commercial software means that most vendors […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: