Open Source innovation vs. commoditization

Matt Asay took me to task for a comment I made at the Enterprise Software Summit this week:

“and had Jeff Nolan
(Director of the Apollo Group at SAP and former VC with SAP Ventures)
challenge open source as a tired commoditizing play. He’s right. It
really is dull if all open source can do (or, rather, what open source
commercial enterprises tend to do, commoditize big, existing markets).”

I thought about this when I read it and replayed the exchange we had during his presentation wondering if I was really saying that. Short answer is that I am but not in the way Matt suggests, and I’m not even sure I’m talking about commoditization.

Matt was suggesting that there is a wave of open source enterprise application companies on the horizon and that they will offer the potential of displacing the incumbants. That could be one scenario. The comment that I made is that nobody in the enterprise world really gives a shit about whether an app is open source or not if the application doesn’t solve their problem. This is the problem that applications like SugarCRM are facing, first they have to build a good CRM application as the entry fee to the market, then they can blather on about open source all they want. If Sugar doesn’t build a good CRM app then what is the incentive for anyone to 1) select them over someone else who has a functional app, or 2) displace an existing vendor from the account.

The point about displacing existing solutions is an incredibly important one because if you spend a lot of money on a traditional enterprise application like Siebel you are looking at that as an investment that you will use for future benefit. It’s far different than saying “well I think I’ll use Microsoft Money instead of Quicken because it’s a lot cheaper”.

The point at which enterprise software vendors are really in trouble is when their customers start looking at that license/services/support expense as a sunk cost… it’s gone and not coming back so there no reason to stay on the current path. I don’t see any indicators that would support this today, but it’s a real possibility that the entire industry needs to be aware of, and for that matter actively create conditions that result in better applications cheaper.

We love open source, we’re big supporters of MySQL, we invested in Red Hat, Sistina, Zend, and most recently Socialtext. SAP was the first major application vendor to officially support Linux as an application platform, and we continue to do things like add development support for Eclipse in NetWeaver. All of the investments in OSS we made were predicated on having an economic model on top of a solid application, the fact that they are open source really was not a factor in our investment decisions.

It’s somewhat ironic that Matt uses Socialtext as an example of a company creating a new market, which I really have to disagree with. Socialtext is using wiki technology to disrupt the existing market for collaboration and knowledge management software, in other words a market that already exists but is poorly served. Commoditization also implies a loss of pricing control, which is something that Socialtext is not experiencing as a result of their success… if anything success in enterprise software actually creates pricing power.

Far too many open source application companies think that the reason why enterprise prospects will get excited about them is because they are less expensive than proprietary apps. No, everyone is cost conscious but I don’t know a single enterprise class buyer that would simply do a price analysis as the sole determiner in their buying process. The app has to fit their functional requirements in order to win, and if it doesn’t then it has to fit a strong portion of their functional requirements and be priced accordingly so that the effort to make it a 100% fit is manageable.

This is where Vinnie or Dennis jump in about out-of-control enterprise implementations that piss away customer dollars. I’ll wait. Okay, for every one of the high profile ERP/CRM/SCM projects that failed there are hundreds – thousands – of implementations that are delivered on time and work as promised. If it weren’t the case we wouldn’t be in business today. (BTW, I’m not picking on Dennis or Vinnie, I have abundant respect for their points of view and agree with them on many points).

The next thing that OSS proponents say is that you get all the source code… big whoop, we’ve been doing that for the 15 years (if not longer). That’s right, you buy R/3 and you get the source code. There might be warranty or support agreement issues if you go mucking around in it, but so are there with Red Hat.

Community developer support is next up. The reality is that the great bulk of OSS projects have a couple of developers and that’s it. The biggies has teams of developers and quite often they work for the same company. There are literally thousands of developers on SAP, Oracle, Peoplesoft, Siebel, etc… hell, you can even find an entire community of developers around AS/400 software and IBM hasn’t sold one of those boxes in years.

My comment to Matt on Monday was that going out and creating an OSS company for ERP is fighting yesterday’s war. I’d be more interested in seeing someone go out and build something better than ERP/CRM/SCM/whatever and use the open source community process to do that. Innovation means going beyond where the market is today, it doesn’t mean simply doing the same thing cheaper or differently.

10 thoughts on Open Source innovation vs. commoditization

  1. I agree with you what he proposes is not innovation – to me web 2,0, telemetry, bio metrics, predictive analytics are innovation. But … ERP is not innovative today with SOA or not. SCM is not, CRMM is not. They are 20, 15, 10 year old concepts. You, Oracle, Siebel recouped your core investment costs a long time ago and still want to charge “innovation” pricing. In doing so you are opening the door for more “utility” pricing. Open source, SaaS, third party maintenance, offshore implementation are all symptoms of that.

    They are not playing the innovation card. They are playing the efficiency card.

  2. 1. I dont think Open source creates new markets. In fact if you have to use open source to create a new product that serves a new market, it is not a sensible strategy as you are leaving money on the table (customers are willing to pay -there is already a need for your product, but you are not getting any money, its free/cheap!). One is better of making a commercial model out of it when you attack new markets, rather than just embrace open source for the love of it.
    2. Open source, till now, has beeen majorly used to commoditize complements. Sap using MySQL (in trying) to commoditize database, or Oracle leveraging Linux to commoditize operating systems (complement to database) are the top of the mind examples. More on complements at my post here:

  3. Some great points, that large companies don’t make purchasing decisions on cost alone, and don’t really care about open source unless the applications meet the business need – if they care at all about it. I made a very similar point at a SDForum discussion as an audience member awhile ago and was practically hissed out of the room.

    Also the point about the size of the development teams is critical. It can be really hard to attract a talented team for free to your specialized, business critical needs. That model just can’t replace or even really threaten commercial software IMO. The good open source guys are too smart to care much about [insert your boring but important business need here that isn’t too interesting as a computing problem].

    Which brings up Vinnies point: this is about efficiency. And efficiency is very important, and not that dull to some of us. My bet: Open Source combined with developing economy dynamics continue lowering costs, but don’t get anywhere near free that you get with the open source backbone. I’ve used a flavor that, solo software development based on open source software, to develop proprietary software – reaching a market segment (Siebel Monitoring) that is underserved by commercial software vendors and open source. It would have been stronger with a large development team, but I would have had to hire developers like SAP does to develop a significant community.

  4. what’s been commoditized has been infrastructure and developer tools. Applications are a whole different ballgame because of how and who you are selling to. Even in the database market the going for OSS has been alot tougher than the proponents predicted. In fact, what MySQL has discovered is that you just can’t go in an switch an Oracle customer to MySQL if what you are offering is at best just as good as Oracle.

    Believe me when I say this isn’t about being for or against open source, it’s more complicated than that.

  5. Jeff,

    MySQl to Oracle is a substitute, not a complement. SAP and mySQL are complements. For substitutes, things like value chain, switching costs and price tag come into play. I think OSS is more about commoditization of software, as my previous comment suggests.


  6. “MySQl to Oracle is a substitute”

    yeah, that’s pretty much what I said and that substituting mysql for oracle is proving a lot harder than they thought… this is fact, Marten is very open about it. You are right about value chain, switching costs, and price tags being components, but enterprise apps are very sticky because of the investment in and support infrastructure that is built around these apps.

  7. Interesting reading. While I don’t disagree with the general points you’re putting forward, I would like to [nit]pick you up on a couple of details.

    “you buy R/3 and you get the source code” – this is a lot different to “_all_ the source code” (emphasis mine). You know as well as I and many others do that SAP is partially source-open, not fully open source. I have always applauded the extent that SAP does deliver most _application_ code (in the ABAP stack at least) as source code, but it is not the whole nine yards.

    “There are literally thousands of developers on SAP” – yes there are, and this in itself is a big issue, and getting bigger, not smaller, as the machine-gun development angle widens, coding and architecture consistency, uniformity and quality suffers (under the hood of SAP CRM springs immediately to mind as a narrow example). All I’m saying here is that you can also have too many developers.


  8. DJ,
    Fair points, but if we’re talking about open source applications why would any of the code in the app server or performance manager be at issue. In other words, what Sugar is releasing in the way of application code is equivalent to what we are doing. I hear you on source open versus open source, they are in fact different things. Personally, I’m an advocate of open sourcing components lower in our stack… but that’s an entirely different subject.

    re developers, what I was saying is not that there are thousands of developers at SAP working on this stuff, but an entire industry that has been built up around SAP that features independent developers working on SAP code or on applications that they themselves built with our tools and are either independent of R/3 or integrated with.

  9. It’s nice to be picked upon.

    The commoditzation angle doesn’t have to rely on OSS. In some cases it can simply be an acknowledgment that certain funtionality doesn’t provide cometitive advantage. Financials is a great example – everyone needs it but can anyone show me a single case where having a SAP GL is any ‘better’ than having one from Oracle, Microsoft, CODA, Agresso…the list goes on?

    If anything I can readily argue that having R/3 GL is akin to being straightjacketed. Innovation in relation to the enterprise comes from how readily processes are assembled, dis-assembled and re-assembled according to changing need. OSS is irrelevant in this context. People will use whatever tools work best.

    Innovation – in my experience – comes from the guy in the corner who asks ‘why?’ and then finds a way to answer the question. In those cases, size matters. Small that is.

  10. Exactly, like I said out loud during the session, anyone who thinks their path to success is doing what Oracle/SAP/PSFT has already done but doing it with OSS i fighting yesteday’s war.

    Innovation comes, as you point out, from someone going well beyond where we are today.

    BTW, I’m an advocate of open sourcing our FI product line… let the community maintain it and build it out. We should be creating new products, not maintaining old ones.

Comments are closed