Wednesday, January 14, 2009

An article last week on got me thinking about open source ERP again. The article highlighted another open source ERP project called Openbravo. Apparently a couple of months ago the number of downloads of Openbravo passed the 1,000,000 mark. That's how the success of these products is typically noted, but it is pretty much a meaningless statistic.

Compiere was the first of these projects that I remember getting media attention. There is some Oracle DNA at the top of the company and in one story the founder was quoted as saying that he doesn't view other open source ERP projects as competition. His competition is Dynamics. Compiere got a bunch of attention as its download number from SourceForge (a site that tracks open source projects) reached into the hundreds of thousands. Like other open source business solution providers the revenue is in the services side of the business. Based on estimates of total revenue and Compiere's service pricing I estimated that as little as 0.1 % of the downloads equate to actual customers.

Openbravo just boasted of passing the 1,000,000 download milestone. Yet from the notes of an attendee at the Oct 2007 Openbravo World Conference (attended by 120 people) there were at that time less than 100 live customers (80% of those in Spain). The download number comes from now and the live customer number from over a year ago, so the ratio may not be exactly 0.01%, but you get the point. Let's say that half of those downloads have come since Oct 2007. That still means the live customer number is just 0.02% of the download number.

Taking this further, if the customer has 20 users, do they download the software 20 times to install on each PC? You get the point, inflated download numbers are indicative of web traffic, not of user adoption. Businesses should take that into account and not be wowed by "1,000,000 downloads!".

I don't think this model is going away, but it's not about to take over the ERP market either. Per the article, Gartner Group predicts that by 2012 more than 90 percent of enterprises worldwide will deploy open source software in one form or another. The vast, vast majority of that will not be in the area of ERP.

And besides, Gartner Group's predictions are only about 71% accurate, and even then only 54% of the time (0.4 probability).


  1. @AbridgedMind,

    You made a good point. I call myself Openbravo the major open source fraud I've ever seen and I'm aware of a bunch of internal infos you might not know.

    Also notice that:

    * I, as an experienced programmer can testify how bad is Openbravo codebase and how functionally non competitive it is on the ERP market in general. Indeed, they are really Oracle rooted despite calling them Java and abusing buzzwords and claimed features like CRM and BI are non existent to say the least. It's also very buggy and non user friendly as it takes lots of clicks to achieve simple things.

    * The people who criticize Openbravo the most are people who know it the best, easy to check: almost not a single third party success story to be found, lot's of deeply involved guys dismissed too.

    * Only new naive entrepreneurs partering who obviously never got their hand dirty can show enthusiasm about the product, and since they invested themselves in that partnership fraud, they hardly can step out of that shit so easily as they discover slowly they got caught.

    * Take Magento, the oss ecommerce, for instance, they claim only 500 000 downloads, but the adoption rate is MUCH higher because the product actually works (I think it's mostly slow shit under the hood but I should admit it works well for what it addresses). Just digg into the Magento blog and related forums to see how many real success stories they have for half the downloads of Openbravo, compare this with the eternal same poor sucesss stories of Openbravo funded by their first fund raising and the noob profile of the average Openbravo forums users who didn't realize the fraud yet.

    * Finally Compiere might tell you that 0.01% of download turn into adoption. But you should absolutely understand that for Openbravo the adoption is even much lower than 0.01%!!!!!!!!!

    Why? Simple my friends: Compiere is an expensive quite closed business model but it actually works more or less. Compiere integration is controlled and expensive but Compiere get money out of that they re-invest in their product.

    On the contrary, Openbravo guys do not even manage to only maintain their crappy Compiere original codebase properly (that's why it used to be so many dead broken Compiere code in Openbravo). Those dead feature are even polluting the user interfaces, just like Openbravo thought it was better to let potential customer think it might be implemented rather than just helping the user experience for doing the little that works.

    So, unlike Compiere, Openbravo have chosen a different path: they decided they would fraud investors making THEM believe they get a high user base and therefor they might be bought later on by Companies like Sun Microsystem, just like they bought MySQL $1 billion (see those faked blog announces: ). Of course not a chance this happens cause Sun strategists are somewhat rational and tech aware even if JavaFX is a dumb move.

    But you get it. This is a muti-agents fraud: they don't really try to make you believe their product works, they tried to make investors believe you and me believe it works...

    That's why they spent all their money and effort on the marketing (last development expenses do not reflect invested money at all), thus artificially increasing downloads figures for the same very low accidental adoptions. So I mean overall that 0.01% adoption rate is even MUCH higher than the reality.

    Finally I think now they are in a dead end. I think Openbravo managers will start having bad sleeps, because now that got the money from the investors, but day after day it's harder to keep them dreaming buy making buzz on fake blogs, faking a community...

    On the contrary, all the Openbravo team is just realizing how far away they stand from a real competitive ERP, some were leaving before the big crunch... Not to say they are some smart organizations like OpenERP or Tryton who are actually making it for real with a tiny fraction of their investment, just because they were much smarter on the tech heart of their product and savy on the marketing.

    But now Openbravo is in a dead end: they know they can't fix their product, even spending their money and they feel uncomfortable spending money they don't really own and will never make back.

    They release too they couldn't hide anymore how much in a dead end they stand so the dream of selling their company and running away with the money is also ending too.

    Still, what is frustrating I think is guys like us had anticipated this to happen more than one year back and made no money out of those predictions while VC's still bought money to burn into that Openbravo dead end. There is definitely something spoiled in the way finance works, I hope in the future the oss market could be regulated buy guys like us who could then bet on the opposite alternative: the burst of the Openbravo bubble.

    Regards, and thank you for your valuable analyze.

    jackthripper (for now).

  2. At least for CRM, it is obvious that SugarCRM has a lot of traction in the low-end of the market. With young companies that may one day become big ones.

  3. "If the customer has 20 users, do they download the software 20 times to install on each PC?"

    Of cause NOT. They install it from LAN. I’m really can’t imagine person to download the same software for multiple installation from the Web.


  4. Dmitry - fair point on the 20 downloads. I guess I don't know enough about Openbravo to know if it would be run in a distributed environment over the web or only with all workstations directly connected on the same LAN.

    jackthripper - Openbravo just added some significant leadership to its board of directors. Hard to say it's a dead end.

  5. "Dmitry - fair point on the 20 downloads. I guess I don't know enough about Openbravo to know if it would be run in a distributed environment over the web or only with all workstations directly connected on the same LAN."

    My comment was not about Openbravo but just about common practice I saw regarding software distribution to each user's PCs inside enterprises.

    Talking about Openbravo it's even easier. It has a Client (WEB client) - Server architecture which in our discussion means that enterprise has to download software package only once to install it on a Server and the users don't need to install anything else on their PCs (ordinary WEB browser is enough) to use it (inside the corporate LAN or from remote locations) thanks to WEB client.