Big Endian

Crosstex Energy Implements Newmerix’s Automate!Test To Effectively Automate M&A Activities in PeopleSoft

Superior, Colo., (March 28, 2006) – Newmerix, a provider of application quality and change management software for PeopleSoft applications, today announced that Crosstex Energy has successfully implemented Newmerix’s Automate!™Test—an automated testing solution for PeopleSoft. Crosstex will utilize the tool for several automated testing projects, including Sarbanes-Oxley testing, but the company found immediate cost-savings by using Automate!Test to migrate employee, vendor and customer data from recently acquired companies into their standard PeopleSoft implementation...

http://www.webwire.com/ViewPressRel.asp?aId=11776

BMC buys Identify

Before I started Newmerix, I took a long time deciding what business segment to get into. I looked at a number of possibilities including SRM, real-estate management systems, charting components, packaged application management, and J2EE management. In the end, my short list was whittled down to J2EE and packaged application management. My final decision to go with packaged application mangement was made along a few dimensions. One of them, was that there was already a glut of J2EE startups out there (Wily, Sitraka, Parasoft, Altoweb, Altaworks, etc..) and I thought the market would consolidate sooner than I could build and real value or differentiation into a product of my own. Well, it looks like I was right to be concerned with the timing of everything - as the market does appear to be consolidating. Today, another major management player decided to stake a claim in the J2EE market. BMC bought Israeli Identify Software for $150MM. This is clearly a response to CA's purchase of Wily only a few months ago for $300MM. All good M&A comes in threes (the most recent example would be the Kintana-ChangePoint-Niku triple play for the ITG space), so any bets on who HP or IBM will buy?

Software 2006

I am heading to Software 2006 next week. I'd love to catch up live with any readers or fellow Enterprise 2.0 bloggers that might be attending. I know a number of folks will be there. Drop me a line at [email protected] if you plan on attending. If there is enough interest, I'd be happy to organize a dinner for us all.

Out-platforming Outlook

I wrote in one of my first BigEndan entries about a number of companies that are chipping away at the Microsoft Office suite by producing online versions of each of these applications. Today, Charles pointed me to Zimbra, who will be presenting at Software 2006 next week. I watched their end user demo and it's very interesting. Not only are they providing an online set of features that compete with Outlook, they are providing direct integration with Exchange servers, supporting key features like Blackberry integration and being able to work offline (caching). More importantly though, they have a built in web services interface for mash-ups directly integrated into the platform. Take a look at the demo to see what I am talking about.

The reason I like this is that I think Microsoft is losing out on a golden opportunity to position Outlook as an application platform. There are some interesting applications that seem to have co-opted Outlook as a platform whether it was really designed for that or not. Familiar examples would be Plaxo , GoodContacts, and Salesforce.com Outlook extensions (which I now use religiously). Some more complicated integrations exist like Attensa  (which i stopped using 2 days after installing it because it made my machine grind to a halt), and the Mendocino project with SAP.

But is this too little to late? A quick search of Amazon produces only a few books really oriented towards developers trying to build complete applications on top of Outlook (and a number of the reviews point to how hard this really is based on the architecture of Outlook). For those that have read my Increasing Tail blarticle, they will know my philosophy that, in the end, the platforms that win are the ones that seed the market with a killer application and then give developers the most tools to extend the value of the platform with their own niche applications and content. Out of the gates, Zimbra gets the developer part. The question is can they find the killer app that will drive their adoption?

ERP Down Under

I read a good article covering the results of a smallish survey in Australia on Oracle, SAP and Microsoft. Not suprisingly, Oracle and SAP seem pretty evenly split, there is skepticism about any up front cost savings of SOA, Microsoft is very far behind in perception, and licensing fees are becoming a much bigger source of concern as folks contemplate how much to pay for licensing before they make any final choices about Fusion or NetWeaver.

For those unfamiliar with the options for Oracle/PeopleSoft/JD Edwards/Siebel - here is a short list to get you started:

TomorrowNow (leading PeopleSoft, JD Edwards 3rd party maintenance provider) [full disclosure, TN is a Newmerix customer]

Klee Associates (JD Edwards)

Rimini Street (Siebel)

BTRG (PeopleSoft)

LegacyMode (PeopleSoft)

Cage Match - Part Two

I an earlier post - I pointed to a fun online debate over who would win: SAP or Oracle. Each participant wrote an article arguing their reasons for SAP's or Oracle's inevitable success or failure. The follow up is a rebuttal from both original authors to their counterpart's first piece. It's worth a read, if for nothing else than the pretty open handed jabs between the authors.

GPL 3.0

Okay - for once I am not trying to tag a version number onto something -this one already has a version number. In Plato's Software Children (PodCast) I wrote a bit about Richard Stallman, the FSF, and the GNU Public License (GPL). Many of are familiar with the GNU open source licensing scheme and some of its nasty viral effects. Well, the folks at the FSF have put out their first draft of GPL 3.0. While it cleans up a lot of the viral dynamics of the 2.0 version (there is specific language around it being able to cooperate with code bases that include multiple licensing schemes), it does introduce some nasty anti-patent suit clauses. With the advent of Creative Commons (which this blog is licensed under) and the growth of the BSD open licensing scheme, why do the folks at FSF always have to whack you with a stick after giving you a carrot if you are trying to build and sell enterprise software for a living? Some excellent coverage on the subject can be found in these articles:

SD Time (subscription required, but this site is worth it)

Forbes on Linus Torvald's response (Linux is GPL 2.0 licensed)

An EXCELLENT article on C|Net about the folly of FSF ideology

The Demise of Friendster

I read an interesting article/exploration by Danah Boyd on the releative successes of Friendster and MySpace. While it is more "anecdotal viewpoint" than classic research (which is what we are all pretty much doing online anyway), it has some interesting nuggets in it. Also - in many ways it is arguing a similar point that I made in the Increasing Tail - platforms who give their users tools to develop the tail end of content almost always win.

Illegal.com?

I read a very interesing post in response to my entry on Salesforce.com sandboxes. It appears that this may have some challenging legal ramifications in the UK and EU related to copying customer data. If nothing else, a very interesting tangential thought on SF.com.

Startup 2.0

There has a been a lot of chatter out there around Startup 2.0 philosophy (I am staking my claim on everything 2.0). At the end of the day, a lot of the discussion revolves around how to make technology startups cheaper to build. The traditional rule of thumb is that an enterprise software start will take about $20-25MM investment to make it profitable. I know this seems like a lot, but when you consider the investment in developing complex enterprise software products, a direct sales model, marketing, facilities, customer support, etc.. all dimensioned by the time needed to penetrate deep enough into the market for brand and maintenance revenue to kick in - the costs start to rack up. Some, of course, have done it for less, and many have spent much more on their way to success or failure.

So, when people talk about the next generation of startups and venture capital, they are really asking "How do we make this cheaper to do?". Well, i wanted to lay out three approaches that I see emerging.

1 - Open Source as a Starting Point (not as a model)

While there are a few successes (MySQL, RedHat, Novell, etc..) of the "give it away for free and charge for maintenance" model, in my opinion is yet to be proved widely as a sustainable model. There is another derivative of the open source model that I am seeing emerge: companies who fundamentally use open source as a core component of the initial solution they deliver. In fact, I am invested in a company that started this way. StillSecure built it's first generation of security products on top of Snort (an open source IDS system). This enabled them to quickly penetrate a growing field of vendors in the security market with a 3rd generation level product, allowing them to moved cheaply towards more classically proprietary extensions and products.

From the entrepreneurs perspective, building on top of open source can reduce cost and increase speed to market. However, i wrote a long blarticle-based cautionary tale on the use of open soure on Parallax. Handing over your core code base to someone else (or a group of someone elses) is a blessing and a curse. Buyer beware.

2 - Offshoring

About 2 years ago when Newmerix was raising its second round, there was a lot of chatter from the VC community around offshoring. "You're going to offshore development, right?" was a pretty commonly heard phrase in pitch meetings. Does Nantucket count as offshore? For many reasons, we consistently argued that doing it "in house" at our stage of a company was a better way to go. In general I think the VC fervor around offshoring for startup companies has died down a little bit as ROI's have proven to be smaller than expected (at least in India, who's salary rate has increased double digits per year for the last few years). In addition, anecdotal evidence is saying that large companies who can open their own shop in India are the ones really reaping the rewards. There is enough chaos in a startup as it is, that waiting 12 hours to ask a pertinent technical question is a pretty debilitating problem.

3 - Saas

While I have a longer piece coming on the "real reasons" why SaaS will take over the world - here is a little snippet related to starup costs. One of the biggest costs involved in a startup is the N to the Nth platform support problem you face when shipping out CDs. Take a normal application today. It runs on a browser (which versions?), a database (which vendors and which versions?), an operating system (which vendors and which versions and which service packs?), and possibly interfaces with something else in the application stack (application server, BPEL services layer, packaged application.. which versions?). This causes an unbelievable amount of cost in the development and support cycle. I'd estimate it takes about 50% of your development time and resources to actually properly QA a piece of enterprise software in all relevant environments for your customer base. Not to mention the cost of managing a lab of all of these permutations (thank the heavens for VMWare). But then, when you move forward, you have the additional problem that customers are going to be using different versions and patch levels in the field. What upgrade paths do you support, which do you not - you have to test every upgrade path. Then, when a support call comes in, you've got to rebuild the exact customer environment to find the problem many times. Its a cost management nightmare, but one we've been living with for so long that we sort of don't talk about it anymore and assume its part of the cost of doing business. Frankly when Microsoft slices and dices the next version of a platform (e.g. Vista) into 6 marketing bundles, my stomach turns at the thought of how many new permutations we'll have to test (or not test and just "hope" it works).

SaaS changes all of that. You only ever have one platform stack to test - the one you run. You only have one version upgrade to test - the one you run. And putting out patches to your software can be done at your own pace, incrementally, and there is only one IT department to wrestle with - yours. It's a phenomenal model for startup companies. The classic virtual urban legend around this is the mere 100K Joel Kraus apparently spent to start JotSpot. Even if he is an order of mangnitude off on the build costs, the maintenance costs alone will be dramaticallly different than classic enterprise software.

So, while I'm not arguing the death of classic enterprise software (some things just need to be written as desktop applications - e.g. Newmerix Automate!Test product), I am arguing that startup entrepreneurs and VCs look very hard at viable SaaS business models, as the risk/cost equation has all of a sudden moved heavily in their favor.