Asay the Alfrescan

In a former life, I did a Masters degree in International Conflict Analysis (a fancy way to say “International Relations”). As part of my thesis (”The End of American Empire,” a discussion of the source of the United States’ ideological hegemony and reasons for its imminent collapse), I read Francis Fukuyama’s Trust. Basically, Fukuyama argues that high-trust societies (Germany, Japan, United States) tend to be more productive and economically prosperous than low-trust societies (Italy, France, Korea). Trust matters, and translates into cash.

Whether Fukuyama’s theory matches up with the real-world economies or not, I do think it applies well to software. Jason Matusow of Microsoft once related an interesting survey that he/Microsoft had conducted. (This is public information - he related it at a conference in Toronto.) 60% of the surveyed customers felt that access to source code was very important. Of these, 95% said they would never look at the source code, and 98% said they would never modify it. So why did they care?

Trust. Or lack thereof. Not lack of trust that Microsoft had their best interests at heart (I actually think Microsoft generally does have its customers’ best interests at heart, but goes about securing them in the wrong way), but just a desire to have a safety net under their relationship with the vendor.

One of the great things about open source is that it heightens product permeability - you can see what you’re getting before you come to the decision point: buy or don’t buy. As in a strong-trust society, the barriers to enter into contracts, associations, etc. are diminished. IT decisions are based on maximum information (perhaps too much information at times?), not marketing messages and sales guys with good hair. That’s why I like this slide we use at Alfresco:

No, trust isn’t unique to open source companies, and an open source company is just as capable of acting in ways harmful to trust as a proprietary software company is. But the foundation is more amenable to trust. And, if Fukuyama is right, this means that over time, trust-based software economies (open source) will trump the closed software economies.

At the risk of raising Larry’s ire again, as the end of Alfresco’s quarter draws to a close (giving me a little more time to think), I wanted to write down some additional thoughts on revenue models in open source. So, in no particular order, here are some important things to consider when building your open source revenue model:

1. Price matters. This seems obvious, but if you have an exceptional product (as an increasing number of open source companies do), there’s a strong temptation to position your pricing just below, at parity, or even a little above (if you believe Microsoft) the proprietary competition. Don’t. Part of the drive toward open source is price. It’s not the most important part by a long stretch, but it is important. By the same token, be careful not to price your product too low. Alfresco has actually lost customer interest in a few cases (and I’m aware of other open source companies who have experienced the same thing) because the price was too low. You want to be in the same ballpark, but markedly less.

2. Open source is never 100% pull. By “pull” I’m referring to Larry Augustin’s excellent discussion on this topic. I agree wholeheartedly with Larry that open source can create an inexpensive mechanism to enable pull: you create a great product, prospective customers download it, they evaluate it, they call you to buy. But this last part depends, to a large extent, on what kind of product they’re downloading, and how easy it is to install, configure, and evaluate. A CRM system is, arguably, easier to evaluate than EAI or ECM or ERP. In short, some products are relatively self-evident, while other products are not.

   So, when do you start to push the pull? At Alfresco we get involved with companies evaluating our products early into the pull cycle - this has worked well for us. We’re able to help would-be customers understand how to derive maximum value from the system, assist them with configuration, etc.

   As a result, however, we don’t get as much of the automated sales cycle that some other open source companies do. As a tradeoff, however, we do tend to get larger deal sizes than the industry average (just as Medsphere does, an open source medical information systems company). We’re still fine-tuning how we interact with open source’s natural pull mechanism, but I’m generally quite pleased with the results. Open source shouldn’t be about push, but it’s never completely about pull, either. Not if you’re looking to grow a serious company.

3. Document everything. Related to #2 above, one way to facilitate the pull phenomenon is to heavily, clearly document your product. This tends to be the last thing that open source projects think about, but it’s the first thing you should do. It will dramatically lower your cost of sales, because it will mean far less “pushing the pull.” This is something that I’ve learned the hard way - don’t follow my lead on this. Document first, then worry about selling.
4. Revenues now. Every good corporation as revenue and profit as its two primary goals. I like the way Jack and Suzy Welch recently put it in their BusinessWeek column (”The Whining Game“):
   

   You are running a company, not a social club or a counseling service. Your No. 1 priority is to win in the marketplace so that you can continue to grow and provide opportunities for your people. Of course, you want your employees to be happy. But their happiness must stem from the company’s success, not from their every need being met. When the company does well because of their performance, they will thrive, personally and professionally. Not the other way around.

   Money matters. It matters now, not two years from now.

   Of course, as Larry wisely suggests, you don’t want to ramrod sales conversions from downloads - if you try, you’ll fail in the short term (would-be customers will walk away) and the long term (you’ll find yourself more sales-focused than product-focused, and will lose your price and product advantages). But you need to be thinking about how to convert downloads into dollars every second of every day - it should be second only to creating an exceptional product (and directly correlated with how you’ll serve/support customers).

   Red Hat isn’t the company it is today because it made a lot of friends gifting great technology to the world (and people happily paid for the brand), as Bob Young once implied. Red Hat became the most successful open source company on the planet (ranking #2 on Business 2.0’s 100 Fastest Growing Tech Companies list) out of dire necessity: it went public on hype and bubble-esque thinking, and then had to hunker down to justify its (then) bloated valuation. In the process it quickly discovered that the we-give-things-away-and-people-pay-us-for-our-brand model simply didn’t work (sorry, Bob). It had to find some way to push all that pull, and came up with an ingenious business model.

   (Incidentally, on that note, it’s strange to me that very few open source companies have followed Red Hat’s business/revenue model. There are at least two reasons for this, I suspect: 1) It’s hard to fully let go and believe that customers will buy a tested, certified (that third-party apps will work with it), supported product. But they do. (Alfresco recently made the move to 100% open source and it has been an unqualified success for us.) 2) The model has been hard to understand, because there are some legal nuances to it that aren’t immediately self-evident. But once you grok it, it’s a thing of beauty. These are just two reasons - I’m sure there are others (including potentially better business models). But I think Red Hat’s success is directly tied to its model - something worth studying.)

5. Be open and permeable. By this I don’t mean your source code. I mean the company itself. Put your pricing, business model, product information, forums, etc. online. Accessible to all. Including competitors (especially them, actually - it will strike holy fear into them when they see customers leaving them in droves with the ability to see exactly why their customers are deserting them, and not being able to do anything about it). You want to lower the bar to doing business with you. No one should have to wonder about your pricing or anything else. John Powell, Alfresco’s CEO, often laughs when we sign mutual NDAs - “Matt, be sure to keep secret what you learn about the other company. They already know everything there is to know about us.” That’s the way an open source company should be: open.

That’s all for now. Of course, there are various other questions to be answered (Per-user or per-CPU/server pricing? Banded pricing or unit pricing? etc.). But many of these are only answerable in the context of your product/business.

Enjoy!

Thus spake Simon Phipps, in this eWeek article. Perhaps I’m being naive, but I think the answer is relatively simple: GPL it. Not one of those newfangled modern corporate-created licenses. The stubborn old GPL.

Yes, from the GPL you still get various flavors of Linux. Ubuntu. Red Hat. Gentoo. Etc.

But the differences between these are more in the assembly of packages than anything else - they’re much more similar than they are different.

So, GPL it, Sun. Set Java free.

I had a funny “customer win” last week that reminded me (and the new customer) that signing on with an open source company can be much more than a passive relationship….

Let me explain.

The customer was looking to cluster Alfresco in a way that we don’t currently support. “We,” meaning Alfresco. But they told me that “Danruss” was working on the code solution to the problem, and would have it by July.

“Danruss?” I queried. “Who the heck is Danruss? We don’t have anyone by that name working for us.”

“He’s all over your forums,” they replied. “He told us he’s backlogged right now, but will be issuing a beta by July 1.”

I couldn’t figure out what (or, rather, whom) they were talking about. But then it hit me:

Russ Danner.

Russ is with The Christian Science Monitor, and has proven to me once and for all that open source enables a highly interactive, dramatically innovative way of developing software. No, Russ doesn’t work at or for Alfresco. He works for The Christian Science Monitor. But he definitely works with Alfresco, and in a way that proprietary software has never been able to replicate.

So, Russ, you won’t be getting a paycheck from us, but you will be getting (and giving) exceptional software. Thanks to you. To Alfresco. To all of us. That’s the power of open source.

Jose Mourinho is quoted today about how pleased he is that the best striker on this planet - Thierry Henry - turned down two record offers of nearly $90M from Barcelona and Real Madrid to stay with Arsenal. Why would the coach of the world’s most competitive team (in the transfer market, anyway) be happy that Arsenal will be stronger next year?

Because competition is good.

Who cares if Chelsea wins if they have no competition? Indeed, Mourinho has bemoaned this fact lately - no one takes him seriously as a coach since he has billions to spend on buying up every good player on the planet.

It’s the same in the open source world. First, it’s becoming a bit distressing to me that the proprietary players have come up with such anemic rationales for why IT buyers should choose them. Innovation? Open source has that in spades. Or how about this one from Microsoft, arguing that open source is not reliable or dependable? I think Redmond has failed to notice the mass exodus away from its rainy shores for stable, reliable, dependable open source in the form of Linux, MySQL, Apache, etc. etc. etc.

In short, I hope the proprietary vendors learn to put up a real fight. Right now, they’re looking pathetic.

Of equal importance to me is that open source breeds an increasing number of quality projects. We need more, not fewer, Alfrescos, SugarCRMs, MySQLs, etc. Too often we think that because one has been funded or grown in popularity, that it’s time to move on. The proprietary world doesn’t think that way - why should the open world fall into this trap? Where would Oracle be without DB2? And Sybase, Ingres, etc.?

Competition is good. Lots of competition is better. Even for you, Jose “Big Bucks” Mourinho.

If I seem dismissive about ODF and Microsoft Office generally, it’s because I am. Over the past 2-3 years I’ve watched Microsoft build a new, growing bastion of lock-in.

The battle is no longer being fought at the file level. It’s being fought at the network level.

The network of files, that is, within an organization. People in the open source world make a fetish out of defeating .doc, .xls, and .ppt with .odf. Fine. But in Microsoft’s new world, even ODF documents would be locked into its network. Of which network am I speaking?

SharePoint.

I first encountered SharePoint at Novell, where Microsoft was using it to nudge Linux out of organizations. See, SharePoint is insidious. The basic version of it (Services) comes free with every Windows 2003 server. It costs departments nothing to deploy it. Given a taste (it’s a decent, though not great product), a significant number upgrade to SharePoint Portal, start storing their content in this SharePoint repository, and kiss their freedom goodbye.

This is the same for any proprietary content repository. Documentum, Vignette, etc. have been locking in customers for years with their respective repositories. But Microsoft is more dangerous, because SharePoint is integrated with Office, Windows, SQL Server, and every other Microsoft product. Once you get a taste for SharePoint you have to keep buying more and more Microsoft product to leverage it, and the more content you store in the repository, the less likely you will ever get it out.

You’re locked in.

This is one reason that companies should be extremely wary about using SharePoint, in particular. It is your content, not Microsoft’s. If you want to keep it yours, you need to keep it in a secure but open place.

There are a range of great open source repositories out there (Alfresco (Truth in advertising - I work for Alfresco), Apache’s Jackrabbit, Plone, etc.). This is where you want your content stored, because each of these offers easy ways to get the content in and, more importantly, out.

So, yes, I am a bit blase about file formats. That is yesterday’s battle - an important one, but an old one. Today’s battle is being fought in the network of files. You may not realize it now - though companies like Novell are already waging a fierce battle on this front - but you will. Get your data/content out now.

I’ve been having a discussion with a group of friends on the advisability of calling out the fact that open source is cheaper. With other benefits (Flexibility. Open standards. Performance. Quality of code. Etc.) arguably trumping cost, I decided to ask the only person that matters in the debate:

A customer.

This particular (Alfresco) customer vetted various proprietary ECM solutions before going with Alfresco. The price for Alfresco’s top competitor in the deal was 20 times Alfresco’s. 20 times. They eventually discounted it so that it was only 10 times more expensive. We sold them a similar system for under $30K (one that actually performs better, is easier to use, and won’t lock the customer’s data into Alfresco).

You do the math.

Now, if I’m a CIO, I’d better have a darn good reason for not choosing Alfresco in this scenario. (And if I’m that other vendor, I’d better be buying indulgences or something to assuage my guilt for over-charging so much for a relatively simple ECM requirement.) Is open source primarily about lower cost? Absolutely not. But should buyers care about cost? Of course they should. And they do.

Price may not be a long-term competitive barrier, but it sure is an excellent short-term opportunity for open source. Use it.

I figure I might as well maintain my status as Resident Inquisitor of Open Source Myths with a discussion on the value of choice. A friend at a Fortune 500 company recently set me to thinking on the problems (and opportunities) that open source affords vis-a-vis choice. (I’ve opined on open source choice before, in case you’re interested.)

I’m very fond of telling enterprises that open source maximizes their choice. I often use one of Larry’s graphics to illustrate how much better off they are:

Look at all that choice the CIO now has! She can spend her money in a variety of different ways.

Oddly enough, that can be a problem. In many ways, it’s easier to be forced into a decision: if I only have $10 to spend, in some ways I’m glad to have $9.95 in Arsenal tickets staring at me. My choice is made. No need to worry about spending $2 to send condolences to the Barca fans.

However, there’s a much more difficult angle on choice for CIOs: how to figure out what to choose in the first place. My friend tells me that it can be hugely time consuming to download and try out open source software. Vendors like MySQL and Alfresco take it as a matter of course that our would-be buyers will first download, test, and evaluate our software, and then opt to buy it. However, what if they go through all that effort only to find out the product is rubbish? Lots of man hours on the road less traveled…and it will have made all the (negative) difference.

Oddly enough, in the commercial world, the buyer gets the product largely “site-unseen” and then has the pleasure of beating up the vendor to get them to make it work. They paid big money for it, so they’re a) not going to recognize the sunk cost and move on and b) have given the vendor a strong incentive to fix the problems or face a battering in the word-of-mouth press.

For many, they would prefer to throw down the cash and pray it works. For others, they’d prefer to invest their time/money in experimenting toward a good fit (though, let’s face it, the process for winnowing down a universe of products/projects to a small group for a bake-off is the same - IT buyers are always going to talk to peers, read what the media has to say, etc., as I’ve noted before).

Over time, I think it will become increasingly easy to find good open source software. So, some of the “cost of choice” will be removed for my friend. But for now, there are real costs associated with choosing open source. Worthwhile costs, yes. But costs, nonetheless.

eWeek has a good overview of Sun’s announcements this week (Operating System Distributor’s License, for example).

One announcement that it didn’t make? That it’s open sourcing Java. My friend Dave Rosenberg gave a great line in this regard (must have taken him days to think of it): “I applaud Sun’s philosophy to open-source all of its software, but the community is asking for Java. Until Sun open-sources Java, their open-source credibility flag will still fly at half-mast.”

However, the company did say that it’s “not a question of whether it will open source Java, but how.” I can answer that: “Very soon, with minimal restrictions.” You’re welcome.

A new project - reasonably far along - just popped up on the SugarForge: Alfresco/SugarCRM integration. I’ve been looking forward to this for some time: the best of open source CRM combined with the best of open source ECM.

We have a few mutual partners working on the project now. I’d highly recommend you take a look, and hopefully get involved.