Comparing NetSuite and Salesforce

I was at a user meeting with NetSuite in Boston earlier this week.  The company has bought two companies since going public — OpenAir and QuickArrow — both of which support the professional services market.  Companies sell things as well as services and CRM has been applied most successfully to the former.  Companies that sell services have been left to their own devices to figure out how to automate and manage sales and delivery and their situation resembles that of the thing-sellers pre-SFA.

NetSuite’s idea is an integrated solution combining ERP and services oriented planning and sales modules going by the name of SRP — services resource planning — and the idea has legs.

As you can imagine there are some significant differences between selling things and selling projects.  Most importantly, services companies have bigger issues with fixed overhead because you have to have smart people on staff if you expect to sell their time.  Economists might say that supply is inelastic or certainly less elastic for the services guys than for the companies that can throttle up or down the manufacturing process.

All this got me thinking not about the two different types of selling but about the two different styles of building a company exhibited by Salesforce.com and NetSuite.  Both companies have purchased other companies when it made sense as a way to build out their offerings.  But each company also has a multi-tenant architecture and a cloud platform, which makes it easy for third parties to build or modify applications.  Nonetheless, if I had to describe each company’s strategy I would say that NetSuite is more likely to buy than make compared to Salesforce — if you include the partners.

Salesforce appears to have decided on an approach that encourages a partner community to build native applications while NetSuite seems to encourage partners to deploy and modify its core solutions though not necessarily build wholly new ones.

Now, this is a rough approximation and it looks more black and white than it is — there is a lot of grey area in all this.  But it drives an interesting question that I believe can’t be answered, at least not now.  Which approach is better?  Should the primary vendor be the only one involved in new product development or should the platform vendor simply let a thousand flowers bloom?  Certainly the existence of the platform makes the second option possible.

Part of the answer can be found in how each vendor views itself.  Salesforce is obviously looking for a big new market to penetrate that’s bigger than CRM and it has selected application development tools for the enterprise and smaller organizations.  NetSuite might have a serviceable platform but for the time being it appears to be more interested in the market for integrated front and back office applications, which is more crowded.

I don’t have any good answers here or prognostications, just these observations.  Salesforce has always been in the business of inventing the future and while they’ve been successful they have had their stumbles along the way too.  Other companies have been content to stick to their knitting, but the future rarely keeps to a script.  There are many markets just opening up, at least in part because there is reliable and low cost software available to support them and that says good things for both companies’ chances.

The big question to ponder is whether there is enough demand for in-house development to support Salesforce’s vision.  It groups are notoriously backlogged and it is unclear to me if the backlog is a result of too much demand or inefficient tools.  For decades we have argued that it is the tools and we’ve seen generation after generation of tools that promised to fix the problem.

Tools are important but if you read “The Black Swan,” which I recommend, you might get the notion that backlogs are inherent in what we do, in part because we do such a poor job of understanding and planning for future requirements.  If so, one of the next logical acquisitions for either Salesforce or NetSuite should be a company that focuses on improving forecasting and planning methods.  Does such an animal even exist?

Adobe on a roll (hold the fries)

For the second day in a row Adobe made an important partnering announcement.  Yesterday the company said it had teamed with Salesforce.com to produce Adobe Flash Builder for Force.com, which will speed development of Flash-based user interfaces for Salesforce customers.  Today Adobe announced that it has concluded acquisition of Omniture, a web analytics vendor based in Orem, Utah for a whopping $1.8 billion.

It seems an obvious strategy to leverage some of Abobe’s ingredient technologies, like Flash, to make a bigger presence for itself in Cloud Computing.  The addition of web analytics is very interesting.

At this point in the evolution of CRM, if you are not already a big player the chances of starting from scratch and getting big are nugatory so the strategy has to be to buy.  But Adobe’s choice of partnering with a leading CRM company for user interface design and following up with buying analytics is intriguing.  With these two ingredient technologies, Adobe appears to be 1) betting on the future importance of understanding customer moves and motivations and 2) clearly understanding that robust simplicity must rule all software interfaces regardless of platform.

If you ask me, these are two good bets.  While there are clearly many good analytics products on the market either freestanding or embedded in business applications, my research tells me that regular users are still too confused about analytics to fully leverage them.  Ask ten people in our industry the difference between reporting and analytics and you will see what I mean.

My quibble with analytics and analytics vendors generally is that few acknowledge the effort required to capture good data.  Too often the MO is to capture large samples and get some averages, a good but not great approach that, in another setting, once left a bemused Benjamin Disraeli to list three categories of lies, “Lies, damn lies and statistics”.  There’s no substitute for understanding demographics, biases, attitudes and the like to better predict behavior.  Here’s hoping that Adobe gets it and uses Omniture to go the more rigorous route.

Fusion applications decide an argument, sort of

With Oracle’s announcement of Fusion applications, you can make a reasonable case that Salesforce.com has won an important ten-year old argument about the future of the software industry.  Notwithstanding SAP, the only significant outlier left, Oracle is the last major software company to adopt on-demand computing as a centerpiece thus awarding legitimacy and critical mass once and for all to the idea.

Amen.

But the Oracle announcement says more about business models than technology paradigms and at the model level it is not clear that Salesforce has won.  Salesforce CEO Marc Benioff’s vision of business applications delivered over the internet has won an important victory but the business model — subscription services — that makes this technology the center of the movement and exclusive delivery mechanism has not eliminated all competition.  Not yet, at any rate.

The reasons are simple enough, market reticence generated by concerns — both real and imaginary — about security or the viability of the technology model still hamper full adoption of the business model.  As a result, companies as diverse as Oracle, Microsoft and Sage have hedged their bets by offering technology that can be implemented in numerous ways including on-demand as well as by conventional deployments.  As a result vendors have effectively thrown the business model decision over the wall to the customer.

With software capable of, shall we say, polymorphous deployment, the ultimate decision about how to deploy now becomes the exclusive province of the customer as the vendors have now turned into Solomon or, in a modern interpretation, Burger King.  Customers are completely free to have it their way or ways.  They can deploy business applications in a fully SaaS configuration or in hybrid ways that are to a lesser extent owned and operated by the IT department.  As I have noted before, this is typical transition state behavior of vendors straddling two diverse paradigms.

It is no surprise that adoption of the business model lags adoption of the technology.  It has always been true that conversion from traditional software licensing to SaaS is a big step and one that for many software companies could lead to financial ruin if not handled expertly.  More to the point, there are customers who, for reasons of security, custom and preference believe that SaaS computing is not for them, at least not now.

So it is no surprise therefore that the most successful SaaS companies are those that, like Salesforce, grew organically from on-demand roots.  Other successful SaaS companies like Oracle bought their way into SaaS computing, a time honored tradition when adopting new models.

Even before Oracle’s Fusion announcement at Open World this month, the company had been a player in SaaS based CRM with Oracle CRM On-Demand due to its earlier acquisition of Siebel Systems.  But it remains to be seen if any software vendor can fully realize the benefits of SaaS — and now Cloud Computing without full emersion into the technology model.

One of the most powerful aspects of SaaS computing is not the idea of subscriptions or even Internet delivery but of a single version of the applications supporting all users.  With a single version of the code, all users have the same foundation on which to configure, modify and build new applications.  The single code set — also called multi-tenant architecture — makes it hugely unlikely that any two independent software makers would develop incompatible applications and therein lays the power of the business model.

This single idea makes it highly likely that applications built to the standards of the foundation — or platform as we like to call it — will be able to inter-operate.  Take this standard away and you have the same Babel of competing standards and proprietary designs that have been the bain of the software industry.  There is a cost associated with this lack of standardization and software customers have been paying it for decades — with rising resentment.

That cost is not measured strictly monetarily; there is opportunity cost involved too.  When everyone played by the same conventional software rules the opportunity cost problem was equivalent to a farmer experiencing bad weather.  But SaaS computing eliminates the weather variable giving a big advantage to companies under its umbrella.  So it is ironic that the decision about adoption is still left to taste.

With most of the hybrid products, the same code can be deployed in a conventional multi-tenant way or as a standalone system behind a traditional firewall.  The segregated system becomes a unique instance the moment a developer modifies the platform.  Doing that makes the idea of standards a waste of time.

But for the time being — and I am still calling it a transition state — we can expect to see a lot of deployments in which the software is SaaS ready but the deployment is decidedly twentieth century.  In the next five to ten years we will see examples of companies trying to back out of their proprietary SaaS-like systems to finally get on board with SaaS or Cloud Computing.  It will all have been avoidable and it will be good business for software consultants.

As Kurt would say, “So it goes.”

Oracle Fusion in (financial) Context

With the introduction of Fusion applications Oracle has joined the cloud community.  You might want to argue that the company has been involved in the cloud for many years as one of the key technology underpinnings of many of the biggest SaaS companies.  That was one of Larry Ellison’s big points at the Churchill Club.  Cloud computing still needs a ground station to serve it and Oracle has been at the top of that market for a long time.  For example, salesforce.com uses the Oracle database to support its service and many other companies do too.

Because so many software company founders and executives got their starts at Oracle it was a natural for them to build on what they knew and they knew Oracle.  That may not sound like much if you stop thinking about it at the database but the community of Oracle people present and past gave many a start-up the intangible resources they needed to be successful.  Knowing who a product manager is or what is generally on a product roadmap can be useful information.  I am not saying that any legal lines are ever crossed but having been an insider and understanding the culture as well as the technology can be very helpful.

At any rate Oracle is, or will be with the full release of Fusion Applications, a member of the Cloud Club.  That leaves stragglers like SAP to still deal with the conversion and I have a notion that the pace will only accelerate from here and I expect that if cloud computing is not the dominant paradigm today it will be in the not too distant future.

Unlike salesforce.com and other companies that entered the market and developed their products as native SaaS applications, legacy companies had a great deal of work to do to convert to the new paradigm.  Changing a technology paradigm is never easy or cheap — the last time we did this we went form mainframes to client-server or was it client-server to thin clients?  But changing the technology paradigm is only half the battle.  The other, harder and more intense job is changing the business model.

Companies do not always have to change their business models when they change software paradigms.  As a matter of fact I am not sure if the majority of software companies today ever had to make a business model shift before the advent of cloud computing.  But there’s really no option at this point if a company wants to get to the cloud.

Oracle’s shift to Fusion, like many other vendors’ shifts to the cloud, buffers the business model change by effectively breaking the shift into two hard but more digestible bites.  Oracle is fortunate in that it has a large number of products and it is acquiring more all the time.  As a financial exercise, when Oracle begins selling fusion applications, at least a part of the revenue will come in the form of subscriptions and I expect that subscription revenue will grow at the expense of traditional licensing over time.

Because Oracle offers multiple products including hardware (Exadata storage systems — and servers if the Sun merger is approved) the company’s financial results will be insulated from a big hit — something that pure software companies dread when they change models, especially if they are publicly held.

So, I am not a financial analyst, but my reading of the situation is that Oracle has positioned itself and its business well with Fusion.  When delivered in a service mode, the applications will be competitive with other market leaders.  The applications ought to find a big market in other nations where software might be too expensive for local tastes.  And there at least multi-tenancy should prove to be essential.

Meanwhile, the conventional license business should continue as a viable option for customers who prefer that mode for as long as those customers want it.  Converting from the license paradigm to a more or less pure cloud paradigm will be a business decision and the required development will consist of rewriting contracts, building web sites and adjusting marketing.

This isn’t all guaranteed to happen but it looks like a promising and logical evolution.  Of course, I am reading “The Black Swan” right now and I am wondering about something highly improbable affecting this rosy scenario.

Key Findings for OOW #3 — The Fusion Edition

Fusion is big, potentially powerful, based on new technology, backward compatible and not available yet.

Larry Ellison began taking the wraps off Fusion at his Open World keynote on Wednesday.  His appearance on Sunday with Sun CEO Scott McNealy was just for poking some fun at IBM, this was the real deal.  Fusion is a big idea and this post will leave something out – that’s a given – but here are some important impressions.

There are ten applications and I don’t write fast enough to have copied them all down from Larry’s slide.  For sure there was CRM but also GL, Deal Management, Territory and Talent Management too.

The applications are based on SOA architecture, the UIs have embedded BI.  There are six thousand database tables, 6,500 objects, 20,000 views, 10,000 task flows and the applications are code complete and being tested by customers.

Fusion is based on industry standards like JAVA Fusion middleware which Larry said is the first such deployment.

Fusion applications are scheduled to debut next year and while that is a little disappointing it is entirely in keeping with the purpose of a keynote – forward looking statements right?  Ok.

I saw some demos but as I said in a previous post, I want to see a birth certificate.

Fusion applications are modular and they are designed to be deployed as full replacements for other older Oracle products.  Modularity enables them to also work side by side with existing applications so that there is no need for a wholesale replacement.  There are also new applications that have no analogs with the older product suites so it is good that Fusion and legacy applications can work together.

Like a lot of CRM products coming out these days, the Fusion applications, based on a SOA architecture are intended to operate behind your firewall in a single tenant manner or at some other data center in either single or multi-tenant mode.  This approach neatly straddles the diverse deployment options that some people feel they need today and gives a company like Oracle the flexibility to support all of them with one code set.  This neatly solves the problem of how to convert Oracle’s product set from premise-bound to cloud resident by leaving the decision to the customer.  That’s good, fine even and it does a lot to close the discussion about on-premise vs. on-demand, or does it?

The trouble with running a private cloud is that as soon as I make a modification to the system I might be making the product unique and unsupportable putting me back into the same version conundrum that many hope to avoid.  I need to know more about this.

Interestingly, in no demo did anyone talk about Fusion code or coding beyond Larry’s statement about JAVA.  I suspect that is not because you can’t get into coding some arcane part of your application but I hope coding is infrequent and at a level of abstraction sufficiently removed from the guts of the operation to make it possible to have one version of code for the whole planet.

The Fusion applications, specifically the CRM stuff, are compatible with Oracle’s Social CRM gadgets and widgets and I expect that it will offer fairly robust support for enterprise computing when the applications hit the street.

The UI looks nice.  I don’t know what the technology is that supports it but it has an Adobe Flex look.  Nice job on that – it will give all those Gen Y people coming into sales and other front office disciplines a feeling that they are using something as modern as the games they play on the home computer.

That’s about all I feel qualified to say.  We need to see more but for now it is very good to see Oracle redeeming a promise it made a few years ago when it went on a buying spree in the front office market.

Key Findings for OOW #2

This time a little more serious.  Ok?

I am a software and CRM guy so that’s the focus of this piece.  Much of Oracle Open World (OOW) was table setting.  It was all interesting and valuable but it was also a lot of independent data points.  There was lots of cool CRM introduced for sure but it all lacked a certain coherence until the Fusion announcements.  That’s not a bad thing by the way, just a reflection of the maturity level of so many products.  Some interesting stuff:

Support for retailers and anyone who wants to support employees who are directly involved with customers.  For example, handheld applications that bring customer information to devices in support of sales clerks.  Not just the customer’s size and account numbers but things like intelligent offers and any loyalty points accumulated so that a pricing negotiation can happen on the spot.  The same technology supports users on their mobile devices to do things like buy train tickets and other self-service purchases.  This was not really new for Oracle but it has been improved and it looks better with each iteration.

I have several problems with the self-service example though.  The demo was from the Swedish Rail Service and it showed how customers can buy travel packages and trade in frequency travel points –  you get the idea.  But there’s no infrastructure for this in the U.S. of A.  For this application to work we need a high-speed rail infrastructure and that will take about ten years (LOL).  Good thing Oracle is looking at foreign markets.

I attended the afternoon half of a Chief Sales Officer Executive Summit or some such thing on Wednesday afternoon.  It was very good.  Lots of high-octane sales executives from billion-dollar (or equivalent) companies talking about their success with Oracle-Siebel and Oracle CRM On-Demand and their successes.  They had an economist give a quick analysis of the world economy and for a practitioner of the dismal science, he sounded upbeat.  I regard this as a contra indicator of something, just not sure what.

No surprises there in the following sense.  CRM and SFA are pretty mature, the biggest gains we are likely to see going forward will likely be in more enlightened use of the products by the high-octane talent.  I am not optimistic in the short term for the following reasons.

Anthony Lye did a good job presenting the state of the union in SFA mediated selling today.  Most people still use spreadsheets to make forecasts — letting error and unpredictability enter every time a sales manager decides to spiff up the data or apply a fudge factor.  Lye’s slides showed only a tiny fraction of users had forecasts that were worth anything (i.e. accuracy rating of 90% or better) yet we keep messing with the data.

I have an idea.  Rather than futzing with the data in spreadsheets, let the forecasts stand as they are and make accuracy a criterion for sales compensation.  Then stand back and watch things improve.  Short-term pain for long-term gain.

Ok, Oracle showed plenty of good technology (including Fusion apps as part of some demos) like that described above, for B2C segments and even more for B2B but the thing you come away with is that Oracle is really focused on the enterprise.  Oracle has a lot for the mid-market user – it’s more about what you don’t buy in that case – but they really groove on the sophisticated and complex selling that goes on in billion dollar companies.  They have apps that solve problems that mid-market companies might not even encounter.

For example, the Oracle deal management application takes a lot of data about what a customer has bought already, their price tolerance, the value of a deal, what other companies might be paying for similar deals and derives a statistically relevant price for the deal.  This happens automagically after the sales representative has fed a few data items into the system.  Management, meanwhile, had set a few parameters in the system and out pops a price that the boss will, if not love, then certainly tolerate.

Smaller companies don’t do this kind of thing much.  They’re focused on closing as much business as they can at the end of the quarter.  Too bad, because Lye’s slide deck included one chart that showed the 30% of companies knew they were leaving money on the table – but they were not sure how much they were leaving.  Maybe there is broad applicability for deal management.  Ya think?

My impression of the related sales applications is that they were best for companies that sell to managed accounts.  You know what I mean, two companies have long term relationships and sell parts, components, raw food and the like to a large group of repeat customers.  That’s not the only kind of business they do but it’s significant and the Oracle apps work well in that large environment.

Oracle also has some apps that I find more interesting for things like territory planning and while they got some attention in the executive summit I could have used to see more.  My favorite is something that helps find the “white space” in a territory.  I wasn’t familiar with the term white space until this week but now I can use it almost as much as the Governator says technology.

What white space refers to is the knowledge that a territory acts like a buffer.  It can absorb only so much before it gets saturated (think of a kitchen sponge here).  Once the buffer is saturated it won’t absorb more so it’s smart to know at the start of a year how full your territory is – especially if you plan to assign quotas that your people have a realistic shot at making.  Ok, the white space application helps the manager figure this out by taking account of things like the target density in the territory, what’s already been sold, the average sales cycle, price points and other relevant information.  That’s cool and my only critique of this and other cool stuff is that Oracle didn’t show it off enough for my taste.

These are some of the applications that separate Oracle from other SFA vendors and they represent some true differentiation.  I hope they do more with these apps.

I was going to write about Fusion now but this post is getting long so the next post will be about Fusion.

Benioff Attends Open World

Well, that was interesting.  Salesforce.com CEO Marc Benioff just completed a speech at Oracle Open World, perhaps the ultimate example of co-opetition.  It was apparent to me early on that Marc’s purpose for being there was to refute Larry Ellison’s rant at the Churchill Club in which he compared Cloud Computing to vapor.

I got a hint earlier today when someone said that the event had been put together quickly, which to me confirmed the need to refute Larry who spoke a couple of weeks ago at Churchill.  Benioff started out by graciously telling the audience that he had worked at Oracle and attended Open World many times in his 13-year career, even presenting on the same stage he was now on.  He went further pointing out that the Oracle database is one of the key components of the Salesforce service and thanked Oracle executives for the graciousness.

But there was little doubt in my mind that Benioff felt he needed to refute Ellison’s off the cuff assertions at the Churchill Club.  He did that with ease and just when you might have thought he’d reached the end of his talk, he brought up the CIO of EMC Corporation Sanjay Mirchandani to discuss that company’s hybrid CRM approach that includes Salesforce and Oracle for on-premise CRM.  It was almost as if he wanted to say that Salesforce can play the hybrid game as well as Oracle.

I guess the Open World setting proved too much of a temptation for Benioff.  It’s in the same city as Benioff’s office.  The venue was easy to get, Michael Dell another big Salesforce customer spoke at Open World this morning and was available to be on stage with Benioff for part of the afternoon.

There’s little doubt that Benioff was able to refute Ellison but the bigger question for me was why he felt he needed to.  We haven’t seen this kind of action for many years and it makes for lively times in these challenging days.

At Oracle Open World

Open World most resembles Forrest Gump’s box of chocolates in that there is such variety that you never know what to expect.  At any moment there is equal probability that you will be dazzled, challenged, delighted and perplexed.

This being journalism, perplexity reins as a dominant topic and perhaps the most perplexing thing about the meeting is the show floor which includes large booths from the heavyweights in the industry a.k.a. Oracle’s greatest competition and greatest customers, for example, SAP and Microsoft.  Salesforce.com’s booth sits long and narrow moored on the show floor like an aircraft carrier in a crowded harbor.

By the time most of you read this Marc Benioff will have spoken and we will at last have an answer to the question haunting the halls of the Moscone Center.  Why would Benioff speak at Open World, the user meeting of one of his staunchest competitors?

You can make all of the arguments you want about how Salesforce relies on the Oracle database to serve its millions of customers, you can invoke arcane game theory to explain this apparent cooperation among competitors if you like – after all the Nobel Prize in Economics was just awarded to two social scientists who studied this phenomenon.  Still you are left with an irreducible Why?

Benioff speaks at one today and may have an answer.

In CRM kudos have go to Anthony Lye and his team for their top to tail work with the Siebel and CRM On-Demand suites and the dogged determination to prove the necessity – even desirability – of hybrid premise-based and on-demand approaches to CRM.  I will not digress into a discussion of my oft repeated belief that this is a transition state on the way to full Cloud Computing in deference to my hosts and I only wish they would give up the sophomoric assertion that cloud computing is simply vapor.

The CRM team is bristling with innovations for large and small customers –announcing twelve new products, eighty customer driven enhancements, thirty-one new features, a REST API, CRM availability in Microsoft Outlook, and a new Siebel version coming this year.  I think there’s more but maybe my note taking is not so good.

Larry Ellison spoke on Sunday night — a cameo in Scott McNealy’s keynote.  Ellison made the expected and highly believable statements that rather than letting Sun sink into the, uhh sunset, once the merger is completed, Oracle would increase its investments in Sun systems beyond the hefty investments that Sun had been making.

Oracle’s stewardship of PeopleSoft, J.D. Edwards, Siebel and fifty-five other acquisitions (according to Safra Catz) provide the needed street cred here.  Ellison even had fun poking IBM about an internal program they call Sunset reminding all that one man’s sunset was another’s sunrise.  He then proceeded to announce significant benchmark superiority over Big Blue.  Some things don’t change, benchmark competition is one of them.

But Sunday was McNealy’s time to shine.  The justifiably proud Sun CEO rattled off a slew of Sun’s leading innovations in CPUs, memory and file management, operating systems, and, of course, JAVA.  Many of us forget how many devices run on JAVA code — without any “JAVA inside” branding — but it’s a lot and McNealy was happy to provide a glimpse.

Ellison will speak on Wednesday to conclude the meeting and my contacts keep telling me that my questions such as those about integrating the sprawling software suite will gain clarity then.  We’ll see.

Perhaps the most interesting moment of the show for me so far came on Sunday at the end of McNealy’s speech.  He showed a slide meant to sum up his experience at Sun as well as the operating philosophy the company has been run by.  The slide said we (Sun),

Kicked butt

Had fun

Didn’t cheat

Loved our customers and

Made money

(I am not a hundred percent on the last bullet, note taking again.)

McNealy concluded by saying of the merger of Sun and Oracle, “Larry’s going to like his new toy.”  The statement immediately put me in mind of Newton’s famous summation of his own career when he said near the end of his life:

“I do not know what I may appear to the world, but to myself I seem to have been only like a boy playing on the sea-shore, and diverting myself in now and then finding a smoother pebble or a prettier shell than ordinary, whilst the great ocean of truth lay all undiscovered before me.

I can’t think of a better description of why these very bright people work so hard to make electrons dance.  Sure, it’s profitable but at the end of the day it’s even better if the ride has been fun.

Both sides now

As luck would have it, I knew nothing about Larry Ellison’s rant at the Churchill Club on September 21 about cloud computing when I wrote last week’s piece on cloud computing.  I saw it on YouTube.

You have to admit that Larry is a heck of a showman and the video is fun to watch.  But whenever someone in that kind of situation starts to nit pick over definitions it says to me that they’re hoping no one will notice that they might be a bit threatened by a next generation technology.

Various analysts have pointed out the cloud is really an amalgamation of several technologies including Software as a Service (SaaS), Platform as a Service (PaaS) and Infrastructure as a Service (IaaS).  I can’t disagree with any of that though it is a rather mechanistic summation, but my reaction to all this parsing of nomenclature is, so what.

Say what?

Really, so what.

My take on cloud computing is different, it’s part cultural and part historic and the two are twisted tightly together.  The history of the technology industry is one of overcoming shortages and cloud computing is part of that progression.

When storage and memory were in relative short supply we kept data on tape and programs on punch cards.  We responded by building more and denser capacity until the shortage went away and we found ourselves with extra capacity at greatly lower costs.  We had capacity to spare and instead of wasting it, clever individuals built the relational database.

Computing power has always been short but once we got into Moore’s Law the curve kept delivering new capacity to burn every eighteen months or so.  Did we waste it?  Some did, but enterprising souls built the graphical user interface of GUI and graphics cards and they changed computing.

Along the way as these improvements became ubiquitous and cheap — then something magical happened.  We started branching out and the improvements didn’t stay in the confines of enterprise computing.  They leaked out into a plethora of unimaginable products.  Consider the iPod, for example, nothing but a tiny computer with a baby-sized disk.  Talk about unintended consequences, I know some people were thinking about these devices but there weren’t that many and look at the effect this one device has had on our culture.

iPod is only the best known example but there are many others.  For example, our cars are now bristling with processors and memory for fuel injection, ABS brakes, navigation systems (probably with a DVD in your trunk) and more.

But we are not done.

Then it was bandwidth.  Ethernet was an interesting standard but it was slow.  Bright minds turned their attention to networking, changed the way our applications use networks so that they use less bandwidth and others made more of it available culminating in the ultimate consumer bandwidth, the Internet.  Tell me that hasn’t changed your life.

What’s different today is that we have a relative abundance of everything.  If it comes to us as a service we no longer think about all of the provisioning and cost that was once part of our calculus.  Platforms and applications coming in as a service ensure that bright minds can dream big and who knows what they’ll come up with?

So far all of that abundance has enabled us to build not artificial intelligence, that appears to still be in the future, but artificial, or should I say synthetic, relationships.  I mean social media here.  The relationships we maintain through social media are available to us because the cost of maintaining them in time and effort and real dollars is so low that it might as well be free.  And the power of these relationships is that at any time, for the right reason, they can become full blown active relationships that provide friendship, information, help or anything you might expect of a friendship.  And, of course, physical distance is much less of an issue than ever.

Our advances in technology have been the sometimes-surprising results of efforts to overcome scarcity and other adversities.  Social media is a child of the ubiquity delivered through the Internet and it is still in its precocious early years.  And logically, social media is not the only advance we should expect from this ubiquity, though I can’t say what the next things will be.

So when I hear someone, even kiddingly, say that cloud computing is nothing new or that it’s just water vapor, it makes me think that someone isn’t getting it.  We only have a few faint ideas about what cloud computing will ultimately yield and that alone is why it is so important.

What is on-demand?

I have an on-going conversation with an industry executive about the nature of on-demand, SaaS and Cloud Computing.  The central question is what is it exactly?  Our conversation is always thought provoking and I come away from it with at least some additional perspective.  Some times I think the discussion is incredibly philosophical like the question about a tree falling in the forest — does it make a sound if no one is there to witness it?

I was always flummoxed by the tree question.  Of course it makes a sound I thought.  But then my kids, who are musicians, explained to me that sound is the perception of certain energy waves emitted by the falling tree.  Kids!  When the tree falls it certainly makes those waves but unless an ear and therefore a person or some other animal is there to hear it the energy is never converted into sound.

So my conversation with the executive goes like that.  Is SaaS just application hosting with delivery through the Internet?  Is it SaaS if the hosting service is completely within the company?  Can a company with such an infrastructure support remote sites that way and claim to have its own cloud?

These are all good questions and they are issues that we are still dealing with.  The SaaS and Cloud Computing market is still so new that there are many issues like this that have not been nailed down yet to everyone’s satisfaction and they may never be.

It is my belief that you can’t separate SaaS or Cloud Computing from a multi-tenant architecture.  Over the years, and as I have written before, the difference between SaaS computing and conventional applications that are hosted have dwindled to a small core largely consisting of the multi-tenant issue.  Initially, conventional applications were client-server and hosting them meant using virtual private networks and high server overhead.  They were definitely not SaaS.

Vendors have done a good job of bringing those client-server applications into the Internet age — the user interfaces run in browsers like SaaS applications do, and some even offer the capability of multi-tenancy.  Some vendors have become good at offering a choice of hosting options to customers, so are then truly SaaS?  Are they candidates for Cloud Computing?

The answer is complicated, like the issue of sound and the falling tree.  First things first.  As long as a customer has the option of multi-tenancy then I think it’s not possible to call a solution SaaS or Cloud Computing.  In an optional setting like that the vendor still has to manage and maintain multiple versions of the application and with that comes all of the overhead and complexity of conventional computing.

Ironically, a vendor who offers the same software as both single-tenant and multi-tenant instances straddles definitions.  A customer using that software as a service in a multi-tenant mode is using a SaaS solution.  But a customer using the same software tucked behind another company’s firewall in single tenant mode is simply using a conventional solution and the same can be said of a company using a single-tenant solution in some other data center — that’s just facilities management.

As for Cloud Computing, I think it’s not possible for a company to have a private cloud.  A private cloud is like an old cell phone gathering dust on a shelf somewhere.  The very idea of the cloud is of a network, a communally accessed resource for accomplishing a growing list of personal and business pursuits.

I suppose a company could have and make good use of a private section of the cloud but there is a big difference here.  Private clouds imply many incompatible little networks and what good is that?  Recall that Metcalfe’s law says that the value of a network is proportional to the square of the number of connected users.

But I think the best way to come to terms with SaaS and conventional computing is through the business model.  True multi-tenant SaaS leaves the user completely unconcerned about the nitty-gritty of system ownership — the licenses, the versions, the compatibility, the hardware capacities — all of that is hidden and immaterial to user decisions about provisioning.

The multi-tenant SaaS business model is simpler and much less costly and for that reason it is catching on globally, especially in places where infrastructure costs are simply not affordable.  More importantly, though, it’s clear that multi-tenancy is the new standard because the business model is a better fit for the times and for the growing world market.

I expect that single-tenant solutions that look very SaaS-like will be around for a long time for several reasons.  Most importantly, there is customer demand — some customers are still not on board with the idea of multi-tenancy either for personal reasons or because of various restrictions.  Second, many vendors have not yet converted their technology and so they will continue selling what they have.  Third, some vendors’ business models can’t stand the strain of converting — it may still be too early to try to convince investors that a business model shift makes sense.

Business model conversion may prove to be the biggest obstacle.