My new position is with Sterling Commerce in Dublin, Ohio, where I am now the VP of product management for the B2B Collaboration product line.  If you don’t know Sterling, check out our pitch at www.stercomm.com.  I am really excited to work in a different part of the software universe – to see the industry and customer needs from a different perspective -- and especially to work for a company that is a strong and recognized leader in its space (as BMC is in its space). 

Sterling and BMC are similar in many ways, and different in many ways.  The similarities are a bit scary – war stories that sound exactly the same except for the names of the products; people in the same job function who appear to have the same personality (even the same name!) as their counterpart at the other company; successful products; large numbers of happy customers; and plans to change and grow in significant ways.  The biggest differences have to do with the markets each company serves – BMC’s goal is to optimize complex IT systems (people, process and technology) to provide the greatest business benefit; Sterling’s goal is to help customers and their trading partners maximize business performance by optimizing complex data exchange and process interactions (like supply chain and other information coordination challenges) both within and between enterprises. 

There are a lot of similarities between systems management and B2B collaboration, and I have been thinking about lessons learned in one space that may apply in the other.  Do the ITIL processes that Service Management solutions are based on resemble a process framework for the (internal or external) “IT supply chain”?  Might multi-enterprise collaboration tools or even true supply chain apps apply here – to enable IT services delivered and coordinated among multiple providers?  The big problem for IT asset management solutions and CMDBs (how to arrive at an accurate “single source of truth” describing IT assets) has a lot in common with the Global Data Synch problem retail supply chains deal with every day.  Identity management and trading partner management seem very similar; so do distributed database recovery and supply chain disaster recovery.  Why do IT assets (even software) not have “barcodes” (so to speak)?  Or maybe shipping containers should have MAC addresses.  Both companies even use the phrase “Business Impact Management” or BIM – at Sterling, this means analyzing the impact on a business or supply chain of a delay or failure of one “link” in the chain; at BMC this means analyzing the impact on an application or extended business service of a performance problem or outage of one IT node / element.  Maybe IT is just one flavor of the B2B or supply chain problem space … as Mark Stabler might say, “This is interesting”. 

Best wishes and best of luck to all my friends at BMC.  In this big / small industry, we will probably meet again.  If you are ever in Columbus, I will buy you a bowl of chili entirely different than you can get in Texas :^).  I hope to have my Sterling blog humming soon.  You can contact me at chris_johnson@stercomm.com.

A new blog on the same topic will be started on TalkBMC by Jay Gardner.  Wiley Vasquez will also start a blog focused on the technology behind on demand software.  Enjoy those.  Thanks!

Today's example: dry cleaning.  The first dry cleaning business was established in France in the 1840's, and the business has not changed much in the subsequent 160 years.  The typical dry cleaner is a large retail store front, with complex equipment and lots of storage space (since they have to keep 1000's of customers' clothes on hand) and a 2-3 day cycle time to clean & return your clothes.  Although the method of cleaning has changed over the years - switching from dangerous chemicals like kerosene and benzene to more friendly stuff like carbon dioxide - the basic business model has not changed too often.

One significant change was a shift from on-premises cleaning to off-premises.  In the prior model, many dry cleaners had the equipment and chemicals in the storefront or nearby, and did the actual cleaning work themselves.  Although this allowed the fastest turnaround time (I once had a shirt "emergency dry cleaned" in only 10 minutes) the economics were wrong - picture 100 dry cleaners in a city, each with a large capital expense for buying the equipment, each needing to train people to use the machinery and chemicals safely, special permits from the city, expensive insurance policies ... a difficult situation.

A generation ago, most dry cleaners switched to being storefronts only, and sent the clothes to a cleaning plant somewhere across town to be cleaned and returned to the retail location later.  This was a classic "economy of scale" play and has some parallels to SAAS.  You can make a good living as a dry cleaning plant - owning, operating and maintaining your own equipment & specialist staff - but only if you operate at volume (preferably 24 x 7).  If you aren't prepared to operate in this highly scalable and efficient way, you are better off buying the service from someone who is already doing it. You can succeed either way, but need to know which role you want to play in the dry cleaning supply chain.

How this relates to software - if a software title is available from a large-scale service provider (including vendors who offer their own software in a SAAS model), they can sell you the service for less than you can do it for yourself, if you consider the total cost of ownership.  Be careful to not look at just the price of the software (like $1.00 per shirt cleaned) while ignoring the surrounding costs (your time, gas to and from the dry cleaners, time delay etc.)  If you take a dispassionate look at what it costs you - in real dollars and opportunity cost - to run your own software, the value of SAAS becomes obvious.

More recently, some dry cleaners have changed again by eliminating the retail storefront altogether, by picking up and dropping off at your front door.  One example, the company my family uses, is 1-800-DryClean.  When I first heard of this service, I was elated.  Although our previous dry cleaner did a great job, the logistics of getting clothes to and from them were tough.  I travel a lot for work, and sometimes work odd hours, so I often could not get to the dry cleaner for many days.  When I was able to visit, I had to park, get in & out of the car, maybe walk thought the rain, stand in line ...

When I heard that 1-800-DryClean would pick up & drop off at our home, I thought "wow, but how much will that cost". Surprise: it costs less (sometimes much less) to have your dry cleaning hand-delivered to your front door. That is counterintuitive and kind of amazing. Think about it - in the traditional model, you are doing half the work and paying more for the privilege.  How can a dry cleaner do more for you (delivery) but charge less ???

The answer is pretty simple - by eliminating other expenses.  The dry cleaner no longer needs an expensive retail storefront, just cheap office space in an industrial park somewhere and a few delivery trucks. By rethinking who-does-what, and questioning some basic assumptions of their industry, 1-800-DryClean offers more, charges less, makes me happier and (presumably) makes more money in the process.

How this relates to software - SAAS allows vendors to reduce their expenses in many categories.  For example, in the traditional model, software teams need to design, develop, test and release products for many different environments - typically several operating systems (Windows, Solaris, AIX, HPUX, different flavors of Llnux ...) and different versions of each; several different database management systems (Oracle, SQL Server, DB2, Sybase, MySQL ...) and different versions of each; and multiple combinations of these (Oracle 9 on Windows 2000?  Oracle 8 on Solaris 9?) etc.  This might have made sense in the old model where every customer made their own "equipment" purchasing decisions, but in a service-delivered world, the software vendor can choose just one platform and use it for every customer - since the customer doesn't see or care about what is back at the "dry cleaning plant".  All the customer cares about is, clothes cleaned and back on time (or software that works from the end-user's perspective) and a reaonable price.

The same is true in many other categories.  No tech support required for old versions of software - because there are none (just one hosted version used by all customers). No need for huge test labs and software testing mechanisms - because you have 100X fewer environments to test.  Which leads to fewer bugs - happier customers and less tech support.  I could go on and on ...

That's the challenge facing the software industry.  How can we do more (useful) things for customers while reducing (non-value-added) expenses?  Can we charge less and make more money in the process?  I think we can, but not by working harder - we need a new business model - in my mind, a combination of SAAS and managed services.  Those who understand this will be able "clean up" (sorry) in the market.

I love this question for several reasons - (a) it is so colorful and misleading! (b) it reveals a lot about the mindset of traditional software companies; and (c) in the end it does more to point out the advantages of the SAAS model than any perceived downside.  Let's break this down.

The question is misleading because it assumes moving customers to a SAAS style of delivery is a bad thing!  The word "cannibalization" gives everyone a mental picture of explorers wearing pith helmets in a giant pot of boiling water, and naturally you picture yourself as an explorer. A better way to ask the same question is, "How many customers (and which specific customers) will prefer SAAS or traditional products, and how is this potentially good or bad for us, and how should we respond?"

Which goes a long way to illustrate the mindset of traditional software companies.  If you have ever worked in an industry that has gone through a period of fundamental change, you will know what I am talking about - when a big change appears on the horizon, the typical reaction is to first ignore it; then deny it; then denigrate it; then worry about it; then respond to it. I think the "cannibalization" idea is somewhere between denigration and worrying on this spectrum of responses.  This makes me think of Detroit automakers who ignored Japanese cars in the 1970's, or telecom companies who sat on packet-switching technology (which they actually invented) rather than upset their circuit-switching empire. Good examples of what NOT to do.

I believe SAAS solves many more problems than it causes for vendors.  Sales can be easier, faster and cheaper when a customer can evaluate your offering online, without any assistance or approval, and make a fast decision based on what they see.  Dev, test and support costs can be much lower when you no longer have to support many different OSs, DBMSs, app servers (and versions of all these, and dozens of combinations of all these).  You can release new capabilities and bug fixes as often as you need without inconveniencing your customers with a huge redeployment.  And on and on and on ...

The SAAS model can be so much better than the traditional model that you actually want to cannibalize your traditional products.  Don't you want to reduce your cost of sales and cost of support? Assuming your software and services are good enough, the lifetime total value of a SAAS customer can be much higher than a traditional model because the cost of sales and support is lower, and the service renewal rate can be very high.  If your customers like you, they will renew service every month or every year and keep on going.

Of course if your software / service is not good, customers will drop you like a hot rock.  That is the core issue of concern, the unspoken fear behind the "cannibalization" question - we understood the old market and did well there. We will have to work hard to respond to the new market and may not win there. Are our products up to the challenge? Can we afford to maintain the traditional product while creating a new one? Can we cross the chasm?  The economics of transition are tough, but if you don't do it, someone else will - another company whose ideas are just as good as yours, doesn't have the benefits of your install base (recurring revenue, customer relationships) but also doesn't have the pains (legacy code, support costs, etc.).

It is up to you to which role your company will play in the SAAS world - the explorer, the cannibal, maybe even the firewood or the giant pot. Eat the other guy before he eats you. "Cannibalization" is only bad if you are the one in the pot.

When a new business model emerges, most people either (a) try to understand it by analogy to the other business model, or (b) freak out and try not to think about the new business model at all. You can see both behaviors in the software industry today. I believe the right approach is (c) understand the new model and how you can best play in it, without getting too hung up on the other model. But that's easier said than done - if you grow up in a certain industry / business model, you eventually become an Old Dog and will find it hard to learn New Tricks.

So maybe one way to crack the ice is with an example from another industry that we can all understand without our software-industry mental filters turned on. One great example is a revolution that happened in the video rental business around 10 years ago.

The established model for doing business worked like this:

• The retailer bought videocassettes from the movie studio for around $70
• The retailer rented tapes for around $3 and kept all the rental revenue
• If a tape didn't rent, that was the retailer's problem (held all the downside risk)
  
• If a tape rented many many times, the movie studio missed out on the upside
  

This created a bad situation for all parties. The retailer would not buy a tape unless they felt certain it could be rented at least 20-30 times and could cover its own cost, so they bought relatively few tapes. If you were the retailer and were sure you could rent 5 tapes but not sure about the 6th, you would only buy 5 (and certainly not 10 or 100). You had potentially many customers who might rent the tape if given the opportunity; but no way to predict that demand. Customers were unhappy because they would go to the video store and not find the movie they wanted (happened to me many times). Studios were unhappy because they could potentially have sold more tapes. Retailers were unhappy because they often aimed too low, missed out on potential rental revenue, and their customers were not happy with them. Overall, a pretty dysfunctional market.

All this changed due to a new idea driven by some visionaries (people and companies) in the video rental industry. A new business model was established that changed all the rules. In the new game:

• The retailer bought videocassettes from the movie studio at cost, around $5
  
• The retailer rented tapes for around $3.50 and split this 50/50 with the studio
• When a tape stopped renting, the retailer would sell it and split that 50/50 as well

This was truly a new business model for the video rental business. Now the retailer could buy a vast number of tapes and not worry too much - if a tape rented only 2-3 times it would cover its cost. As a result, retailers bought MANY more tapes (I remember walking into Blockbuster and seeing hundreds of copies of a new release). As a result, customers were more likely to come to the video store since they expected to find what they wanted. This created a boom for the movie studios and the rental stores, and fueled the growth of the industry - for example, Blockbuster experienced 30% growth in number of retail outlets, and 50% growth in revenue in the 3-year period following this change. Revenue for the movie studios also boomed. Retailers who did not switch to this model lost share in the growing market.

Apply this lesson to software -

1. A new business model might be a great thing for you
2. Trying to get all your money up front may actually hurt you

Asking $70 per tape up front (that's a perpetual license) makes the customer less willing to buy from you, because their expense is certain but their return is uncertain. You will do some business that way, and it feels good to get your money up front, but your customers is not stupid - they will tailor their buying behavior to the risk / reward economics of the situation. If you create a way to share the upside and the downside with them - you will make more money.

SAAS is a great way to share risk and reward with your customers, because you don't ask for money up front, and if you don't perform your customers can move on. This is nearly heresy for an industry that has always focused on large financial commitments up front and ways to "lock in" customers to your solution. But if you want to experience "Blockbuster" results you may have to move to a different risk / reward model. SAAS is a great way to do that.

One aspect of which is, assumptions about the size & shape of your customers compared to your software. In the traditional model, the starting price of your software - the purchase price plus unavoidable related costs like hardware, 3rd party software, human effort to install and configure the system, floor space, rack space, power, sys admin time, backups, disaster recovery, on and on - was usually pretty large. Think of the usual suspects in the ERP, CRM and (dare I say it) enterprise systems management space, and you can find many data points to support this. Analysts will commonly say, the annual cost of owning any enterprise application is between 2X and 4X the initial license price - every year, for as long as you own it.

As a result, the only customers who would consider using that software (the only ones who could form an economic rationale for why the software would provide more gain than pain) were, naturally, big organizations. With a large start up price per customer, and a large annual TCO figure, traditional software only makes economic sense if its costs can be spread over a large number of users, or servers, or whatever the unit of commerce is for that particular kind of software. As a result, almost all of the buyers were big companies.

This led to two unfortunate behaviors:

• Software vendors have come to believe that large enterprises are their natural market, as opposed to just the only market that could afford the current form of their products

• Smaller companies have been underserved, being offered either dopey cut-down versions of the enterprise products (useful features removed) or just special pricing (overlooking the TCO which dominates the cost equation in the long run).
  

This was a lose-lose situation for the software vendors and the potential customers. Very sad because it didn't need to be this way, but perhaps inevitable since the traditional choice of software implementation style made high TCO almost unavoidable.

SAAS changes the equation to a win-win, because the startup cost for a new customer is, zero. This is made possible via a clever but (in my opinion) misuderstood design concept called multi-tenancy. More on why M-T is great, and why it is misunderstood, in Part 2!

Think about other services you use in your daily life. I have a person who mows my yard for me. Guess what he calls his business - it is a Lawn Service, not Lawnmower as a Service (LAAS). I don't care how he cuts my grass, as long as the result is good - he could use a rotary push lawnmower, a gas-powered lawnmower, a pair of scissors. I would object if he used an unsafe or ridiculous tool - no flamethrowers please - but within the bounds of common sense and accepted best practice, the lawn guy can cut my yard any way he sees fit.

This is obvious in the context of real-world services we all know outside the realm of software. Do any of these make sense?

To my ear, these all sound ridiculous, and actually so does the phrase "software as a service". Forget there is software in the mix at all. What service (what result) are you trying to achieve?

Maybe the right way to achieve your result is via services delivered hands-on by human beings. There are problems that are not worth generalizing and capturing as the expensive intellectual property we call "code". In my years in the software industry, I have seen many man-years of effort wasted trying to write code for operations that would have been better done one-off by an expert. This is not "Expert as a Service" (EAAS) but just, consulting.

Or possibly the right way to achieve your result is via services that include some SAAS, some software and hardware that really should be on your premises, and a certain amount of human-powered expert knowledge as well. This is what I call a managed service - people in the loop to manage the things you don't want to so you don't have to.

Although I believe in "software as a service" as a great business model and design model for many kinds of problems, it isn't the only way to solve the broad range of issues our industry is suffering from. One of the tricks I learned a long time ago as a product manager is to listen to the words customers use, but dig beneath the exact words to get at the underlying motivation. I think in many cases when a customer says "I want a SAAS solution", they may actually be saying, "I want a solution that is convenient and cost-effective, that does not give the vendor all the money up-front so they lose their motivation to help me succeed, that does not assign me most of the risk and the vendor most of the reward". Of course in some cases the customer is literally saying, "I want this app in SAAS mode of delivery, stop screwing around" and we need to be prepared to hear that too.

Here's to software as a service and any other model that improves on the "put a CD in a box and mail it to the customer" model. Here's to problem solving with the full range of tools and ideas available to us.