Chaotic Flow

Streamlined angles on turbulent technologies

I couldn’t let the much anticipated announcement last week of the final draft of the new Affero GPL go by without making at least a few comments. First of all, I love open source. But, I am unimpressed by the new GNU Afferro GPL for three simple reasons.
1) It is a weak effort
2) It is already obsolete
3) It is irrelevant
While there are 10 pages dealing with everything from patents to liability, largely legacy from the original GPL that was designed for installed, perpetual license software, there is fully ONE SENTENCE that attempts to deal with what is clearly the biggest disruption to the software industry since the personal computer.

“Notwithstanding any other provision of this License, if you modify the Program, your modified version must prominently offer all users interacting with it remotely through a computer network (if your version supports such interaction) an opportunity to receive the Corresponding Source of your version by providing access to the Corresponding Source from a network server at no charge, through some standard or customary means of facilitating copying of software.”

This is a band-aid and indicates a lack of interest in exploring the issue. The elephant in the room is that real value is shifting from the software itself to the fully integrated service offering, and the two are intimately intertwined. The delivery, scalability, integration, configuration, usability and administration of the software have as much influence over the Web user experience as the source code itself. Put another way, whose interests are the FSF really representing, developers or consumers?

While I think it is a truly beautiful philosophy that software should be “free to run…to study…to improve…to redistribute,” it is in reality free because its fundamental economic value is dropping to zero! If you buy into the idea that software-as-a-service is built around the delivery of a commodity offering, then you understand that real economic value and competitive advantage lie outside the functional capabilities implied by the source code, i.e., proprietary assets or IP in service delivery, brand loyalty, or switching costs. Any competitor worth its salt can duplicate another’s functionality in a matter of months, with or without the source code, but can that competitor deliver it to the user.

Alternatively, Web-based SaaS applications bring capabilities to consumers and SMB’s that they would not be able to afford or are perhaps even impossible under an installed, license distribution model. Different proponents of open source (myself being one) each have their own philosophy and motivations, but one underlying theme is making good software available to everyone. While much of the controversy surrounding this new license has been about stopping evil Web companies from taking advantage of open source, I will submit that it just don’t matter. Free software continues to make strong inroads into the installed, enterprise software space (and I will arbitrarily include Web servers in this category), but it is in danger of losing its philosophical underpinnings on the increasingly syndicated, service-driven Web.

Many B2B SaaS businesses will succeed or fail based on the volume, cost and quality of inbound leads. The reason for this is straightforward: subscription price. If you are like most SaaS vendors, you are probably offering an annual subscription price that is 5%-25% of the equivalent upfront perpetual license of a traditional enterprise software vendor. There is a lot of analysis by VC firms etc. regarding what the real benchmarks for these numbers should be, but I’ll suggest we just go with the theory that you have to generate leads for A LOT LESS. If your subscription rate is $5000/year, then you need to acquire customers for anywhere from $2,500 to $7,500 total absorbed sales and marketing cost (the lower value if you don’t have a lot of cash, the higher if you can absorb a three year payback timeframe).

Therefore, you can’t afford expensive lead generation efforts in the $1,000/opportunity range, or direct selling costs in the $10,000/deal range, which are the bread and butter of traditional enterprise software, i.e., outbound telemarketing, print advertising, direct mail, trade shows, and heavy field marketing backed up by field sales reps making $200K/yr on $2M/yr in quota. While you may be able to apply some of these methods selectively for concentrated, high-need, high-value market segments, most SaaS startups will find them unprofitable. You must strive to develop marketing campaigns that bring customers to you and a sales model that can close a deal for around $1000. What are you to do?

You are a Web-based business—act like one! Your first plan of attack should be Website search engine optimization (SEO) and search engine marketing (SEM). Your Web site should be built for SEO from the ground up on the day you launch, not optimized (or sub-optimized) later on. Your competency in online marketing and public relations will be as critical to your success as your engineering effort. You should be leveraging every possible low cost, online vehicle including search optimization, search advertising, directory listings, affiliate programs, email, online articles, newsletters, online interviews, Webinars, blogging, social networking, banner advertising, podcasting, flash tours, free mash-up widgets, viral customer referral programs, free trial, and online purchase before you even consider direct mail or an expensive trade show. Why? Because you don’t just need leads, you need the Glengarry leads. Internet-laggards will be too hard and too expensive for you to close.

After this, your second order of business should be offline PR—because it is free. Opt for evangelism over outbound calling, speaking engagements over booth space, and online case studies articles, and interviews over advertising. Many startups fall into the trap of over-analyzing their highly-fragmented, SMB target markets and coming to the conclusion that a heavy field marketing effort is required to get in front of these hard-to-reach customers. This is not the case. The right answer is that your most important prospect qualification criteria is not industry sector or application need, it is comfort with the Internet. If your target industry sector is food service distribution, then your target prospect is food service distributors that are online, with broadband, and Web-savvy. You cannot afford to communicate and educate with late-stage adopters who are not 100% online today. These customers will come along as they educate themselves. If you visit the Web sites of successful SaaS companies like www.salesforce.com and www.taleo.com, you’ll see ten times as many online events as offline ones, and the offline events lean heavily toward road shows, local seminars, and user groups, i.e., prospects that came there just to see them! If you find that you are spending too much on offline marketing to generate leads or you aren’t getting enough leads, the problem is not the efficacy of your offline marketing program; it is your qualification criteria and your allocation of marketing expenses.

There are many roadblocks that can get in the way of getting leads in the door. But, before we consider these, ask yourself this: “Is my market really big enough?” You would be surprised at the number of vendors jumping on the SaaS bandwagon without a real handle on the size of the market they are really addressing, or the impact market size will have on their ability to deliver their service profitably.

Given that most of the Fortune 1000 IT departments are still trying to swallow the alphabet soup of enterprise software (ERP, CRM, BI, RDBMS, WMS, RFID, ASP, .NET, J2EE, etc.) they have purchased over the last fifteen years, you are unlikely to make much headway with this group as early adopters. For this reason, most B2B SaaS vendors have been honing in on the small to medium-sized business (SMB) market and small, independent functional groups within larger companies. This is the tried and true model of the companies that have been successful securing customers, such as salesforce.com and NetSuite. However, many SaaS applications are not as generic as CRM or ERP, and even these applications can have dramatically different requirements across SMB segments.

Most SMB’s do not consider themselves “SMB’s,” but instead identify closely with a specific, and generally highly fragmented industry sector such as retail fashion, toys, food distribution, hospitality, moving and storage, plumbing, HVAC, instruments manufacturing, and software startups to name just a few. Therefore, careful market segmentation analysis is often required to identify those customers with sufficient business need-ROI and business process-product function alignment with your SaaS offering to drive a purchase at a low enough acquisition cost that is profitable for your business model. In addition, all these “sweet-spot” segments must add up to a big enough mass market for you to achieve the inherent cost advantage of SaaS. It is your cost advantage that ultimately will motivate your prospects to switch over to your SaaS from their current way of doing things, be it manual processes, simple desktop applications, or a traditional enterprise application.

The overriding problem with most B2B SaaS vendors today is their inability to reach profitability. Customer acquisition costs in particular often run 50%+ of revenue, and while logic says that one day subscription revenue will overtake this as the market matures (i.e., all prospects are now customers, so they just pay maintenance and we collect the cash), this plan is not playing out in practice. Welcome to the Silicon Valley world of VC-backed software where growth is EVERYTHING. You will never reach a time where you have all the customers. Even if you do, no one will recognize it and there will be unrelenting pressure to spend MORE on acquisition and new product development to gather up that one last customer to increase top line revenue. You can do all the break-even and NPV analysis you like, the psychology of the industry will not let you follow through on this strategy.

A paradigm shift is required in the minds of both SaaS executives and VCs if on demand enterprise software is to reach the level of success of traditional licensed software: growth follows efficiency, not the other way around. Everyone gets excited about a stable subscription revenue stream, but they forget that the requirement of profitability implies a stable cost structure that is below that revenue line. If you’re annual subscription fee is $1000, then your average annual cost of sales, marketing, support, development, operations and administration better be under $900/customer. Too many companies and investors are trusting too much in the inherent technological economies of scale of SaaS and are waiting to see how the rest of the SaaS cost structure shakes out. This is foolish, because the subscription fee that your target customer (usually SMBs) is willing to pay for your service is completely unrelated to your technology or your cost structure. It is your job to make them meet at a profitable price. I believe SaaS vendors should be targeting and driving toward a stable, lower cost structure throughout their operations—from day one, or if not day two.

If you annual customer acquisition cost exceeds 50% of sales, you should focus on increasing the efficiency of your sales and marketing processes to reduce cost/lead (or increase leads/cost!), shorten sales cycles, and reduce or eliminate labor costs. The days of the high-flying, all powerful enterprise sales star are over. SaaS is a cost-reduction play and your organization must strive to be a customer creation and product delivery machine. Drive leads through low cost Web 2.0 PR and viral marketing techniques, develop an effective, low touch sales process with extensive self-service, and most importantly avoid prospects that are too expensive to acquire—they may look like your target market, but they are not, because you cannot sell to them profitably. Chasing risk-averse buyers in an early adopter market, over-marketing to prospects that are too difficult to reach, and wasting precious selling time on prospects that have requirements beyond your current product offering are all scenarios that must be routinely qualified out of your sales process.

One example I like to share with SaaS executives who believe this is a pipe dream is Atlassian. Demonstrating strong year-over-year growth, this company is selling what might be thought of as traditional license enterprise software under a shrink-wrap model at SaaS prices! How do they do it? By applying Web-centric, low-cost, prospect-driven, grass-roots sales and marketing techniques that creates a strong inbound lead flow and keep out-of-pocket marketing and sales labor costs to a minimum.

There has been a fair amount of pondering over how Chris Anderson’s thesis on the Web’s transformation of the Long Tail in consumer markets might be applied to their business-to-business counterparts, and to B2B on demand software in particular. I’d like to take a slightly different approach and try to clarify where it does NOT apply, in order to better highlight its potential impact in areas that it does.

In my first post to this Blog, I stated that SaaS was fundamentally a mass market, commodity business. On the surface, this statement seems to run completely contrary to the Long Tail thesis. How can I possibly cite Google, Amazon, eBay and MySpace, the most notable enablers of the Long Tail phenomenon, as commodity SaaS providers. The answer lies in which products/services we are talking about.

The table below illustrates the potential source of confusion:

Put a different way, while every Web page, song, and friend may be unique and appeal to a specific group of individuals (niche market), HTML code, MP3 files, and XML profiles are essentially the same and shuffling them around cyberspace is a mass market, commodity business. In fact, this inherent sameness is a critical element of the interoperability that has driven down the cost of information distribution and processing to nearly zero and thereby enabled the Long Tail phenomenon.So, how does the Long Tail fit with SaaS? In his thought-provoking book, Anderson identifies three primary forces driving the Long Tail: democratization of production, democratization of distribution, and connecting supply and demand. Or, in terms of on demand software: tools ala TypePad, aggregators ala Amazon, and filters ala Google respectively. All you have to do is ask yourself one question: “Am I a tool, an aggregator, or a filter?” For the majority of B2B SaaS application providers, such as Salesforce.com or WebEx, the answer is obvious: “I am a tool.” (Sorry, too easy ;)) In fact, under this framework, it becomes clear that B2B SaaS mashup and mult-tenant platform plays, such as Salesforce’s force.com, are not only tools, but they are tools designed specifically to go after the Long Tail opportunity by democratizing production through self-service mass customization.

The core value proposition of the vast majority of enterprise SaaS application vendors, who are uniformly targeting the SMB market, lies in their ability to deliver basic (read commodity), enterprise-class capabilities ala to Siebel, Oracle, and SAP at a fraction of the cost. Attempting to address the Long Tail too early in their evolution could easily undermine this cost advantage and drag them into the “customization hell” that plagues traditional enterprise software vendors. That said, SaaS vendors are by far the best positioned to go after the B2B Long Tail as the sector evolves, because their base offering is a commodity. Into this plain vanilla code base, they can easily inject any number of unique flavors through mass customization techniques, such as extensive self-service configuration and standards-based interoperability.

There is a corollary to this proposition. Tools by their nature are flexible. They can be used to build a lot of different things, even other tools. If you are providing a B2B SaaS application that can easily be duplicated by more general, democratic tools, such as mashups, force.com apps, etc. Get out! You are about to get wacked by the Long Tail, because you will soon be competing with a multitude of do-it-yourself SaaS applications, widgets, and mashups. Better yet, take your thinking to the next level and become an aggregator or a filter. If the Long Tail phenomenon has any legs at all in B2B SaaS, one day there will be more on demand micro-apps than we know what to do with. (BTW, if you think this sounds crazy, get your creative juices flowing by paying a visit to sourceforge.net to see what a real, Long Tail software aggregator looks like in the open source world.)