New Model Software Startups: Two Stage Ventures

How will entrepreneurs wanting to build value in software startups, and venture funders requiring both defensibility and eventual liquidity, respond to the rise of open standards, open source, and offshoring? As I've written before, these are all faces of the commoditization of information technology.

I think I see the shape of one response emerging from the fog. This is partially based on companies which have pitched to our team - and will remain nameless - and partially on a rather inchoate impression of chatter on blogs, tech sites, and open source sites. I'm not sure this is particularly unique or profound, but perhaps it's an early indicator of the shape of symbiosis between community and commercial processes.

Let's call this pattern the two-stage software startup. Unlike a two-staged rocket, the first stage is light and runs on little fuel, it's the second stage that has the big burn, if it ignites.

The first stage of the new model software venture builds a useful product as cheaply as possible. Actual engineering is focused ruthlessly on the unique value and differentiating features. In most cases, open standards are exploited to address as large a market as possible using off-the-shelf APIs and libraries. In many cases, the software is written on top of open source platforms, such as LAMP, to keep down development and initial customer costs. Code is usually written to published interfaces, rather than integrated into the open source itself, to avoid 'contamination' by GPL and other OS licenses. Often, a portion of the development will be sent offshore, particularly if the founders have prior experience or cultural connections with a reliable venue.

Build as little as possible, as fast and cheaply as possible, while demonstrating some unique value. From the classic VC view, many of these efforts will result in a product, or even a feature, rather than a sustainable company. But that may be OK for the first stage, because the development time and expense are small enough to be funded by the founders, friends and family, or a few angels. The go-to-market is similarly light. Rather than a sales channel, the venture will buy ad words on Google, promote itself via word of mouth on blogs and via user communities, and penetrate enterprises by pricing low enough to fall within the purchasing power of a department, or even an individual. Being in early and continuous touch with its market, the venture can course correct early and often.

Assuming this promotion strategy reaches interested buyers, the venture can go cash flow positive fairly quickly. This doesn't mean that it's a sustainable business: Barring some strong intellectual property defenses, the costs to an imitator may be equally low, and the venture's reward for proving a value proposition may be to draw competitors with greater sales and engineering resources. However, the time value of having a functioning product with newly proven value may be sufficient for a quick sale to a larger company which has sales channel synergy, or products in a touching function which can quickly integrate the new functionality. While the sale may result in only a few million dollars, that outcome may be quite profitable to the founders and the individual backers. This may even be true on a risk adjusted basis, and that may be a new thing.

The second choice after a successful first stage launch is to light the second stage, which will require venture funding. Second stage activities will consume cash in advance of the sales to fund them, as they must occur before imitators arrive. They may include adding functionality to meet customer requests, rebuilding parts of the product for greater efficiency and defensibility, adding the necessary sales force, scalability, and system integration to be able to sell to a higher end market, such as the CXO enterprise level, or carriers. At the point of making the second stage decision, the technology risks have been greatly reduced, and a portion of market risk eliminated. The company has already been learning from the market, though it will undoubtedly need to relearn some things as it shifts focus. Entrepreneurs who choose to enter this stage will receive valuations well above what they would have commanded before achieving a first stage takeoff, though perhaps not as much as they might hope: The VC will understand the business still requires significant investment to become sufficiently defensible for an eventual exit, and that the management team which succeeded in the first stage may need to be supplemented for the second phase.

Have we seen this pattern before? In a way, yes. Those of dating back to the beginning of the PC movement remember that a few coders, in less than a year, could knock out an acceptable product. The machines of the time had limited capabilities, customer expectations were modest, networking and enterprise integration were years away, and the early PCs provided embedded code for the machine-specific and performance critical input and output functions. Many of the early companies went the whole way without venture funding, and others bootstrapped their proof of concept stages. These options slowly eroded as the power of PCs grew, customer expectations for functionality and integration of software expanded, and the manpower and capital required to achieve initial release increased. Now we've taken a step back closer to those early days, as the combinations of open source and standards, and commoditization of hardware and software, once again make a low rent market entry feasible.

Comments

[Jeff Clavier]

Couple of observations:
- I agree that this model is increasingly used by software startups, excerpt for the ones where the management team has a track record giving them support from VCs very early on (e.g. SocialText vs JotSpot ?). However, it would be fair to say that this fits more easily a consumer play than an enterprise one, where the sustainability test has to be passed (I am not saying that this is impossible, it is just more difficult).
- In early stage plays, I would submit that LAMP and open source infrastructure, standards and the ever dropping cost of computing and bandwidth has allowed startups to get off the ground for a few tens to a few hundred of thousands of dollars. I see offshoring coming later in the process, once the development is ramping up, but at the very beginning (less than 10 people) it introduces an extra dimension of stress to the venture - in most cases.
- Isn't this model essentially linked to the "return of the angels", who will fund and support these companies to reach the first stage milestone ? And in some cases, angels will imply/involve small (seed) funds ?

[Charlie Kemper]

Great post! But, I would argue that the development of this new model is more likely in response to the Internet - easier & less costly distribution of product. Whereas selling software ten years ago required physical store distribution and the physical publishing of diskettes, now functional software can be sold online and marketed through cheap ads.

[Tim Oren]

Jeff - some of your observations about penetrating the enterprise market sound reasonable on their face, but I have direct evidence to the contrary, in the form of companies fitting the two stage model that have accomplished that entry, and have pitched to us. I'm not going to use their names, since that would violate my blogging rule of confidentially for those who pitch, but I assure you they exist. The keys seems to be pricing to the individual or at most departmental budget, and being a clean add-on to existing products and processes, at least at the onset.

Likewise, we are regularly seeing seed or A round companies that already have an offshore presence. It's usually facilitated by a cultural tie (e.g., outsource to your old college classmate from IIT), or a team member with prior experience in offshoring (e.g., outsource to the Russians who did a good job for you at Sun, or wherever). This is common enough now that it's not remarkable, instead we collect the most exotic offshore locations: Armenia and Bulgaria, for instance.

The advent of the net is definitely part of this; open source itself would likely not exist without the Internet, since it it radically reduces the coordination costs and time for a distributed team. I think you could in fact argue that my entire observation is simply a time-lagged economic adaption to the Internet.

[Brian Flanagan]

Tim, If you have the time, would you like to give us a 200-250 word summary of your 2-stage model which might be less software-specific for inclusion as a weekly business planning tip at "www.planware.org/businessplantips.htm" ?? See any "Suggest a Link" there for more info. Your two stage approach has been greatly facilited by the 'Net which allows many businesses (not just in IT) to dip a toe and pilot a concept with minimal funding.
I am associated with a program here in Ireland which works with high-potential startup mainly in IT and have asked all the participants to consider relevance of your model.

[Elisha Klein]

Tim,
Can explain to me what this sentance means:
"Code is usually written to published interfaces, rather than integrated into the open source itself, to avoid 'contamination' by GPL and other OS licenses." ?

I understand what GPL is I just don't understand this tactic.

Thanks a lot,
Elisha.

[Duncan Krebs]

I can't wait to see how things in the software industry are going to shift moving forward. I'm getting closer to getting a startup to market which ships with a set of Java SWT based tools deployable over the web that enables users and managers inside an organization to quickly deploy structured data and directories to the web and intranet.

This blog kind of hit home in the sense that I have not doubt I can make it to market and probably make some nice cash but being able to sustain myself in light of the changes that are happening is what I question. Sooner or later, someone will be able to go to apache.org and download for free what I’ve spent the last two years slaving over.

But I have to think there are some things in the software industry that remain constant and because they are based on capitalism I don’t see them changing anytime soon. First, companies need software and second they need someone to implement or develop that software. Even if all the software in the world is free there is still a long term business model in packaging software as a service. In fact, look at this site that I'm writing this on right now. Blogging software is out there for free but typepad.com bundles it as a serivce and it looks like they are doing alright.

Its also funny to see IT managers in middle sized companies fumble trying to go from IIS to Apache Tomcat. They just don’t get it and probably don’t want to.

All of this leads me to believe there is still economic value and potential in the industry and its interesting to see how big blue and others have made an economic model out of open source software. I thank them for eclipse.org in which I’ve built my software foundation on.

My burning question is however, how long will Microsoft Windows remain the dominant client computing platform? I give Bill Gates about 20 more years.