By Michael Woloszynowicz

By Michael Woloszynowicz

Saturday, October 30, 2010

Freemium, Panacea or Parasite?

Based on my observations there is a huge dichotomy in the freemium camp, there are those that swear by it, and those that swear at it. Given this level of divisiveness, how do you decide if it is right for your business? I've been wrestling with this notion for some time now and would like to share some thoughts on the subject.

The evangelists will argue as follows:
  • By taking a freemium approach you can rapidly build up your user base which increases awareness of your product thus further driving growth
  • It allows people to get familiar with your product over a long period of time and as they get more entrenched they will need more features, thus compelling them to upgrade
  • You are less likely to loose customers to other companies as paying users will at worst downgrade to your free service
  • It makes life easy for sales as they have a large pool of users that they can upsell to
Sounds great, but what about the bad:
  • Your free customers may never convert (and most don't)
  • You have to support a large user base that pay's nothing and uses up the bulk of your computing resources
  • Conversion rates can be dismal, less than 1% for many companies and up to 5% if you're lucky, that means your 1% has to pay for the remaining 99%'s usage
  • You have no real gauge as to how valuable your service is since the bulk of your users have not attached a monetary figure to it
  • It can be easier for your customers to switch to another provider since they have no financial buy-in and are not locked into a multi-period contract
  • You could be leaving a ton of money on the table
First off, freemium is not a fad or new concept, it's been around quite some time. In fact it's not even unique to web based businesses. The best example of a freemium model outside of tech is a bookstore, let's pick on Barnes and Noble. Bookstores have a number of users that frequently come into the store, read the magazines, and never purchase anything. Those that wish to read the magazine in the comfort of their home pay the premium price, while those that don't care pay nothing. Nearly a decade ago, in an effort to bring in lots of foot traffic, Barnes and Noble tried to make the experience of reading in-store as pleasant as possible by providing ample comfortable seating. Over the years we've seen the amount of seats diminish to zero and it is clear that the company is trying to make in-store reading as uncomfortable as possible. The reason I bring up this example is because this is where many web companies are going with freemium as well. Many are either eliminating freemium outright (e.g. CrazyEgg) or they are burying their free plans in hard to find places with fewer and fewer features (e.g. 37 signals). That's not to say that freemium is wrong for all companies, there are many that have been extremely successful and happy using it (e.g. Box.net). The important point however is that it's not for everyone and you need to determine which camp your product/business fall into. So let's begin looking at some critical considerations.

One of the most fundamental reasons for considering a freemium model is the presence of network effects. If the value of your product increases exponentially with the number of users (see Metcalfe's law) on the system  then freemium can be a good way to go. Similarly, if your product is a platform supporting a two-sided network, a freemium model can be applied on the subsidy side to generate added revenue through premium features. You can think of freemium on the subsidy side as a sort of bonus as these users typically pay nothing in an effort to grow the size and value of the network. If network effects are not a factor in your product you are almost always better off with a traditional pricing model and free trial. An exception to this is if your product has multiple revenue streams whereby non-paying users can engage in individual transactions. For example having a photo sharing site that also offers a printing and framing service, so although a user may not pay for a premium account, they may place periodic orders for prints. In this case it is clear that having a large user base is also beneficial. In the same vein we can think of multiple revenue streams as non-monetery currencies. Your free users can effectively pay you by submitting data and increasing the value of your offering. For example, free users tagging photos on Flickr creates value for the entire site. Ideally this value can be converted into a monetary figure to see if it justifies the cost of supporting your free users.

If you are considering switching from a paid model to a freemium model it is crucial to ensure that your websites architecture and infrastructure are scalable so they can support the rapid increase in users that will hopefully ensue. If it isn't you risk loosing your paying customers as well as the free ones that come in. You also have to ensure that you have enough cash on hand to support your freemium model. For example, if you're currently breaking even, or are only slightly cash flow positive and you introduce a freemium model, you will immediately go cash flow negative. Freemium not only increases your expenses (if you don't have excess computing capacity or are you're using cloud service such as Amazon S3) but also decreases revenue as certain paid clients may drop to a free plan. If you don't have enough cash on hand to ride out the dip then you may end up in flow-based insolvency. The important thing to note is that it may take a year or more to convert a free user so the cash-flow negative state may persist for some time. The same lessons applies to a new venture, if you are hoping to bootstrap your startup then freemium is not a good starting point. Due to the long conversion times it is also important to assess your retention rates. If users only use your product for a short period of time the it's unlikely to work under a freemium model. Simililary if your users are casual and only use the product on a sporatic basis, they are unlikely to become a paying customer.

Below is a table summarizing the factors that should weigh on your decision:


As a general rule of thumb, unless your product demonstrates a tremendous degree of network effects, you should charge for the product from day one and experiment with different price points, subscription levels, and free trial periods. Diligently track customer conversions with metrics, and monitor these metrics over a longer period of time to gauge the price sensitivity of your users. This will give you a good measure of the value of your product and how attractive your proposition is to your clients. Now that you have some financial information to work with you can use the above criteria to determine if freemium will increase the value of your product, or have long-term strategic effects that justify the added costs. Finally I will re-iterate that the decision to use freemium should be made very carefully as it can have a significant and lasting impact on your business. 

If you liked this post please follow me on Twitter for more

Wednesday, October 27, 2010

The Minimum Viable Product Dissected

One of the fundamental and most misunderstood tenants of the lean startup is the minimum viable product (MVP). The misunderstanding breaks down into three key areas which are described below. I've decided to address these three issues in one consolidated post as I've arrived at the solutions through various sources, including asking the lean startups founder himself, Eric Ries.

The first misunderstanding relates to the MVP concept itself. Many people are unclear how an MVP can be created for their particular product and what components they need to integrate into it. For starters you must look at the MVP not as a product but as an experiment. The prerequisite for any MVP is a hypothesis that you want to test as this drives the scope, nature, and content of the MVP. Once you have decided what it is that you are trying to learn or validate from the MVP you can then begin constructing it. This is generally the stage where confusion sets in. What exactly do you put into the MVP and how far do you go with it? Again the answer is driven by the hypothesis you are testing, but the important thing is to be creative. Many people think that if they have a very complex and extensive product, an MVP that tests whether clients will buy it would still require a great deal of functionality, even if it is the minimum. The fact is that the actual product does not have to exist at all. Don't lock yourself into the mindset that your MVP is some early version of your final product, in fact it is likely that you will throw the MVP away completely. With the expectation that you'll discard most or all of your MVP, you should be motivated to make it as basic as possible, this is where the creativity comes into play. Rather than build a whole system, build a simple landing page with some screenshots from your graphic designer (think of about.me), a value proposition, and a call to action. This call to action can be something as simple as "register for our beta launch". This will give you an idea as to whether anyone is interested in the concept. If this is the route you take then you'd be wise to A/B test the landing page as your value proposition can be strong but your message or delivery may be flawed. If you want to go further you can create a functional html prototype driven by static content that terminates at a form that tells the user "our product is currently in a beta stage, submit your email for an exclusive invitation to preview the system". At this point you may be thinking, "but what if my product is not done for a year, this person will be very disappointed, we've effectively lied to them". The fact is that the number of people that will arrive at your demo site is quite small relative to your target market which is hopefully big, so even if you do disappoint a few people, there will be many others to pursue. Besides, thanks to their interest the product may indeed one day come to fruition, and if it solves an important problem they will return once it’s done. Other options for your MVP include a PowerPoint prototype that you demo in person, a video of screenshots that you've pieced together, etc. Generally the more concrete your MVP the more accurate your feedback will be as you leave less to the imagination, hence you should continually test your MVP at each stage of development. The key takeaway is to be creative and remember that it's OK to fake it and fib a bit.

Similar to the above issue, I've often wrestled with the question of how the user experience (UX) plays into the MVP. How far should we go with UI design and creating an engaging UX in an MVP prototype? The answer here is to once again focus on your specific hypothesis and do as little as is necessary to validate it. Eric Ries has aptly pointed out that at the customer discovery stage you really can't create a good UX since you don't know who the user is yet, you simply don't have enough information. I would therefore lean towards a clean but minimalist UI in your MVP. Once you have moved on to the customer validation stage you can focus on perfecting the UX through more in depth research and tools such as A/B testing.

The final misconception lies with intellectual property. Most people feel that by throwing an MVP on the web they're giving away their idea and others will be tempted to steal it. There are several rebuttals to this claim. First off, Sean Ellis will tell you that most of the people that have the capability to copy your idea are way too busy with their own ideas and tasks to bother with yours. Secondly, Eric Ries will tell you that in practice it is very difficult to copy someone else's idea, go ahead and try it yourself. Finally, David Cancel will tell you that nobody gives a f*** about your stupid idea. The fact is that nobody knows if your idea is any good at this point, not even you do. If you'd heard something about a site called Twitter several years ago would you have been compelled to copy it? You'd have probably though "that will never catch on, who needs that?". The reason you would think that is because you are interpreting the idea from your perspective rather than the vision of the creator. In fact if two people pursued Twitter simultaneously it is unlikely that both would be successful or end up with the same result, as it is the execution of the idea and the pivots throughout its inception that define the outcome. The important thing to keep in mind is that the learning you gain from the MVP far outweighs the chance of someone stealing your idea. After all there is about a 10% chance (and that is generous) that your startup will be successful and less than 1% chance that someone will steal it so you decide which number it's best to dwell over.

If you liked this post please follow me on Twitter for more

Saturday, October 23, 2010

Lean Office Design for the Bootstrapped Startup

So you've launched your self-funded web-based startup and can no longer keep working in your basement. You need to setup a working space for your employees but don't have a fortune to spend on furniting it. I've recently been tasked with this very challenge and here are my suggestions to anyone in this situation.

First off, why is your office setup and design important? Well, for starters if your not a terribly well known startup and can't afford to pay extravagant salaries, attracting talent will be one of your greatest challenges. If a prospective employee comes to your office and see's a poorly furnished space, their immediate instict is that the company is cheap and doesn't care about providing its employees with a comfortable work envorinment. One of your key differetiators from a large corporation can be a fun office that people love to come into every day. Secondly, since you're a startup, it is likely that the people working for you will be there for 10+ hours a day so why not make it enjoyable for them and keep them as productive as possible.Fortunately all this can be accomplished at a fairly reasonable cost.

Let's begin by considering what elements your office space should have. Ideally the office layout will be fairly open to promote collaboration and offer an ample amount of natural light. If light is scarce, invest in some good quality bulbs as this can make a huge difference when one is staring at a computer screen for 12 hours.

Once you've found the ideal space, the challenge of furnishing it begins. As you begin your search for office furniture you begin to realize how expensive it is. Since you are dealing with an open space, the electrical is likely to be provided by a ceiling drop so you'll need a workstation to feed the power into. New workstations range from $1500 to $5000 each, which certainly adds up when you are buying 5-10 workstations. Workstations can be found used for $600-$1200 each, but I don't recommend going this route. The second hand market is littered with mismatched, damaged, and out of style units that are never quite what you want. What I therefore recommend is the route I have chosen, purchase a set of 4 six foot panels and arrange them to form a cross configuration. Then purchase 4 free standing 6 foot x 6 foot wooden L shaped workstations to arrange in each corner of your panel setup. The panels will feed power and you are free to choose whichever desks you want. Surprisingly this turns out to be much cheaper as one panel with power will cost around $350-$450 (depending on height) and a wooden workstation will run about $500-$600 bringing your total cost to around $900 per workstation. You will end up with a modern looking setup at almost the same cost as a used one. As an added bonus, installation of this setup will cost about $100 per workstation vs. $200 if you buy a standard panel based workstation. So besides the obligatory workstations, what else should you get for your space? I suggest the following items:

  • A couch/lounge area, sitting in one place is never fun and if your team works with laptops why not let them move around, plus a lounge area gives them a place to sit, think, and brainstorm
  • An espresso machine, programmers work late into the night, we drink a lot of coffee and really appreciate having an espresso machine in the office, plus it will save time as your staff won't have to drive to a nearby coffee shop
  • Tons of white-boards, this will help people brainstorm and plan, and is an absolute must for any software company
  • Comfortable chairs, although expensive at $400+ it will keep your employees happy and productive
  • Dual monitors (21" or greater), monitors are dirt cheap and will make your employees at least 10% more productive so they will pay for themselves in a week 
  • Make it fun, paint the walls a cool color, hang up some pictures, add your logo as a mural, this makes it an inviting place to come in, plus colors can stimulate different parts of the brain (blue makes you creative, red makes you more careful)
As a budget, if you are setting up your office for 5 programmers I would expect to pay around $15k. It may seem pricey but a nice office can be a strong tool in your arsenal for attracting and retaining top talent.

If you liked this post please follow me on Twitter for more 

Wednesday, October 13, 2010

Porter's Five Forces, Web 2.0 Style

Michael Porter’s Five Forces framework has become the de-facto tool for performing a strategic analysis of your industry or business and determining its overall attractiveness/profitablity. While it is quite universal, its application varies across industries, particularly for a web 2.0 business. This post is intended to be a brief primer on the strategic elements to consider when applying the framework to your web business. It is by no means all-encompassing as many of the standard measures in the framework still apply and are already well documented.

Degree of Rivalry
Typical considerations such as excess capacity and exit barriers can be ignored here so the only factors are: is the market you are service particularly popular and thus saturated with lots of new entrants, how similar are the offerings provided by them, and how similar are their strategies or target markets (e.g. demographics, location, etc.).

Supplier Power:
Since web businesses revolve around virtual goods or services there are very few suppliers involved, thus this category can largely be ignored. The one exception is if your application built on top of a 3rd party API (e.g. Twitter or Facebook). If this is the case then you have to be aware of the possibility of your supplier forward integrating and encroaching on your business. We’ve seen this happen a great deal recently as Twitter has increasingly added functionality that is currently being provided by third-parties (e.g. link shortening, photos, and multiple clients).

Barriers to Entry:
This is the largest category to consider as there are numerous elements in the Web 2.0 space that increase barriers to entry. Simultaneously there are also many elements that disguise themselves as barriers, but aren’t necessarily so.

What applies?
First off, does your application require a great deal of time and money to build? More specifically, does the minimum viable product version of your offering require a significant amount effort to bring to life, and there are no pieces of your product that can be separated from the whole. If this is the case then any competitor would have to dedicate a great deal of time to match your offering, at which point you are moving forward with more functionality, improving your offering, and learning about your market, thus creating a moderate barrier. A similar but more powerful barrier is created if your product’s technology is difficult to replicate (e.g. Google’s search), or if you hold patents on key aspects of your offering. Marketing can also become a barrier if your brand name becomes synonymous with a particular market or product type as you become the logical choice to service a clients needs. This is especially true in enterprise software for as the old adage goes, “nobody ever got fired for buying IBM”. If your product exhibits a high degree of network effects (e.g. Twitter, Facebook), namely the value of your offering grows exponentially with the number of people using it, and you have established a sizable user base, a very strong barrier is formed. While we have shown that building on top of an API can be risky, being the API provider has a number of advantages. If a number of businesses are based on your platform then value of your product is greater than its core functionality, the third parties are indirectly promoting your product on an ongoing basis, and user switching costs increase drastically.
Some other quick barriers include:

  • Does the value of your product rely on a great deal of data that is acquired over time?
  • Is the market that you serve highly complex and takes years to learn? If so then you have a path dependency (learning curve advantages) that others would have to go through in order to match your value proposition.
  • Owning a key distribution channel (e.g. being first in the organic search results of Google)

What doesn't apply?
Infrastructure requirements are no longer a barrier to entry as cloud computing services have removed high upfront costs and turned them into variable costs that grow along with the success of ones product. Similarly, economies of scale don't apply as cloud services scale and reduce automatically. While it’s natural to think that having a highly viral application is a barrier as it helps you to grow your user base rapidly, it also helps your competitors grow. Therefore it is more logical to think of a lack of virility as a barrier as organic user base growth is much more difficult and time consuming to replicate. Another non-existent factor is high marketing/promotion costs. Although it may have cost you a great deal of money to market your business, social networking tools have drastically driven down the cost of promotion, allowing bootstrapped startups to effectively advertise their product with only a bit of patience and creativity. This may not apply as much to enterprise software providers as social media is not as powerful in this space and sales are made largely on reputation. In case you haven't noticed, enterprise software is a difficult market to service.

Buyer Power
Buyer power boils down to more traditional considerations such as the number of substitutes available for your product. When thinking about this it is important to look beyond your way of servicing the user’s need, perhaps there is a completely different way of achieving the end goal. In addition, if you are in the enterprise SaaS space, is your business reliant on a small number of large buyers? The final consideration is the switching cost of your product. These don’t necessarily need to be monetary costs, as the “cost” of switching can be a loss in value such as having less people to connect to, no longer being part of a community, or having to build up your persona or position/ranking from scratch. For example, if a user has a high standing on a site like stack overflow, he/she is unlikely to switch to a new and similar offering as they would have to build this position up from scratch.

Threat of Substitutes
Largely similar to what has already been discussed, threat of substitutes effectively reduces to the switching cost of your product as well as it’s price and performance compared to other offerings that solve the same problem.

Keep in mind that most Web 2.0 firms do not benefit from any of the above barriers but still remain successful so don’t fret if you score poorly on these measures. A low barrier to entry is simply a characteristic of the web based software industry, as open source tools and cheap outsourcing allow nearly any entrepreneur with a great idea to bring it into fruition. Your main defense against competition is to stay ahead, create a product that your clients love to use, and most importantly know your customer and market. Always bear in mind that when performing the above analysis, you have to correctly identify the market your are analyzing. Look at all firms that help service a customers “job” in any way possible, not just your way of doing so.

If you liked this post please follow me on Twitter for more

Tuesday, October 12, 2010

The Technical Co-Founder Search

I was recently asked for advice on finding technical co-founders for a startup, so I thought it best to share my response with a wider audience. 

In order to keep this post generic I'll break it up into two stages as the advice varies depending on where you are in your startup.

The Early Stage
By early stage I am referring to the point in your business before product-market-fit is established and a verified business model has been put in place. To be more specific, you have tested one or more minimum viable products, found the element of your product that people love, and chosen a route to focus on. If your startup is not yet at this point then I suggest reading an excellent article by Charlie O'Donnell, Need a Technical Co-Founder which will help you determine if it is the right hire in the first place. Charlie suggests looking for a Product Manager at this stage, and it is in fact a suggestion that I touch on later in this article. I would add to Charlie's post that this is a good time to learn some basic programming. Get your feet wet with HTML, CSS and back-end languages like PHP, C#, or Java. The goal is not to become a professional programmer, but to understand the lingo and technical discussions your staff is having. Many non-technical co-founders can often help with early prototypes that have limited functionality and are intended to test the hypotheses you have about your product. Also, at this point try to find a part time technical advisor to give you some early guidance. This is much easier to do than finding a full time partner as many currently employed business leads will make time to help out other startups. 
The Growth/Expansion Stage
So you have a product that people like, that addresses a market need in a market that is big enough to support your business, great! By this point you have either built or are in the process of building your working prototype, likely through a combination of in-house and outsourced resources and are beginning to think of the next stages of your business. You realize that your lack of technical abilities can only take you so far, and so the search for a technical co-founder begins.
So who are you looking for? As with programmers, a technical co-founders match is based 50% on personality and 50% on technical abilities. This presents a challenge as personality is not always easy to gauge in the short term (more on this later). From a technical standpoint your interview should focus on problem solving skills, breadth of knowledge, and vision. Is the person feature focused or more appropriately, product focused. Do they view the application as a product or an ecosystem from which new products, ideas, and sources of value can grow? Have they worked on routine engineering projects or have they encountered a good breadth of technical challenges? Your focus should not be their knowledge of your specific programming language (although this helps), but rather their ability to architect a well-functioning, highly usable, and robust application. Are they familiar with the latest computing trends and technologies such as web services (e.g. SOAP vs. RESTful), distributed computing (e.g. SOA), API's, and open source technologies (you want to keep your costs down)? Ask them to take you through their application design process, what considerations do they make? Do they use the out-dated waterfall model or the better suited agile methodologies? See if they think of the application as it is today, or as it could be in the future. Ask about their failures and struggles, how they recovered, and what they have learned. Do they give their team as much credit for their success as they do themselves? As far as eduction goes, a Bachelors degree in Computer Science/Engineering is a must and an MBA is often a benefit as they should be able to weigh technical vs. business trade-offs. 
As mentioned above, product managers make good candidates, provided they started their career as programmers. If you choose to hire a PM in your early stage, search for one with a CS background and groom them with this future position in mind. Project Managers with a CS background are also wise choices as they've been faced with many of your challenges on a smaller scale. On the personality side, look for passion and excitement about the work they've done and an interest in your product. Look for someone that is personable and inspiring but also strong willed.
So where do you start looking? As anyone will tell you, a great starting point is tapping your network and asking around for people they would recommend. If this fails it's wise to begin attending technical conferences such as the Web 2.0 Expo which offers an attendee directory. Scan the directory for people with the right skills (as described above) and engage a meeting with them. Scour LinkedIn and become an active member of technical groups, see who stands out. Another option is to contact job banks at MBA programs as a good proportion of MBA students have technical backgrounds. Find candidates that have several years of technical experience and if they fit a good chunk of the above criteria, give them a chance. Once a suitable candidate is found, bring them on as a lead developer or project manager with the prospect of becoming the companies CTO. Offer them restricted stock units with a vesting period of at least one year. An interview will never reveal if the two of you work well together so test things out before you issue equity. Your candidate should be willing to take a lower salary in exchange for ownership in the company, thus proving their dedication and belief in your product, and their drive to make it grow. 
The search for a technical co-founder is much like the search for talented programmers and designers, only more difficult. There is no magic formula and it requires you to keep your eyes and ears open, interact with lots of people, be patient, and never settle. As your product develops and your user base grows, your sales pitch to talented individuals will becomes more enticing so don't get frustrated if you can't immediately find someone. Remember that anyone you hire will heavily influence the direction of your business so ensure that they are good fit for your company and that you share similar values. 

If you liked this post please follow me on Twitter for more 

Monday, October 11, 2010

The Great Outsourcing Scare

I recently read an article on Yahoo once again claiming that software development jobs will decline in developed countries as outsourcing will become ever more prevalent. 

I for one wholeheartedly disagree with this claim for a number of reasons. First off, we've been down this road before and when I went to take my Computer Science degree more than eight years ago, the .com bust and ensuing rise of India and China as outsourcing destination led many to question me for pursuing such a degree. So where are we now? Well outsourcing certainly has changed the way many industries work, particularly IT and manufacturing, but there is also no shortage of jobs for talented developers. As the person responsible for hiring programmers at my company, I have discovered that the pool of developers with creativity and problem solving skills is getting smaller and smaller. This trend is not only evident in the top percentile of the talent pool, even mediocre developers are hard to find. The same sentiment also extends to tech hubs such as Silicon Valley as talks with several startup companies reveals a similar struggle. So why could this be you ask? First of all, students are being fed the same nonsense that I listened to prior to undertaking my degree and are pursuing different fields. This is evident in the steady decline of CS graduates over the last few years. Next, the recession has brought with it a demand to lower costs and automate processes, something that computers are great at. Similarly the rise of new hardware devices such as tablets and smartphones have created new markets for software, so while some development is being outsourced, the increase in demand has more than offset the negative effects of offshoring. 

One of the central arguments of Yahoo's post was that software development is a simple candidate for outsourcing as it is virtual in nature. By the same argument we could extend the outsourcing scare to professions such as architecture, engineering, business analysis, financial analysis, and so forth. With the North American manufacturing sector essentially dead, and the predicted outsourcing of all "virtual" industries soon to follow, can it be posited that the job market in developed nations is effectively doomed? By their argument there will only be work for doctors, top-level managers, and low level customer service representatives. Wouldn’t this also increase wages in developing nations thus reducing the benefits of offshoring? In fact this trend is already occurring as outsourcing firms are already looking towards new destinations in Eastern Europe to remain competitive. Sure the above example is a bit exaggerated but the fact is that you cannot outsource everything. As we all know, the true success of any company stems from its employees. Outsourcing has an important place in a company’s toolbox but it is not a panacea for IT companies. For a company that strives to be the best, there is no substitute to a locally present, dedicated, and energetic team.
I write this post in the hope that students will continue to view computer science as a lucrative and growing field rather than one destined for decline. 

If you liked this post please follow me on Twitter for more