451 Research: Zombie accounts and 'freemium' business models

InsightaaS perspective: 451 Research is one of the world’s leading sources of insight into cutting edge technologies — including areas that are important to InsightaaS and our principals, such as cloud, collaboration, analytics, and sustainable IT. As regular visitors to this site are aware, InsightaaS.com works with 451 Research to bring occasional thought leadership pieces to our readers. In fact, we featured a post on M&A involving ad attribution startups two weeks ago, and interest was so high that we thought we’d run this piece as well.

One of the appealing parts of the ad attribution report — and of 451’s research generally — is its ability to connect technology and business perspectives, and this piece provides an even clearer example of 451’s leadership in this area. In it, authors Alan Pelz-Sharpe, Owen Rogers and Matt Mullen take aim at the ‘freemium’ business model frequently used to launch products (and even entire companies) in the cloud. The authors examine how freemium has worked in a specific product segment — social business applications — and though they are quick to note that what is true here “may not be true for other technology markets,” InsightaaS agrees with their belief that there are “lessons, or at least observations…for others to consider.” For example, the five forces that they identify in the body of the text — cost per acquired user, the volume of users acquired, the time required to convert a freemium user to a paid subscription, the value of that subscription, and volatility within the freemium base — apply broadly to many SaaS business models. The Dropbox-based example furnished in the text is similarly valuable, demonstrating the power of a perspective that bridges technology and business issues.

The concluding sections of this report — “advice to vendors and investors” and “the 451 take” — crystallize the value in the content. “The freemium model,” the authors concluded, “can work…[but] in reality, it does not work for many vendors.” The report urges vendors and investors to ensure that they are “in control of the throttle and brakes,” noting that while it is possible to build business through a freemium approach, “overly optimistic conversion estimates and an inability to jettison poor-quality accounts” can make freemium a perilous proposition.

Report by Alan Pez-Sharpe, Owen Rogers and Matt Mullen of 451 Research

Over the past five years, offering free versions of enterprise applications has become commonplace. But as the current startups prepare for potential exit points, the value of free versions is starting to come under renewed scrutiny. In this report, we are focused solely on the relative values of the freemium models for social business applications. It’s important to state up front that what is true for social business applications may not be true for other technology markets. Even so, we believe that there are lessons, or at least observations, here for others to consider.

What are ‘freemium’ services?
There are various definitions for freemium, but they all essentially boil down to offering (for an unlimited time period) a free version of software or a service, with a more feature-rich premium version available at a cost. There are numerous examples of this being used in the commercial world — for example, Dropbox has a free basic service that offers 2GB of storage and the ability to share and sync files. If you sign up and use the service, and in time decide you need more than 2GB of storage, you can either successfully refer others to the service and get a credit, or sign up for a premium version of the service for a monthly fee.

Why go freemium?
There are three main reasons why vendors offer freemium versions of their products:

1: The hope of attracting many customers quickly and at low cost — customers who will then be impressed by the service and convert to paying customers and a premium version.

2: The hope of attracting many customers quickly and at low cost — customers that can then be bundled and sold on to paying parties as targetable market segmentations (both business and consumer in nature). This is popular with social media platforms.

3: By offering a freemium version, a startup can gain some early market traction. This helps the firm’s marketing, sales, and ability (in many cases) to raise money and visibility. For example, Box can honestly claim that it has 15 million users of its service; similarly, Twitter can claim more than 645 million users.

There are different ways to monetize these nonpaying customers. In the case of Box, it is to move them to premium models and convert related groups of users into enterprise accounts. In the case of Twitter, it sells the ability to promote and advertise to its more than 645 million users. Ultimately, though, the goal is the same: to monetize nonpaying customers. The big difference between the likes of Twitter and Box is that the former is a consumer-oriented service, the latter an enterprise-oriented firm. Consumer-oriented firms, by default, have the potential to load free services with adverts to either cover or contribute to the costs of running the service. If they get a sufficiently large or high-quality free user base, they may also be able to monetize the data that this group generates. Such routes to profitably are far more difficult for enterprise buyers that are not keen on either advertisements or having their data exploited.

But what about Google?
Providing a freemium service can be a high-risk strategy for any vendor, no matter how large or small. A large and established vendor like Google, Amazon or Microsoft has the resources to fund freemium products, whereas startups often find the cost of such services cripplingly expensive. Even for large vendors, there can be costs. Google, for example, has angered many by closing freemium services that it previously offered — services such as Orkut, Notebook, Jaiku, Google Labs, etc. Although it can afford to do so, and is able to take the flak from closing these services, there is a lesson to be learned here.

It’s also worth noting that, in many regards, Google does not, in fact, offer free services. It mines the content of users and leverages this in its core advertising business. If it cannot generate sufficient customer volumes or rich enough data to make the resulting financial reward great enough, it walks away. Walking away for Google is one thing — for a small startup that has bet the farm on a single service, it means closure of the business.

Freemium’s five forces
At 451 Research, we see five key forces in play when evaluating the relative risks and rewards of launching and running a free version of a product.

Figure 1 – Five forces of the freemium business model

Cost per user
Probably the most important force at play in evaluating a freemium service is the actual cost of providing a service to a customer who does not pay for it. The fact is, computing power, storage, R&D, administration, etc. are not free to the provider. We know of a number of early stage startups that launched free services only to pull them almost immediately.

Volume
Volume is a tricky but essential parameter in the equation — having too few freemium customers provides too low a number of potential conversions; alternatively, too high a volume can mean a drain on resources to fund nonpaying users. Generating the right volume for a particular service is a major challenge. Some providers, including Dropbox, incentivize users to help increase this volume, by offering extra free storage in return for referrals.

Time to conversion
A major factor in assessing the viability of a customer (if not the service as a whole) is the time it takes to make a user of a free service either convert to a paid version or otherwise generate revenue to cover the costs of the service. For example, a user of an EFSS (enterprise file sync and share) system that uses the maximum amount of free storage for four years without converting to a paid version is a drain on resources over a user that converts within just a few months. Microsoft incentivized OneDrive users to help reduce this time to conversion by offering extra free storage when a device is configured to back up to the cloud platform; as a result, it’s likely more storage will be consumed by a user, forcing an upgrade to a paid package.

Value of conversion
Moving a freeloader to a fee-paying position is one thing, but once converted, the revenue the customer generates needs to not only cover current costs, but also the cost involved in formerly providing the free service; a profit margin would be good, too. Similarly, one has to consider the cost involved in converting free users when considering the overall value of a conversion.

Volatility
A factor that can cut two ways — high volatility or turnover of customers can mean that poor-quality leads qualify themselves out of the sales process relatively quickly, but on the other hand, it can mean you never really get the stickiness required to convert to paid versions. Many providers incentivize users to reduce this volatility by rewarding consumers with a discount for annual contracts.

Worked example
We can do a simple calculation to show the profit-generating power of freemium pricing. The graph below shows the price of FSS services offered by Dropbox, against the cost of AWS S3, which we know is used by Dropbox for storage. We assume S3 is Dropbox’s largest cost of goods sold, and others are negligible, which is a crude assumption, but nevertheless reasonable to demonstrate a point.

Let’s assume the worst case: that every customer using the freemium model is consuming the most capacity possible without paying a cent — 2GB; there are F free customers, costing a maximum of $0.06 to support on S3. We will also assume the worst case: that every customer using the paid model is consuming the most capacity they can without upgrading — 100GB; there are P paid customers, costing a maximum of $3 to support, but bringing in $8.25 per month.

For a profitable business:
Cost of supporting all free users < profit achieved from all paid users
F x $0.06 < P ($8.25-$3.00)
F / P = 87.5

So, in other words, in the worst-case, assuming other costs are negligible, Dropbox only needs to get one in 87 customers to pay across its user base to profit — a conversion rate of just higher than one percent. With regard to specifics related to Dropbox, we have to take this analysis with a pinch of salt, but it does show how relatively small costs provide the potential to upsell higher margins.

This example demonstrates how volume of users, the cost of delivering the service and the value of the conversion are all major factors in ensuring a profitable (and sustainable) service. What it doesn’t easily show is the impact of volatility and time to conversion. In the Dropbox example above, the company must ensure that it always has at least one in 87 customers paying. This is an ongoing challenge: Volatility means that customers may suddenly stop paying, meaning the provider must find new paying customers to keep the conversion rate acceptable; time of conversion means that even if new customers are found, it can take months before they will pay. These are all financial risks.

Lessons from the frontline
Over the past couple of years, a large number of startups have moved away from freemium pricing models and embraced limited-time trial periods. These vendors include Redbooth (formerly Teambox), Basecamp, Huddle, HootSuite, SugarSync and KnowledgeTree. It’s too early to know the real impact or value that this shift has brought about, but in our conversations, no one was looking to move back to the freemium service, and what data is available suggests no impact in terms of adding to the number of new paying customers, but a significant impact in reducing the bottom line. Some who are still offering freemium options have no plans to change; others are debating the long-term value. For example, Box noted in this year’s S1 filing that it was spending more on sales and marketing than it was making in revenue. A significant portion of these expenses (as for many other startups) can be attributed to the cost of supporting a ‘successful’ freemium model while converting a very small percentage to paid versions. We expect Box to address this in the short term at least by exploring means to lower hard costs by moving some free accounts to lower-cost storage, in time divesting itself of inactive and unconvertible freemium users.

On the other hand, Dropbox has strongly defended its use of the freemium model, and from what we can see pre-IPO filing, it does seem to have done a better job than some of limiting free options and providing enough incentive for a viable number of users to move to paid versions. The key here has been in providing a better mousetrap, insofar as the user experience and add-on features have been driven by customer demand while keeping a firm cap on the 2GB storage limit. Similarly, Evernote has seen success with the freemium model, converting (by some accounts) around 3% (and increasing) of users to the paid model. The connection between these two success stories is volume — both provide low-cost (even the premium version) applications that have exceptionally wide appeal to potentially any information worker.

Interestingly, one startup we spoke with told us how, one year ago, a leading venture capital firm was demanding firms have a freemium version of the product. A year later, the mood has changed considerably, with VCs becoming very wary of freemium versions. Indeed, it has become a common cautionary discussion point in Silicon Valley, both among startups and investors. Even so, we still regularly see firms coming to market driven by freemium models.

Advice to vendors and investors

• Consider carefully the real costs and risks of freemium versions of enterprise products and services; early ‘success’ in generating legions of users can become a longer time drag on the bottom line.
• Don’t base the company’s future on the assumption that gross margins will always be the same, particularly if the service is easily copied by a competitor. Giving away a lot when you have high margins is fine today, but when your competitors force you to cut prices, those free services will become a cost that is unsustainable. When this happens, cutting free services is only likely to result in backlash.
• If you have a freemium offering, at least consider changing over to a free trial period. This offers the same benefits, but with an automatic off-ramp for those users that are unlikely to convert to paying customers.
• Recognize that five customers paying $5,000 a year is better than 50 customers paying $100 per year. The bottom line is the bottom line, and a large volume of low-paying customers do not represent a viable long-term customer base.
• If you don’t have a plan to disembark unresponsive freemium customers (i.e., a plan B), then consider carefully the merits of embarking on a plan A of perpetual free services to freeloaders. Freemium can be good for startups looking for traction and validation, but you need to be in control of the throttle and brakes.
• Corporate needs differ widely from the needs of individual end users, who are solely interested in functionality to benefit them. Corporate needs to consider very carefully include security, compliance, integration, administration and a whole host of back-end critical requirements. A product that works well for one may not work well at all for another.
• Converting a number of disparate users within a single company into an enterprise-wide deal is not an easy undertaking. There is a substantial disconnect that needs to be overcome in taking unauthorized and unbudgeted applications and bringing them into a controlled and budgeted environment. The use of such applications can prompt an internal corporate discussion that a need must be met, but that is no guarantee that it will be your application that wins the day. Just as likely, an RFP-driven procurement process will identify higher-cost but more established alternatives that IT is more comfortable with.
• The freemium model can work very successfully, but you need to be realistic about potential conversion rates and recognize that, today, 1% or less is the norm, with anything above 3% considered highly successful. We regularly hear small startups telling us that they expect much higher conversion rates; in almost all cases, they are likely to be sorely disappointed.

The 451 Take
In summary, we can say that the freemium model can work, and there are examples of vendors that have seen some success. But even those that claim success are remarkably reluctant to share the full metrics externally. In reality, it does not work for many vendors, and through a mixture of overly optimistic conversion estimates and an inability to jettison poor-quality accounts, it can be a deadly trap to get into. For most vendors, a free trial period would make for a lower-risk option. Once the individual companies in this current crop of social business application startups get acquired, dissolve or go public, the true metrics of freemium models will see the light of day and become a deterrent for the next round of startups.

Note — For more information on this report or on subscribing to 451 Research programs, please contact 451 Research, or contact InsightaaS at reportinfo@insightaas.com.