The following is an evaluation of the Carbonite business model relative to that of Dropbox. It is an attempt to draw comparisons between two very different approaches to the same problem; consumer/SMB online data protection and file sharing. There is no attempt to compare relative performance, reliability, availability, restore time/quality, or security of the offerings or related service level agreements. These are extremely important factors that will manifest themselves in brand value and reputation. However, attempting to quantify the key differences in these factors with the limited information available is of little value (at least from this author). Simply stated, these online storage services are good enough until they aren’t. As quickly as these services can build affinity with their respective communities, it can be taken away even faster with one minor outage or mishandling of a critical customer issue. Ironically, Dropbox is battling through a data privacy policy within its community as we speak. From my perspective, and given the data available, the more interesting exercise is evaluating the strategy and economics of one model versus the other. That said, it should be clear that if either company were to encounter issues around any of the aforementioned factors, it would clearly skew the offering's attractiveness, independent of the economics.
Opening Up the Kimono
The recent Carbonite IPO filing, and related S1 documentation, has opened the kimono on the economics of online backup. Until now opportunities to get a real handle on the business model of the online storage/file sharing business have been limited. Vendors, both new and old, are happy to provide customer adoption metrics and the number of objects/files stored as an indicator of uptake. But extrapolating more precise details about the business model has been nothing more than “fun with numbers”. Fortunately for us, Carbonite’s S1 filing (shown here) is chock full of data points and financial metrics that enable a much deeper glimpse into the economics (or lack thereof) of this business. The data seems to reveal a stark contrast between the traditional online backup model and that of the newer generation of “freemium” file protection/sharing services like Dropbox.
Different Solutions to the Same Problem
Although Dropbox isn’t considered “backup” under the traditional definition, it is still being used to solve the core customer problem of data protection. Interestingly, Carbonite seems to disagree with this notion, omitting Dropbox from its list of “direct competitors” in its S1 filing. The document instead references Prosoftnet, CrashPlan, Mozy (a division of VMware), Symantec’s Norton Online Backup, McAfee Online Backup, and SOS Online. However, one of the working assumptions behind my comparison here is that average consumers who need secure online file sharing, viewing, and data protection aren’t discriminating between backup and redundancy anymore. Instead they want an easy on-ramp, seamless user experience, and constant data access and availability. As long as the service delivers on these promises, its formal definition matters less.
Old School vs. New School
For the incremental customer evaluating the different options for data protection, a tale of the tape provides some useful insight. The table below breaks out some important metrics to help outline the key components, and resulting adoption profile, of the respective business models. The graphic presents a seemingly uphill battle for Carbonite. Despite being in business two years less, customers are opting for the freemium model of Dropbox at a 25:1 clip over Carbonite. And Dropbox is driving this adoption with 70% fewer employees and 90% less venture funding.
The Tale of the Tape
Source: Wikibon 2011
Carbonite Financial Details
Carbonite was founded in 2005 to provide online backup for consumers and SMBs through its Carbonite Personal Cloud infrastructure. After a recent 7% price increase, Carbonite’s entry-level consumer offering starts at $59/year for unlimited capacity. Customer count at the end of the recently ended March quarter stood at 1.043 million, up 49% from the same period last year. Since 2007, revenue has grown at an impressive compound annual rate of 162%. The company generated $38.6 million of revenue in 2010 (+102% over 2009) at 57.8% gross margins. Bookings, a more accurate gauge of traction, were $54 million in 2010, up 65% year-over-year (119% CAGR since ’07).
Unfortunately, it cost the company $65 million in COGS and operating expenses to achieve this revenue and bookings growth. This heavy expense structure resulted in an operating loss of -$26 million. The bulk of this expense came from branding and customer acquisition. Sales and marketing spending composed 85% of sales or $33.1 million, including $23.6 million on advertising alone.
The Cost of a Customer
Carbonite relies heavily on driving customer acquisition through on-air endorsements from 49 national radio talk show personalities, including Glenn Beck and Rush Limbaugh. Under this unique strategy, the company added 361,000 new customers in 2010. At the above mentioned sales and marketing spend ($33.1 million), this equates to a staggering new customer acquisition cost of $91.68 per new customer. More surprising, this metric is actually increasing, +34.5% compared to 2009 and +16.0% versus the ’08 level. The per customer advertising component of this spend was $65.37 in 2010, +87.0% over the prior year. Calculating a breakeven timeframe using these per customer investment assumptions yields an investment recovery period of more than three years (see Figure 2 below).
Carbonite – Simplified Per Customer Breakeven Analysis
Source: Wikibon 2011
Customer Lifetime Value
The Carbonite model appears dependent on buying as many customers as they can through advertising and then maintaining its 97% retention rate against this installed base growth. Ideally brand value increases over time, driving a more-than-commensurate decrease in customer acquisition cost. Clearly the company must have some lifetime customer value math that rationalizes the investments it is making. Unfortunately we are not privy to these calculations in the most recent S1 filing. The biggest concern with any set of assumptions here is the company never gets off the acquisition cost treadmill, because the competitive intensity continues to ratchet higher. In this scenario, financing the continued revenue growth in search of eventual profits is unsustainable.
The Cost of Freemium
In comparison, the Dropbox model relies on driving as many users as possible to the service and then converting them to paying customers as they exceed the 2GB capacity threshold. With seemingly little advertising spend, the key customer acquisition cost for Dropbox is the referral bonus they offer to existing customers to entice others to join (presently 250mb free space per referral). At published Amazon S3 storage rates of $0.14/GB/month this free capacity bonus prices out to a “sales & marketing” cost of $0.035/month or $0.42 per year. To be conservative, this amount needs to be tripled to account for the fact that every one copy of user data is stored 3x by Dropbox (ignoring backend de-duplication). Based on these estimates, the burden of the Dropbox referral program nets out to $0.105/month or $1.26/year. This compares to the previously calculated new customer acquisition cost for Carbonite of $91.68.
Ripe for Conversion
With 25 million customers in hand, Dropbox has a ripe conversion opportunity. Every 1% of this current base it converts to paid (at the minimum subscription level) is at least a $30 million/year revenue run rate (250k users * $9.99/month * 12). Using the S3 pricing from above, the maximum cost to maintain a free customer (assuming 2GB storage per customer) before he becomes a paying customers is $10.08/year (for 3 copies of all S3 data). If all 25 million customers were 2GB or less (i.e. not paying customers) this annual storage burden would equate to $252 million. Said another way, to break even under this set of assumptions would require an 8.4% conversion rate from its existing 25 million customers. One potential flaw in this sequence is that S3 has a tiered pricing model based on capacity. With significant capacity consumption it is reasonable to assume that Dropbox is getting at least a 25-30% discount on the above pricing of $0.14/GB/month. These discounts would decrease the infrastructure cost burden and conversion rate needed to break even.
Cost of Service
From an outside perspective, another apparent competitive disadvantage for Carbonite is cost of service. Carbonite runs in its own equipment collocated at two sites in Boston, MA. The company has also built and maintains its own file system. Interestingly, Carbonite provides only disk-level redundancy of a customer’s data (I assume this means mirrored RAID). Instead of keeping a separate redundant copy of data, Carbonite instead relies on the primary copy on the user device as its redundancy strategy. By comparison, Dropbox is based on Amazon’s S3 object storage service which is known to make at least three copies of user data. I assume Dropbox then layers its file system software and user interface on top of S3. While lacking the data to make a more precise conclusion here, there would appear to be a material cost-of-service (including opex and capex) discrepancy between the Carbonite hosted SaaS offering and the Dropbox public cloud-based SaaS approach.
Development Leverage
From a feature/function development standpoint, Dropbox has chosen to leverage the developer community to enhance and extend its offering. This effort has yielded more than 200 add-on utilities, including note/text editors, email integration, password management and others. Carbonite does not appear to have any such open developer program in place. As Salesforce.com, Apple, and Google/Android have demonstrated, creating a vibrant developer following provides meaningful leverage. With the base level service established, the opportunity exists for Dropbox to now layer on internally developed and third-party utilities to deliver enhanced functionality and close any perceived feature/function gaps that may exist.
The Carbonite Response
Beyond increased investments in advertising, Carbonite’s response to competitive pressure has been to expand its customer focus into SMBs, extend its functionality to more supported platforms (e.g. Apple), mobile viewing applications, and protectable devices (external hard drives). Carbonite is also making a push outside of the United States into Europe and China. Additionally, the company has increased its focus on providing the highest quality customer support. Related to this, it is bringing its customer support offerings back onshore from India to Portland, Maine. Collectively, these initiatives appear to be focused around increasing the Carbonite TAM (total addressable market) while ensuring customer satisfaction and sustaining retention at an impressive 96%-97%. However, the key to the company’s future prospects is an ability to improve, or alter, its customer acquisition cost model.
Give Freemium a Chance
Given the investment burden of the existing model, it makes sense to evaluate the impact on Carbonite of cutting all new sales & marketing spending and harvesting its core customer base. Applying some reasonable assumptions for this exercise (97% customer retention rate and per customer revenue, gross margin, R&D and G&A equal to 2010 levels) turned up a very interesting conclusion. The resulting model yields a projected 2011 operating profit of $11.6 million. This compares to a 2010 operating loss of $25.90 million. This begs the question, why not ditch the old model and go freemium. Sure this transition, and related messaging, would need to be handled delicately, but it is worth thinking through. All customers below a certain threshold would now be free. This would certainly impair near-term revenue; however it would also help alleviate the seemingly unsustainable customer acquisition cost burden and potentially result in a more attractive financial model.
Why Now?
So with these types of questions at hand, why is now the right time to IPO and subject the company to the scrutiny of the public eye? Outside of investor liquidity, and before the economics of this model are further compromised, why is Carbonite going public? Two viable conclusions are cheaper access to capital (i.e. lower cost of money needed to finance more customer acquisition) and improved brand awareness (i.e. reduced customer acquisition cost).
Regardless of how cheap capital becomes, does the math behind this model ever evolve to the point that this is a profitable long-term business? The pricing models and architectural approaches of the newer generation of designs are vastly different now then just six years ago when Carbonite started. Might it be better to stay private and adjust to the reality of the online data protection business? Dramatic business model pivots never go over well in the public eye, especially for volatile new issues.
Stay Tuned
In conclusion, while some will argue the usefulness of comparing these two services, I think this is missing the point. The consumer/SMB data protection market is evolving to one that is driven by experience, with users much less concerned about the definition of the approach. And while services like Dropbox are still quite new, they have rapidly achieved real scale in a very short period of time. It will be very interesting to watch as Dropbox now looks to evolve and monetize this core footprint with enhanced services. Meanwhile Carbonite, under even greater pressure to respond, surely can’t stand still. If nothing else, going public will force it to realign its business with the current market realities faster.
Disclaimer: I use Dropbox for my personal and business use. At the time, the service represented an easy and cost-effective (i.e. free) on-ramp to a data redundancy strategy. Part of the reason I recently undertook this analysis was to make a more informed decision in developing out a comprehensive data protection strategy. Regardless of the underlying brands, I set out to better understand the major differences between the approaches taken by the more traditional online backup vendors and that of the newer generation of file/data protection offerings. For customers, analysts and investors evaluating these companies and services, I believe it is critical to take a long-term view of the viability and extensibility of the respective business models and all that they entail. This is what I have tried to accomplish here. Please let me know where I missed the mark.
Action Item:
Footnotes: Dave Cahill is a Wikibon contributor and founder of Diligence Technology Advisors. Diligence is a boutique strategy consulting practice with a focus on storage, cloud computing and emerging technology opportunities. Dave can be reached via email (dave@diligencetech.com) or twitter (@dcahill8).
