In January and February, I attended two major technology conferences - the Needham Growth Conference in New York and the Mobile World Congress (MWC) in Barcelona. The Needham investor conference is a quality, three-day venue to have meetings with/attend presentations of the leading emerging public and private technology companies. The MWC is the largest annual mobile industry trade show, whose size and scale can be compared to the Consumer Electronics Show in Las Vegas.
Cloud computing has been a theme that has struck a chord with many investors for several years (see here for our first Cloud Computing article). Cloud Computing can be broken down into two main categories: Big Data and Software as a Service (SaaS). Big Data replaces more expensive/less flexible in-house servers and storage with public and private out-sourced storage facilities (a.k.a. "clouds"). More recently, Big Data has incorporated software companies with mega-data analysis tools. Part of the "big" data generated is machine-to-machine (M2M) data coming from sensors or meters that today can be stored and analyzed in the cloud.
SaaS replaces more expensive/less flexible enterprise software with software subscription platforms that are hosted in the cloud. Salesforce.com (CRM) typifies a SaaS company. Platforms are also hosted on the cloud (PaaS) with development tools, databases, etc. like Microsoft (MSFT) Windows Azure. Infrastructure-as-a-service (IaaS) with servers, storage, and load-balancers like Rackspace (RAX) are more akin with Big Data.
Cloud Computing Segments
Big Data starts with current powerful network processors such as the Romley chip from Intel (INTC) and super-fast interconnects from Mellanox (MLNX), that are run on EMC (EMC) servers, using VMware (VMW) operating systems, sitting on huge data farms from Amazon (AMZN) and Google (GOOG), that are transported on high-speed broadband networks. The establishment of this fast and powerful infrastructure is now enabling mega-data analysis tools from companies like Splunk (SPLK), Sumo Logic, Teradata (TDC), Attunity (ATTU), Datawatch (DWCH), Factual, and Recommind to analyze/replicate this disparate and often unstructured data in real time, and create an ROI for customers. In fact, companies like EMC Greenplum, Amazon Redshift, Microsoft, Google, IBM (IBM), and Oracle (ORCL) are combining data repository capabilities with data analytics. The strongest growth potential lies within these Big Data software companies.
The key term for data analytics is real time. A study below by Nucleus Research showed that half the value of tactical data is lost within an hour, and the half-life of operational data is about eight hours. Only strategic data is less time-sensitive with a half-life of two days. Until recently, it did not matter, as we did not have the tools to transport and analyze such massive amounts of data in real time. But today, for the first time, we actually can analyze mega-data in real time.
(X-axis is hours)
The size of this nascent Big Data software opportunity is substantial. Sumo Logic's CEO estimates that only unstructured data (mainly machine-generated) for security and infrastructure/application management has a TAM of $10B - $12B, which companies like Splunk and Sumo Logic are only scratching the surface of today. From the viewpoint of value created by Big Data software, Nucleus Research comes up with a figure of $106B that would be shared with the end client. Thus, this is a large market opportunity.
Three interesting Big Data software companies that I met at the Needham conference were Attunity, Datawatch, and Recommind. Attunity is an enabler of the real-time analysis of mega-data. The company seamlessly transports large data files via data replication, file transfers, and data synchronization. Attunity has partnerships with Microsoft, EMC Greenplum, and Amazon Redshift, and works with competitors Teradata and Oracle Exadata (the main competitor being Oracle's Golden Gate division). Attunity's competitive advantages are 1) fully-automated, in-memory processing, 2) Attunity Replicate is the leader in a potential $1B data replication market, 3) independence (versus Oracle/Golden Gate that competes with its customers), and as a by-product 4) key partnerships with Big Data industry leaders.
Attunity's revenue jumped 68% in 2012 to $25.5M, with GAAP operating margins of 10%. The business model scales, so margins should expand as revenue ramps. The first half of 2013, however, will show slower top-line growth and higher marketing and related support costs. The recent commercial launch of Attunity CloudBeam, a managed data transfer platform initially targeted for Amazon Web Services, will take time for revenues to ramp due to its SaaS subscription model. The EMC Greenplum partnership, will also take time to accelerate. For the full year 2013, revenue should rise 20% to 30%, albeit back-end loaded, to over $30M. Longer term, enabling/riding the Big Data wave plus partnering with the industry leaders, will have an outsized impact on a small micro-cap like Attunity. Risks include growing a global sales and support team to meet the needs of this fast-moving market, and pricing limitations as Attunity's main role is to transfer data, which has a lower place in the value chain than data analytics. Software analytics retains the data, and creates the final solution for the end customer. Still, enabling the real-time movement of petabytes of unstructured and structured data is a key value-add, that combined with limited competition, will leave Attunity with quite comfortable margins.
Datawatch is in the more crowded but higher-value space of Big Data analytics. Information optimization (previously referred to as Business Intelligence) increased from 67% of revenues in FY 2008 to 95% in FY 2012 (September year-end), as content management and professional services faded away. Datawatch provides access to all types of data within the enterprise, as well as report analytics, whether that data is structured, semi-structured or unstructured. According to Gartner, 80% of all data generated today is unstructured (which includes machine-generated data). Traditionally, large relational databases dealt mainly with structured data. Datawatch can fill this gap with unstructured and semi-structured data access and analytics.
Management pointed out that highly-touted Splunk deals only with unstructured data, as opposed to Datawatch's full gamut of data capabilities. In Splunk's defense, however, its use cases are wider in scale and fairly simple to use, with highly evolved data logs and indexing that allows charging mainly on a data volume basis. Datawatch, however, is developing "solution blueprints" that target customers' data pain points, which should be easier to use than previous versions.
Datawatch is already in most Fortune 500 companies, which is a distinct advantage. The challenge is to change its low-price, limited functionality image. Most customers take Datawatch's software results as a base input for further business intelligence analysis. 70% of revenues are customers paying $2k/seat for desktop-only usage. 30% of revenues are faster-growing $50k - $100k/server plus $500/seat, using newer products. If management can convert more users over to newer products and successfully launch "solution blueprints," this stock can go higher. But the jury is still out, and competition is intensifying.
Aside from Splunk, other competitors include Actuate (BIRT), Informatica (INFA), Pervasive Software (PVSW), Progress Software (PRGS), Qlik Technologies (QLIK), Tibco's (TIBX) LogLogic product, Hewlett-Packard's (HPQ) Arc Sight product, Sumo Logic, and Cloudera, among others. Waiting in the wings are a host of larger business intelligence software providers that may not yet have full unstructured data capabilities, including BMC Software (BMC), CA (f.k.a. Computer Associates) (CA), Oracle/Hyperion, SAP/Business Objects (SAP), and IBM/Cognos.
Datawatch revenues were in a downtrend from FY 2007 until FY 2011 for two reasons. First, the company transitioned out of its slower-growth non-Business Intelligence segments. Secondly, Datawatch changed over to a more direct sales effort under the new management team starting in FY 2011. FY 2012 sales, however, rose 45% to $26.0M, with license revenue jumping 70% to $16.8M. FY 2012 operating margins were only 6%, mainly due to a doubling in sales & marketing expenses. FY 2013, however, started out slowly with only a 9% y-o-y revenue advance in the December 2012 quarter. FY 2013 revenues are projected to rise 19% to $31M, but margins will contract further as sales and marketing expenses rise again. Management is willing to sacrifice short-term profits in pursuit of long-term growth, given the large untapped market potential today.
Big Data SW Cos
On a revenue value/growth trade-off, Attunity is the most attractive stock in the group, followed by Datawatch (NB: profit bases are too small and volatile to use for PEG ratios for most of these stocks, plus top-line growth is more highly-valued as a proxy to best capture this hyper-growth segment). A few caveats - both Attunity and Datawatch are hitting short-term transitions, which puts them in the penalty box until visibility hopefully improves for 2H 2013. Secondly, these two stocks are micro caps, and may warrant a small discount for their lower liquidity. Attunity has no institutional research coverage, partly because it was on the OTCBB market until last summer. Attunity should go higher once coverage is initiated, and institutional investors pile in.
Recommind is a private company with a new database format that is not only relational, but is columnar, machine-learning, and contextual. Its Context Optimized Relevance Engine (CORE) solves the problem of how to efficiently process and create action items from the current avalanche of diverse data sources, including multimedia and text. Management then built applications on top of CORE, creating end-to-end eDiscovery solutions. Management estimates the eDiscovery TAM to be $3B to $10B. The first vertical addressed was the legal profession. The ROIs are impressive, especially on time saved on legal searches. Banking is now coming to the fore. Revenue jumped from $23M in 2010 to $46M in 2011 to $70M in 2012, and the company is now profitable. An IPO is likely by 2014.
An impressive Big Data storage network company is privately-held Data Direct Networks. I serendipitously met the COO, Erwan Menard, on a very long lunch line at MWC in Barcelona. You know a company is doing well when it has more attractive business requests than resources. Finally, the potential business was large enough for the COO to come to MWC, even though Data Direct Networks was not exhibiting. Data Direct Networks provides secure, extremely fast, high-capacity storage networks. Data Direct Networks' first customer was NASA in 2000. The development effort to meet the large data demands of the US government before the Big Data rush, provided the company a huge lead over competitors like EMC today. The company excels in data-intensive environments like financial transactions, homeland security, etc. 2012 revenues were $265M, up from $226M in 2011. The company is profitable, and does not need to do an IPO, according to the COO.
M2M is a subset of larger Internet of Things, as well as the Cloud Computing movement. M2M uses sensors to measure something (temperature, physical stress points, inventory levels, traffic flows, etc.) which is then transmitted through a communications network onto a software application that analyzes the captured data. M2M has been around for decades (e.g. industrial "automation" from Honeywell (HON)), but is now becoming viable with the advent of high-speed, Internet-protocol broadband networks, faster, lower power-consuming chips, and huge cloud storage facilities. The M2M market, defined as modules, value-added services and connectivity, is expected to grow at a 30% CAGR from $4.0B in 2010 to $17.1B in 2016. The majority of the revenues come from connectivity (+24% CAGR), with modules (+13%) and value-added services (+43%) growing from a smaller base.
Telit Communications (TCM.L, OTC:TTCNF) and CalAmp (CAMP) were two interesting M2M companies that I met at the Needham conference. Telit is a pure-play M2M sensor module maker with 20% global market share and $207M in 2012 revenue. Telit has many blue-chip customers ranging from Telefonica (TEF) to CalAmp. Telit M2M module shipments jumped 46% last year. Revenue rose a still impressive 17%, despite price declines. EBITDA rose 30% last year; but while EBITDA margins improved, they were still only 8% in 2012. In order to boost margins and recurring revenues, Telit is targeting M2M connectivity and longer-term, value-added services, which will begin contributing to revenue in 2013.
CalAmp derives 77% of current revenue from Wireless Networks and Mobile Resource Management (MRM) with its mature, legacy Direct Broadcast Satellite (DBS) business accounting for the remainder. Revenues for the fiscal year ended February 2013 are projected to advance 29% to $179M, aided by acquisitions. FY 2013 EBITDA margins should rise to 12% versus just 7% in the previous year. CalAmp provides total network connectivity solutions via its traditional hardware/wireless systems expertise plus acquired software applications.
CalAmp is morphing into a whole new company, with 20% of revenues coming from SaaS mobile data applications ranging from fleet tracking to usage-based auto insurance. The biggest risk is the integration of its several current and future acquisitions. In particular, one needs to be comfortable that management can execute on this major transformation from a hardware supplier to a solutions company, including software applications serving many vertical markets. To achieve this goal, management is acquiring much of this software competence, as it is not in the company's DNA. Since this transition started in 2007, management has executed. But the transition is only halfway there.
If management succeeds in finishing the transition, the stock can go higher. MRM (just under half of revenue today), which includes SaaS mobile data applications, is expected to grow 2x faster than wireless networks verticals and 5x faster than DBS, with well-above average margins. Management told me that CalAmp potentially can be a $500M revenue company in a few years with EBITDA margins of 20%. A few years may actually be four years at a 30% annual growth rate; but the potential is there.
Investors just need to be aware of the challenge. This is not a simple digestion of acquisitions. This is a total company transformation, increasingly competing against native software companies that are more focused. IBM pulled off this transformation from a hardware company to a solutions (software and services) firm. So there is a precedent, and CalAmp management has succeeded so far. But the risk remains (please see here for a more in-depth discussion on technology transformations and disruptive technologies).
Year-to-date, the stocks of both Telit (+46%) and CalAmp (+32%) have had good runs (CalAmp's price already rose 96% in 2012). Yet their valuations remain reasonable, with CalAmp trading at EV/Revenues of 1.8x and a P/E of 24x CY 2012 EPS, and Telit at 0.7x and 20x. The closest comparable to CalAmp's new business model would be Honeywell, which trades at 1.6x revenues and 20x 2012 EPS.
Software-as-a-Service is an investment theme that is starting to get long in the tooth. There are always new verticals that have not been penetrated like real estate - Trulia (TRLA) - comes to mind. But most successful SaaS companies are not cheap. Some corporations have already switched over to subscription-based services, instead of more expensive annual license fees, at least on non-core applications. In baseball terms, I would put SaaS in the fifth inning in terms of grabbing the low-hanging fruit. Big Data is in the top of the second inning.
SaaS is disruptive. Several software companies with up-front licensing business models are going through painful transitions. Companies like SAP are at risk. SAP recognized this dilemma by paying $3.4B for Success Factors at the end of 2011, but this will only soften the blow.
An interesting SaaS investment strategy is to find companies with evolving SaaS models, including growing recurring revenue streams and rising deferred revenue balances that are not recognized as such, and are undervalued. One company that fits the bill is Esker (ALESK.PA, ESKEF.PK). Esker provides document process automation solutions, which automates billing with the ability to create a paperless work flow. The company's SaaS business, Esker on Demand, has now grown to half of revenue, excluding a large deferred revenue backlog. Normally, investors pay high valuations for the visibility and profitability of such a recurring revenue stream. But Esker is only valued at EV/Revenue of 1.3x and EV/2012 EBITDA of 8.1X, despite solid top-line growth, which is even higher if one factors in the growing deferred revenue backlog, and EBITDA margins that have jumped from 8.1% in 2010 to 12.6% in 2011 to approximately 16.5% in 2012.
Additional disclosure: I am also long ALESK.PA