Fusionex International — Update 15 September 2016

Fusionex is a global big data software business headquartered in Malaysia. It listed on AIM in 2012, and the £12m proceeds of the IPO have been used to support the development and launch in 2013 of GIANT, its big data analytics solution. The company now also has operations in Indonesia, Hong Kong/Macau, Thailand, Singapore, the Philippines and soon Vietnam, Greater China and Australia. GIANT enables organisations to collect and process their data no matter what the format, and to present the information via powerful and useful visuals for analysis. This is a market that is widely forecast to show dramatic growth over the coming years and GIANT has had considerable early success. Notable industry relationships include those with Microsoft, Dell, Hortonworks and Cloudera, as well as its indirect channel partners Avnet and Mesiniaga Berhad. Globally recognised customers include Intel, DHL and Las Vegas Sands Corp.

GIANT reported 79 customers in June 2016, up from 68 in March 2016, leaving it on track to hit its targeted 90 for 2016. Having proven GIANT can compete at the enterprise level against the larger vendors, management has now committed to scaling the product. Following the £14m fund-raising in October 2015, investment has been increased to enable Fusionex to raise its profile through additional sales and marketing initiatives, expand its international reach with new offices planned, and invest in the acquisition of storage to ensure it has the appropriate capacity to take on new customers at short notice. We believe the recent launch of GIANT 2016, a significant upgrade that includes a number of functional and user interface-related improvements in June of this year, will provide an additional impetus to growth and we forecast a CAGR in revenues of 35% for the FY15-18 forecast period. The increased investment means EBITDA margins will be compressed this year and next. However, we forecast margins to rapidly rebuild and a return to profitability at the net earnings level by FY18.

The shares have been marked down since the group announced the increased investment required to execute its more ambitious growth strategy. However, given the potential opportunity, we consider this strategy appropriate. We believe that EBITDA margins have the potential to grow to c 35% on a revenue base of c MYR500m. If this is the case, and assuming our forecasts to FY18, the current share price implies a revenue growth (FY19- FY25e) of c 10% pa.

Given GIANT is only three years old, the product is gaining recognition in a rapidly growing industry (20% plus pa) and customer wins are showing strong momentum (we expect GIANT sales to double this year), 10% growth appears to be rather conservative. Taking our forecasts to FY18 followed by 15% revenue growth to FY26, our DCF returns a value of 325p. We believe that the market and current momentum could support even more aggressive growth than this. However, before using this as a base case, we will need to see clear evidence that GIANT 2016 is well received by the market and that the indirect sales channel strategy, which is the lower cost route to scaling, is delivering a strong pipeline of new customers.

Key sensitivities include the MYR exchange rate against local trading currencies as well as sterling. There is also a trade-off between revenue growth and margins and we do not regard the exact margin structure as being set in stone. In addition, the big data analytics arena involves a number of evolving and interwoven technologies and Fusionex must keep up with the changing requirements of the market.

Fusionex provides big data analytics software to enterprises and SMEs, both within its local markets in South-East Asia and globally. Its main product is GIANT, a user-friendly, competitively priced product that gives users the ability to collect, process and analyse their data. Since its launch in 2013, GIANT has had good momentum, deployed by 68 customers in March 2016. The launch of version two of this software, GIANT 2016, in June, which added additional features and opens up the market to SMEs, is already seeing very encouraging early commercial success, as illustrated by the announcement of two major contract wins worth several millions, and we expect it to give further impetus to this growth.

Fusionex was founded in 2005 by a group of entrepreneurial IT professionals led by Ivan Teh. The initial focus of the business was on consultancy, systems integration and solutions, but soon Fusionex had begun developing its own software. The first product was the Core Transactional Engine (CTE), launched in 2009, followed by the Business Intelligence (BI) offering in 2011. In December 2012, Fusionex floated on AIM and management stepped up the geographic sales expansion and the investment in new product development. Subsequently, the company began building a global distribution capability with partners and, most importantly, it launched the GIANT big data analytics solution in late 2013. Fusionex no longer markets its BI and CTE products. The company’s focus is centred entirely on its big data platform.

The group is run by Ivan Teh, CEO and managing director (46.15% shareholding), a well-known figure in the Asian technology sector having won numerous awards for entrepreneurship and leadership. He has almost 20 years’ experience in the IT sector which, prior to Fusionex, included positions at Hewlett-Packard and Accenture. The CFO, Yuen Choong Lai, is a qualified chartered accountant with over 20 years of finance and investment experience. Fusionex has four non-executive directors who combine UK, Asian, financial and technology industry experience: John Croft (non-executive chairman), Robin Taylor, Calvin Chun and Alan Lim.

With more computing and mobile devices in use, the amount of data being generated is growing exponentially. This data is potentially incredibly valuable to companies. If properly curated, they can enable companies to develop an integrated view of customer activities and business operations, which can lead to competitive advantage, or operational efficiencies.

However, much of these data are hard to store, access and process with traditional database management tools. The data generated may be in such high volumes, in a wide variety of formats and received at a very rapid rate. Big data technologies can help release the value in these accumulations of data.

Big data solutions are of use across many industry sectors; however, according to global technology market data and analysis by the International Data Corporation (IDC), the largest industries for big data spending in 2014 were discrete manufacturing ($2.1bn), banking ($1.8bn) and process manufacturing ($1.5bn). The industries with the fastest growth rates include securities and investment services, banking and media. IDC forecasts that the big data technology and services market will grow at a CAGR of 23.1% from 2014-19, with annual spending reaching $48.6bn in 2019, and software and services accounting for roughly half of this. It predicts that the Internet of Things (IoT) will be the next primary area of focus for data/analytics services with a 30% CAGR forecast over that time period. Although definitions and levels of enthusiasm vary across the forecasters, these figures are indicative of the size and expectations for big data technologies.

GIANT gives users the ability to collect, process and analyse their data. Whether structured, semi-structured or unstructured, data can be consolidated in one place, processed (in parallel and on a real-time basis if required) and accessed in a user-friendly way, removing the need for any real knowledge of the plethora of file structures, processing, scripts or languages that are being applied.

GIANT has three key elements:

GIANT can be provided to customers either on their premises, on a hosted basis (by third parties or by Fusionex) or a hybrid of these options. It is sold either with an upfront licence fee plus ongoing maintenance, or on a SaaS subscription basis. Contracts are generally between three and five years. Fusionex has increased the effort to raise the proportion of its income from periodic subscriptions, rather than on a traditional licence basis and we understand that the vast majority of new GIANT customers over the last 12 months are contracting on this basis. Although this move depresses revenues in the short term, it moves the company to a smoother revenue and cash flow path in the longer term.

Three years after the launch of the original GIANT product, Fusionex has launched version two of its platform, GIANT 2016. Improvements include simpler and faster navigation, with natural language features added to the interface making the product easier to use. Support for R, a specialist programming language for data analytics, and Python, a general programming language that is increasing in popularity, have been improved. Fusionex has also added smart maps features, the ability to make parallel queries, along with streaming insights and more charting functions.

With GIANT 2016, management hopes to open up the big data analytics market to companies of all sizes by enabling functionality to be tailored appropriately (eg some of the high-end features like real-time processing are not required by all companies).

Launched in June 2016, GIANT 2016 has already announced three significant contract wins with an integrated holiday resort in Asia to use GIANT 2016 to optimise marketing strategy, a Japan-based international travel, leisure and hospitality group, to use GIANT 2016 to support its operations in the Philippines and, most recently, with Bursa Malaysia Berhad, Kuala Lumpur's Stock Exchange, where GIANT will enable it to keep track of the multitude of stock and price movements in the market.

Big data solutions are built on a growing number of crucial and evolving software elements (Exhibit 3). Typically, the cost to the user of setting up systems and understanding the various software elements is considerable. Big data projects with the major vendors, such as IBM, SAP, SAS and Oracle, generally command multi-million dollar price tags and yet the customer can still be left with a solution that is difficult to use and adapt. GIANT has been designed to be self-serve and intuitive to use by anyone in an organisation – not just those with database management or coding experience. It was recently made even easier by the integration of natural language processing (much like Google Search).

Based on average customer numbers, we estimate that Fusionex charges average annual subscriptions of approximately $300,000 per implementation compared to millions of dollars charged by the US and European industry majors to set up and license a big data solution. The SaaS version targeted at SMEs will have a considerably lower entry level price point. A key part of Fusionex’s ability to create and deliver products that are very cost competitive against the US and Europe-based major software vendors has been its lower cost of product development and, most importantly within this, the lower cost of local labour. Management states that it does not experience any significant problems in recruiting staff of the right calibre either in Malaysia or internationally, but some training is often required for some of the more specific applications and areas, particularly in the big data environment. There is also likely to be a cost advantage flowing from the fact that Fusionex’s product includes a number of layers in the big data supply chain.

There are many competitors for all of the elements of the GIANT product. However, products from the market leaders such as oracle, SAP and IBM combine a number of solutions. We are not aware of any similar solution that covers such a range of big data system elements in one package. In the area of data management tools companies like IBM, DataStax, Informatica and Syncsort have strong positions. In the visualisation and analytics area Tableau, Karmasphere, Datameer, Tresata and SAS Institute are among the leading names. These players in general have narrowly defined products and reach their markets only via larger resellers or software companies. These resellers will in turn place the product alongside other technologies, from their own library, from other vendors or from open source.

Fusionex’s customer list is a mix of quality names across a wide range of segments ranging across finance, electronics manufacturing, hospitality and media. Customers are both local and international names and include DHL, Las Vegas Sands, Starwood Hotels & Resorts, Lotus, Intel, Yeo’s, Aeon, Dell, TNS Group (Taylor Nelson Sofres), Brother Industries, Canon, Siam Cement, Air Asia, Hewlett-Packard, CIMB Bank, Takaful Insurance, MetLife, Resorts World Singapore, Volvo, Carrefour. GroupM, Jones Lang LaSalle, UEM Sunrise and Mitsui, among others.

Many of Fusionex’s implementations have been for local subsidiaries of global businesses that would typically be expected to choose solutions from the majors. Having proved itself capable at the local level, Fusionex is increasingly providing similar services to other local operations of these customers across South-East Asia and globally.

It has also received a number of accolades from industry bodies including Most Outstanding ICT Company in Asia by the Asian-Oceanic Computing Industry Organisation (ASOCIO) in 2015 and the Innovation Award at the 2015 Microsoft Worldwide Partner Conference. The inclusion of Fusionex as another “relevant vendor” for the first time in Gartner’s February publication of Magic Quadrant for Business Intelligence and Analytics Platforms is an encouraging signal that the company is making its presence felt on a global basis, despite its relatively small size.

Fusionex asserts that for a typical customer, it will derive approximately 10-30% savings on down time, 10-35% operation expenditure efficiencies, or drive double-digit increases in sales growth via customer profiling and product analysis. A good way to illustrate the power and potential of GIANT is by describing some of its applications.

The work undertaken with Intel is probably the highest profile and best documented of GIANT’s applications, with the implementation being the subject of a white paper. Intel worked with Fusionex, Cloudera, Dell, Mitsubishi Electric and Revolution Analytics (now part of Microsoft) on an IoT big data analytics project.

Semiconductor manufacturing is probably the most technically demanding manufacturing environment. Semiconductors, and particularly those that Intel makes, are high-value, ultra-high precision items that are manufactured in bulk on extremely expensive capital equipment in a highly complex process. Downtime can cost millions of dollars per hour, reworks and repairs of defective products are impossible and manufacturing equipment failure can be not only expensive but also dangerous. There is a vast amount of data generated within a semiconductor manufacturing facility. Historically, however, much of this data has been applied in a single use, for example to find out if a temperature is too high or too low, but big data technologies provide the ability to better consider these variables over time and to understand how they relate to each other.

Together with its partners, Intel gathered the data from its plethora of sensors over a period of time and looked at it alongside metrics such as equipment fail rates and product testing accuracy. A number of uses were found and applied over the life of the project, including predicting the failure of key components in manufacturing equipment, identifying where chips have not been successfully attached to substrates and using image analytics to identify and predict faulty products.

Intel’s use of Fusionex technology demonstrates that, despite its substantially lower price tag, GIANT can deliver an application where one might normally expect to find a US or European industry major’s product and, just as important, that global corporations with bleeding-edge technology skills recognise this.

Although the applications made by Intel are industry specific, it is clear that the streaming analysis of volumes of data and allowing users to simply manage and apply data is applicable across many manufacturing, process or extractive industries from semiconductors to food to oil. In fact, it is difficult to identify a large-scale manufacturing process for which it would not be relevant.

Fusionex has a GIANT customer that is a leading player in the Asian hotels and resorts market offering travel, hospitality, theme parks and conference facilities in a number of countries. Handling hundreds of thousands of visitors every year, both corporate and private, the company has a vast amount of information about its customers’ behaviour.

There are clear peaks and troughs in demand, across each day, week and year, and a high fixed-cost base. This means that there are significant potential benefits from managing occupancy and utilisation in hotels, theme parks and resorts. With GIANT from Fusionex, the client was able to bring together the masses of data to which it has access, from bookings, customer preferences, social media and even weather data, to analyse it and extract useful information and insight. As a result, the client is better able to predict and manage demand, yielding occupancy rate improvements of over 20%, with resulting operational efficiency gains of over 60%.

As for Intel, this is just the start for this client and, more importantly, for this application. We believe that this client intends to roll out the GIANT solution to a number of its sites across Asia as a result of the success it has seen with its initial implementations. Although we believe the scale of this client’s sites is large (including some of the biggest and best known in Asia), the relatively low cost of GIANT suggests that it is a relevant solution for worldwide hotels, resorts, conference facilities and theme parks.

The market opportunity is significant. Management is widening its sales capabilities by increasing its international presence directly, as well as by establishing an increasing number of indirect sales channel partners.

Fusionex makes most of its sales directly and works in tandem with Dell, Hewlett-Packard, Accenture, Microsoft, IBM and EMC for projects where these leading vendors see opportunities to increase their sales by working alongside or promoting Fusionex. These sorts of relationships now extend deep into the big data ecosystem. Fusionex also has partnerships and certifications with both Cloudera and Hortonworks, the leaders in providing big data platforms based on Hadoop, and Revolution Analytics, the leader in software and services relating to the R data analytics language. Fusionex’s c 15 sales representatives work directly with these partners and these sales are regarded as direct sales.

Since 2015 Fusionex has also been focusing on increasing its use of indirect sales channels as a way to cost-effectively scale its reach. The proportion of sales coming from indirect channels is currently 30-35%, and in the past management has indicated that it hopes to see this increase to 70%. The impact of this on gross margins could well be negative, if margin is sacrificed to partners, but the increased revenues should more than offset the added cost. In mid-2014, it was announced that the global technology distribution giant Avnet had joined Fusionex in a partnership arrangement and it has since announced partnerships with Mesiniaga Berhad (one of the biggest ICT solutions provider and systems integrators in Asia) and VADS (one of Malaysia’s leading integrated managed ICT service providers).

Establishing channel relationships is a key step in growing the overall customer base. However, motivating these channel partners to actively market a vendor’s products can be harder. Fusionex has decided to support its product visibility by launching a significant digital marketing campaign, and enhanced use of conferences and exhibitions to drive awareness and sales.

In FY15, 24% of sales were from beyond the ASEAN market, compared to 21% in FY14. While Asian markets are considered to be at a fairly early stage of development when it comes to the deployment of big data solutions, it is evident that interest for big data solutions in ASEAN nations is growing, as evidenced by the first Strata Hadoop World, the leading global big data industry event, in Singapore in December 2015. Since the IPO in 2012, the company has established offices in Indonesia, Hong Kong/Macau, Thailand and Singapore. We believe that the recent fund-raising will be applied in part to taking the international expansion a step further, with the establishment of a significant sales and development presence in the Philippines and plans to expand into Australia, Greater China and Vietnam.

Fusionex is still a relatively small company, operating in a rapidly evolving industry, so it is difficult to predict its future path or its share price. There are several key sensitivities for both the company and its shares that should be considered.

Over the last three years Fusionex has transformed its revenue profile from predominantly the legacy CTE and BI solutions (100% of revenues in FY12) to being almost entirely from its big data analytics solution in FY16e (a breakdown of revenues by product is not given, but we estimate c 80% of revenues will be from GIANT in FY16). Management has taken the strategic decision to no longer actively promote the legacy products – despite this drag on revenues, Fusionex has achieved a CAGR of 35% over the last three years to 2015.

Progress with GIANT has been impressive. From 12 customers at the end of FY14, it reported 36 at September 2015, 68 in March 2016 and 79 in June 2016, and with “no sign of slowing down” (27 June statement) Fusionex appears on track to hit its targeted 90 for FY16. With the additional impetus that we expect to arise from GIANT 2016 and the SME market, coupled with the more ambitious sales and marketing effort, we expect the strong growth of GIANT to continue and forecast a CAGR in revenues of 35% for the FY15-18 forecast period despite the impact the discontinuation of the legacy products will have in FY17.

The shift in product mix towards the “easy to implement” GIANT would suggest that services revenue (12% FY15) might decline sharply, but this has not entirely been the case to date. Bringing on board new channel partners does require a significant level of initial support from Fusionex. Our forecasts assume services revenues continue at broadly the same absolute level going forward.

The majority of costs relate to staff (c 60% of which are devoted to product development which is capitalised), amortisation of R&D expenses (principally staff costs), data hosting costs, marketing and general overheads. Hosting and marketing costs are set to see a steep increase in 2016 as Fusionex raises its profile and invests in the acquisition of storage to ensure it has the appropriate capacity to take on new customers at short notice.

Operating expenses more than doubled in H116 and we expect a full year increase in operating expenses of 120%, pushing Fusionex into losses this year. Looking ahead to FY17 and FY18, we do not expect this kind of step change in the cost base to be repeated. Upfront investment in hosting may be required each time Fusionex enters a new market, otherwise costs should be more closely aligned to the increase in sales. With a c 74% gross margin, this profile of operating expenses should support strong operational gearing effects and we forecast a return to operating profitability by FY18.

As a growing technology company that is investing heavily in product development, Fusionex benefits from a number of tax and other financial incentives, notably the tax exemption that goes with its Multimedia Super Corridor (MSC) status in Malaysia. The MSC tax status is due to expire this year; however, in our forecasts we assume that this will be extended and forecast a 12% effective tax rate, in line with FY15.

Fusionex reports in Malaysian ringgit and the share priced in sterling. It bills its clients in local currencies, while most of its costs are ringgit based. Over the last year the ringgit has appreciated against most of its trading currencies, as it has against sterling. The accounting impact of this is to reduce ringgit revenues, with a significantly smaller movement in the costs lines. This may add to margin pressure, and potentially erodes some of the price advantage Fusionex has versus the majors (which typically base prices in US dollars). However, as this appreciation has been gradual and reverses the sudden depreciation seen over the summer of 2015, we are not overly concerned.

In addition to the forecast operating losses, cash absorption is principally affected by:

Fusionex reported net cash of MYR105m at March 2016 (£19m). We believe this should provide adequate funding through the forecast period.

For high-growth technology companies that are in an investment phase, we prefer to use a DCF as the basis for our valuation.

We use a DCF to estimate what the shares could be worth under varying medium- and longer-term revenue and EBITDA margin assumptions. Our DCF model assumes:

Exhibit 6 presents what different medium term revenue/growth and EBITDA margin mix scenarios could mean for the share price.

There are of course many ‘ifs’ in this kind of analysis. We believe it is important that investors consider the following:

Margin profile: in the past, Fusionex enjoyed EBITDA margins of 45% to 50%. Given management’s ambitions to scale Fusionex, returning to margins of this level on a sustainable basis is unlikely. However, as Fusionex scales, efficiencies are likely in the cost of hosting, sales and in the more discretionary marketing costs. We believe EBITDA margins of c 35% could be comfortably achievable on a revenue base of MRY500m (approximately £100m). 35% is consistent with the EBITDA margins seen by a number of larger global software vendors and the medium-term target figures referred to by leading pure-play SaaS vendors. Under this scenario, the current share price appears to be discounting that revenue growth can be maintained over the medium term at c 10%.

Early stage of the market and Fusionex: GIANT is a relatively young product and Fusionex is early in its strategy to widen its reach through marketing, geographic expansion and increased use of indirect channels. We forecast a relatively high share of R&D capital expenditure and it is probable that additional products and feature sets will be launched. If GIANT 2016 proves a success, as we anticipate, particularly with the SME market, we would expect to see considerably higher growth sustained through the medium term than that implied by the current share price.

Revenue visibility: with contract lengths of three to five years, and an increasing share of customers signed on a subscription basis, there is relatively good revenue visibility.

Base case DCF of 325p: while deciding the mid- to long-term revenue growth rate is fairly arbitrary, as outlined above, 10% seems low for the stage of development of GIANT (where we expect revenue to double this year) in a rapidly growing market. If we assume our forecasts to FY18 followed by c 15% CAGR in revenues to FY26, with margins peaking at 35% (in FY25), our DCF returns a per share value of 325p. We believe that the market could support more aggressive growth (at 20%, our DCF returns 474p), but we need more visibility of partner bookings coming through strongly before using this as a base case. Peer multiples comparison has only have limited use

Like Fusionex, many of its listed peers are currently in an investment phase, and so peer multiples comparison adds little, other than to show that on EV/sales multiples of 2.3x in FY17, Fusionex trades at the lower end of the range (Hortonworks at the lower end on 1.5x and NetSuite at the higher end on 7.2x).

While both valuation methods are highly subjective, they serve to demonstrate the significant upside potential to the share price if Fusionex can execute on its strategy to scale GIANT.

Given GIANT 2016 has only recently launched and the new emphasis on the SME market, the coming months will be very important in determining where medium-term growth expectations should be set. Initial evidence is encouraging; in its 27 June statement, Fusionex announced 79 customers stating that “it is showing no signs of slowing”. It has subsequently announced three new customers, putting it on track to reach its target of 90 by the end of 2016.