Software Development in Pakistan: Lessons From The Trenches

Software Development in Pakistan: Lessons From The Trenches
By: Athar Osama

In the first of this series of articles, we looked at a statistical snapshot of software development in Pakistan. While our sample is not representative of the entire software industry in this country, it nonetheless includes a vast majority of the more significant players on the Pakistani software scene. Hoping to develop a better understanding of the structure and dynamics of the Pakistani software scene, not only facilitate investment and entrepreneurial activities but also to better inform public policy debate, we reported a number of fairly interesting findings. For example, although last year has seen considerable growth in both revenues and employment, Pakistan’s software industry is still going through a period of early-stage development and shows lack of maturity in many respects—that is only recently beginning to change. The most significant evidence in support of the above, that the study was able to bring to light, was lack of differences between domestic and export-focused software operations, between product and services-focused operations, and between pre- and post-dotcom operations—differences that one would have expected in a relatively mature industry.

We also found the prevalence of several best “managerial” best practices that differentiates the relatively better performing software houses from the not-so-good performing ones. The study, however, failed to find significant differences in technical practices of the better performing software operations as compared to not-so-good performing ones thus calling into question the assertion whether quality of software development processes makes a difference in the successfulness of a software operation. Indian software firms have been renowned in their ability to continually improve their processes through wholesale adoption and implementation of western process management approaches, like Six Sigma type approaches of the General Electric and Texas Instruments fame. No wonder then that major portion—as many as 50-75%--of the global CMM-SEI level-5 certified software operations are based in India. Indeed, a recent study by Dr. Michael Cusumano of MIT—a keen observer and researcher of software development best practices around the world—highlighted the fact that Indian software development operations have been especially successful in melding together older software development process approaches with the relatively newer ones in a manner that have made distinct and competitive.

While our own lack of finding is certainly a shortcoming of the study it cannot be perceived as an indictment of the industry. Certainly “lessons from the trenches” belie the simplistic conclusion of the statistical on that particular point. More research is certainly needed to understand the role of process quality in the development and evolution of Pakistani software industry. In this second of the two pieces drawing upon the results of the study, we highlight some of the lessons learnt from over 65 interviews of the leaders—CEOs/CTOs/Directors—of at-least 50 of the industry’s major players.

Before we do so, however, we must lay out an organizing framework for this discussion. We attempted to organize software companies within a smaller number of distinct groups—based on their salient characteristics. The purpose for this classification was to allow us to draw generalizations across the experiences of a large number companies into a set of prescriptions for the industry as a whole. Drawing from the qualitative insights of interviews and statistical findings, we devised a four-part taxonomy of generic software business models. The four sub-classifications were named after their most prominent examples, including: Export-focused Local Firm (“Systems” or “Netsol” Model), Domestic Focused Local Firm (“TPS” or “LMKR” Model), Export Focused Foreign (Expatriate) Firm (“Techlogix” or “Etilize” Model), and Dedicated Development Center (“ITIM” or “Clickmarks” Model). These generic models are explained, through the help of a graphic, in Figure-I. It is also important to mention here that none of these categories are as black-and-white as they seem. There is considerable gray area in between and a company might show characteristics of more than one type of generic software business model at a given time.

We present a snapshot of each of these generic software business models and identify key strategic challenges for each—13 in all for the entire industry. While we discuss each of these strategic challenges in the context of the generic software business model to which it is most relevant—they may be common across several (or all) of the business models. Readers seeking to thoroughly understand the strategic challenges faced by Pakistani software industry are hence advised to read through all four descriptions and their respective strategic challenges (thirteen in all) to get a better feel of the landscape as each contains some elements overlapping with others.

Before we begin this discussion on the taxonomy of generic software business models, it is important to highlight its several qualities and characteristics—as well as its proper and improper uses—to allow the readers to put this taxonomy in its proper perspective.

Firstly, the taxonomy gives us a relatively easy and comprehensive way to classify a particular software operation into a broad enough category of organizations, giving us a reference point to compare ourselves against, and quickly begin looking for certain organizational features, managerial characteristics, strategic challenges, and critical success factors that might apply to our own experience. This act alone somewhat simplifies the complexity of the Pakistani software scene narrowing down the search for comparables to look at or seek advice from.

Secondly, it is important to understand the fact that none of these generic software business models are inherently good or bad—just that each has its own place in the overall scheme of things. It is somewhat meaningless to compare firms across business models. It takes a fairly different set of initial conditions and skills to start, and overcoming a different set of challenges to successfully execute upon each. Each of these models, however, have better-performing and not-so-good performing firms within them and what one can do, again to a degree only, is to compare the performance of a firm against another within the same category.

Thirdly, while transitions between the generic business models are possible—they are not automatic. Depending on what a firm intends to do (its unique idea-offering-destination mix), there is a right model to look at and adopt. Although it may seem natural for a specific firm to transition from a domestic-focused firm to an export-focused firm, it is not automatically logical for every firm and certainly not for the entire industry to make that transition. One can remain within a particular model and aspire to be best in class within that model. An entrepreneur-investor must, therefore, clearly understand model implications, scalability challenges, and transition possibilities, before starting a firm.

Finally, it is important for aspiring entrepreneurs and business leaders and managers of existing ventures to understand the strengths, weaknesses, pre-requisites, and structural limitations of each of the generic software business models. They must vet their ideas through the lens of these business models and ensure that they fully understand their various dimensions and then adopt one that bests fits the idea-offering-destination profile of their venture and their short and long-term aspirations from the same. Understanding the model limitations is critical to the long-term growth of firms and the industry as a whole. Depending upon the circumstances and the goals and aspirations of the founders, many firms—trapped in the structural limitations of a particular model—try to outgrow it by doing more of the same.

Following is a brief discussion on each of these four sub-categories of models.

1—The Export-focused Local Firm (“Systems” or “Netsol” Model) is one founded by a predominantly Pakistan-based entrepreneurial team (that may or may not have been aided/encouraged by a group of expatriates), but with an explicit purpose of exporting software products or services. Within our own sample, it accounts of about a third (32%) or 16 of the 47 firms in our sample. Majority of the firms established in pre-DotCom Bubble burst era with an expressed purpose of exporting services to North America and Western European countries fall in this category. Although there are some that have taken the products route, their numbers are relatively smaller than those focusing on export of services. Many such firms have been created post-DotCom Bubble burst as well. Some salient examples of this type of business model in action are: ThreesixtyDegreez, Post Amazers, Advanced Communications, Makabu, Netsol, and Autosoft Dynamics etc.

The most defining feature of this class of companies, namely, the local-presence of their founders and the export-orientation of their products/ services, brings a number of unique and important challenges to this type of a firm. The most important of these strategic challenges, or critical success factors, is the ability to establish a first contact and a long-lasting linkage with the foreign customer. Gaining a foothold in the foreign market, signing up a first customer, and then quickly diversifying ones’ customer base has been a major challenge as well as a differentiator between successful and not-so-successful companies. A somewhat related challenge has been the setting up of a front-office abroad, generally for the purpose of marketing but also customer support etc. The dominant strategy here has been to set up a marketing office—manned by Pakistani or foreign staff. Not only is the cost of setting up a marketing operation in a foreign market prohibitively high (in $0.5M a year range for the smallest of the operations), it also hasn’t always paid off as per the expectations. An industry good practice is to consider other, more cost-effective, approaches like marketing alliances and reseller agreements as possible alternatives to front-office-type arrangements. These are especially suitable for smaller startup type operations. The aspiring entrepreneurs must look for such partnering opportunities and use them to their advantage.

A third strategic challenge (or critical success factor is the ability to understand a foreign domain and context—a factor made complex in this context by the geographical and cultural difference between the customer and the seller. For majority of these challenges, a regular expatriate connection (through a foreign/expatriate co-founder) and/or face-to-face visits to foreign clients and their markets certainly help. But more importantly, even the realization that one faces a particular challenge (e.g. understanding a foreign domain or diversifying the customer base) that needs to addressed would go a long way in getting one to the path of seeking the right kind of help/partnerships/arrangements. Having definite, well-thought-out, strategies in place could determine the difference between success and failure of an export-focused local software operation.

2—The Domestic-focused Local Firm (“TPS” and “LMKR” Model), with an exception of a few companies, is really one because of circumstances rather than choice. 14 of 47 (or 30%) of our respondents fall into this category. More often than not, and logically so, the domestic-focused local firm plans to export its products or services abroad and is merely using the domestic market as a vehicle to gain a track record with real life customers. Whether a firm is in this category by choice (“I’ll do domestic first, export later”) or by circumstances (“Since the export market doesn’t seem very good right now, I’ll survive by selling at home”) the strategic challenges are quite similar. Among the companies being established in this category, we see a much greater domain focus than the earlier category (discussed above). The primary reasons for this is that, unlike the exports segment, there is little or no lure of pure software programming (or coding) “services” on the domestic front. Thus everyone must be in the business to solve a fairly well-defined problem or not be in the business at all. This has led to a much faster accumulation of domain knowledge, maturity and perhaps, to some degree, better and more focused business strategies to start with in the first place. Some salient examples of this type of business model in action are: 2B Technologies, ZRG, TPS, Lumensoft, Yevolve, SI3, Softech Systems, AppXS, and Genesis Solutions

That this type of firm enjoys some advantages vis-à-vis the export-focused firm does not, however, mean that they don’t face serious strategic challenges. The most serious challenge this firm faces has to do with operating in an under-developed domestic market. There are many manifestations of this problem, namely, a psychological price barrier, lack of buyer-user awareness, lack of business volume, dealing with the seth mentality, and little appreciation of the value of software etc. We allude to a number of critical success factors and industry good practices to counter this strategic challenge, namely, avoiding the herd mentality, focusing on better developed segments of the market (e.g. financials, telecom etc.), correctly estimating the difficulties of “creating” a demand, making a business case rather than selling a technology, using innovative pricing strategies, and operating on the innovative end of the product-services spectrum etc.

Another major challenge for this category of firms is getting access to venture, expansion, and working capital. For a variety of reasons having to do with lack of demand, perception of success, and inability of financial and software communities to understand each other’s perspectives, domestic-focused local firm suffers from serious shortage of capital. Consequently, a realistic estimate of capital required to execute a business idea and a realistic fundraising strategy, through external or internal sources, is a critical ingredient of successful operations. A third major challenge for a domestic-focused firm is having a clear domain focus and a solid businessplan—and adaptable but solid businessplan—to start with. Box # 1 clearly identifies some industry good practices used by relatively more successful companies to address these issues.

3--The Export-focused Foreign Firm (“Techlogix” or “Etilize” Model) is one founded abroad (or jointly, in Pakistan), by a predominantly foreign (usually, an expatriate) entrepreneurial team, with an explicit purpose of using the Pakistan-based offshore development facility to deliver a product or service demanded by the foreign market. 7 of the 47 (15%) companies in our sample fall in this category. This type of business model has been adopted by services and product-focused companies alike. Some salient examples of this type of business model in action are: Elixir Technologies, Etilize Inc., Ultimus, MixIT, TechLogix, Prosol, and Xavor etc.

While companies in this class of software business models enjoy several advantages over those in earlier discussed categories, namely, quality of due-diligence on the basic idea, foreign contacts/networks of founders, and better access to capital etc., there are significant challenges as well. The most important challenge facing this class of companies is to deal with the “image” problem. Simply speaking, foreign entities—for reasons having to do with negative perception of the country or geopolitical environment of the world/region—are reluctant to do business with a Pakistani entity. While there are a number of things that can and must be done at the public policy level to address or atleast counter some of these concerns, several good practices of companies trying to address these issues on their own also emerge from the study (see the study for detailed findings).

Another significant challenge is the geographical shifting of the “labor arbitrage” argument whereby as the outsourcing movement matures, newer (less problematic) regions with even cheaper labor come into the competition with the more established ones. The example of Dubai (for it accessibility and investment and HR-attractive policies) and China (for its cheap and abundant labor and large domestic market) are cases in point. Consequently, the Export-focused foreign firm must adopt active strategies to continually counter the threat of becoming non-competitive in the market of cheap labor. Several good practices emerge from our sample of respondent, including but not limited to, developing domain expertise to lock-in customers, relocating part of the production activity to cheaper havens (e.g. Beijing, Dubai, the Philippines), climbing up the value chain, and automating their own systems to remain cost-competitive.

A third key challenge that the export focused foreign firm shares with companies in other generic business models is that of scalability of Pakistan operations. There is a widely perceived and talked about psychological barrier of 200-people in Pakistani software industry. Although the exact number is immaterial, what matters here the fact that the 20,000+ employees of the top companies of our closest competitor—India—dwarf the size of our entire industry. The largest Pakistani company in our sample is not even 2.5% the size of the largest Indian company—putting a big question mark on our ability to deliver serious large-scale projects. Companies face formidable supply (HR availability, lack of middle/project managers) and demand (inability to predict demand and hence scale up) side challenges in this regard.

4—The Dedicated Offshore Development Center (“ITIM” or “Clickmarks” Model), as the name suggests, is a fairly limited offshore operation of a foreign company. 7 of the 47 (15%) of the companies in our sample fall within this category. The dedicated development center is different from the Export-Focused Foreign (Expatriate) Firm in the sense that it is often an “add-on” to an already existing company whose strategic and managerial processes and controls are quite well-established. Due to its unique nature (i.e. limited scope and say within the parent’s strategic processes) it faces a number of challenges that are distinct and different from the earlier-discussed category. The nature and extent of the challenges also differ between the various phases of the entities life-cycle. Some salient examples of this type of business model in action are: MetaApps, ITIM Associates, Clickmarks, Trivor Systems, and Strategic Systems International etc.

The initial set of challenges encountered by entrepreneurs trying to set up a dedicated development center generally have to do with the actual setting up of a Pakistan operation and doing it in a manner that minimizes the disruption to the foreign operations of the parent company, putting in place a quality software development team that understands the foreign domain, quickly learns the parent’s corporate culture, and is both capable and motivated to deliver a quality product or service to an unseen entity. These challenges, however, are not insurmountable and a number of companies are using fairly good and innovative practices to effectively respond to the above-cited challenges.

The most important of the strategic challenges facing a dedicated development center type operation, however, is that of effectively managing the parent-subsidiary relationship. This is essentially a challenge that, unlike those discussed above, presents itself in the later (more mature) stages of the subsidiary’s life-cycle. It can, however, be addressed somewhat painlessly by planning for it at the very inception of the parent-subsidiary relationship. What actually happens is that, owing to the rather limited scope of the development center operation, there can come a time in its life when its interests may diverge from that of the parent company resulting in organizational confusion, serious management tension, and a loss of employee morale.

We find a number of companies in our sample suffering from such a “midlife” crisis and struggling to remain “purposeful” in the overall scheme of things. Despite being the most technically sound operations, they also faced serious bottlenecks vis-à-vis revenue and employee growth. There are many industry good practices that one can adopt to effectively counter this serious challenge, namely, diligently working out a “written” parent-subsidiary agreement to avoid later confusion and heartburn, providing the subsidiary with a voice in the parent’s strategic processes, allowing the subsidiary the freedom to seek other business, having separate top-management for the parent and subsidiary etc. In addition to these model-specific challenges, this model shares several other challenges and critical success factors with the rest of the industry but more specifically with the export-focused foreign firm model.

By using the lens of this four-part taxonomy of generic software business models, we can greatly reduce the complexity of the relevant organizational landscape. Pakistan’s software industry comprises a handful of different types of firms that face a remarkably common set of strategic challenges. While some of these challenges may be unique to the particular idea-market mix that a particular firm is targeting others are common across firms within a generic business model and the industry as a whole. As Pakistan’s software industry attempts to hasten its move up the “learning curve”, its executives must pay attention to the shared experience of the successful and not-so-successful entrepreneurs among them. While many of these problems are substantial and hence may requires years of hardwork, planning, and execution, it is refreshing to know that a number of companies have found ways to get around these in one way or the other and are willing to share these insights with others. The study highlights the thirteen strategic challenges and twenty-five industry best practices that industry leaders would find worth their while to think about and implement.

Athar Osama, the author of the PSEB Study, is a Science and Technology Policy Analyst based in Santa Monica, California. He maybe contacted for comments and queries at Athar.Osama@gmail.com

Software Development in Pakistan: What Do Statistics Tell Us?

Software Development in Pakistan: What Do Statistics Tell Us?
By: Athar Osama

The software industry—widely seen as the “great enabler”—provides an opportunity to the developing countries to play a greater economic role in the fast globalizing world. The example of neighboring India—whose ambition and progress towards becoming a “mini (software) superpower” is no mystery from the world—is often cited in the development literature as an evidence of the fact. Pakistan’s software industry—widely perceived to be sharing a number of key factors with India (e.g. a supply of English-speaking, technically trained, and cheap manpower, a favorable public policy and infrastructure environment, and a government willing to help private interests etc.)—has embarked upon an ambitious effort of its own to claim its share in the riches of the world’s software markets.

Pakistan is currently viewed as a tier-3 country (defined as having around $25 Million in export earnings, tens of software companies, and upto 5 years of industry maturity) in a widely quoted taxonomy of software exporting nations. It is widely believed, in both government policy and entrepreneurial circles, that with the wealth of talent and strengths available, the country deserves a better place in this global pecking order of software exporting nations—atleast a tier-2 status (defined at $200 Million or more in software revenues) like Russia and China, or even a tier-1 status (defined at more than $ 1 Billion in export earnings) alongside archrival India. Whether or not the Pakistani software industry is able to capture its “due” share and make a mark on the major software markets of the world remains to be seen.

While Pakistan’s software industry has been a subject of the curiosity of interested by-standers—both local and expatriate entrepreneurs—industry analysts, and potential investors alike, lack of credible data on the current state and competitive dynamics of the industry has often been a hindrance in engaging these individuals and materializing many prospective ventures. In the absence of relevant data to back these assertions or visible success stories of appropriate stature, these strategic conversations revolve around many tough questions about the current state and future prospects of the industry. For example:

Why hasn’t the Pakistani software industry been able to produce a single world-class software firm (e.g. Wipro, Infosys or TCS of India) in the last 10-15 years?

Why haven’t we been able to grow Pakistani software exports beyond a certain level ($30-60 million per annum) for the last 5 years?

Does Pakistani software industry merely represent a lower level of development or an altogether different development trajectory as compared to known peer nations?

What constitutes a generalized set of best practices in the local software industry (i.e. what differentiates better performers from those that don’t perform that well)?

Answering these questions require insights and understanding of the local software scene. In October of 2004, Pakistan Software Export Board (PSEB)--an entity charged to promote the local software industry--funded a 3-month long preliminary research study aimed at developing these insights and understanding of the local software scene. While several factors are widely believed to be a hindrance in the country’s aspiration to become a significant software exporter, not the least important of which are macro- and geopolitical in nature (e.g. law and order and security situation, image of the country etc.), the study adopted an inside-out approach that asked: “What can the various players, essentially software companies, in the industry learn from each other?”

The benefits of the above study, we believed, would be two fold. The primary motivation for undertaking this study is that of within industry learning. To that effect, this study aimed at developing a comprehensive snapshot of software development activity in Pakistan and catalyze a learning process for the software entrepreneurs, executives, financiers, managers, and professionals. The secondary motivation for undertaking this study is to promote and facilitate investment in the local industry. In that context, the study results would be of value to investors and financiers, local and foreign, as well as those on the sidelines, contemplating starting a software venture and looking for a good sense of what they can learn from the collective experiences of tens of successful and not-so-successful entrepreneurs.

The study was concluded in December of 2004 and briefed to industry leaders through a series of consultative meetings in January 2005. It was formally published and launched by Pakistan Software Export Board (PSEB) on April 26, 2005. The detailed report and related material may be found at the study website (http://paksoftwarestudy.vttp.org) and the study blog (http://Pakistan-Inc.blogspot.com). This is a two-part article series that draws upon the results of the PSEB study to create a picture of the Pakistani Software Industry. In this first of the series of articles, we would look at the statistical snapshot of software development activity in Pakistan. In the second article of the two, to be published next week, we would attempt to develop a qualitative snapshot of the Pakistani Software Industry.

The study draws upon an “on-the-spot” survey of around 40 of the most prominent and largest software companies in Pakistan (60 in all), as identified by PSEB and P@SHA. To ensure homogeniety of results, the sample focused on "pure" software development activity and purposefully excluded BPO and IT-enabled services (ITES). We conducted organizational interviews with senior executives (CEOs/CTOs or Local of Heads of Operations) of these companies to supplement the statistical data with qualitative insights. These interviews focused on understanding these organizations, their business and revenue models, competitive drivers, strategic challenges, and policy bottlenecks. We also conducted interviews of opinion leaders, policy-makers, and senior executives of other organizational entities (e.g. IT MNCs, financial institutions, and academia) that had a significant bearing on the local software industry. In all we conducted over 65 interviews between Oct.-Dec. timeframe. The results of the statistical analysis are quite illuminating.

Pakistan’s software industry is still going through early-stage growth with only few large players, but it is growing at a fairly decent rate. On the whole, the 60 software houses included in our statistical sample employ over 4000 technical and professional employees—for an average of 62 employees per organization. Roughly one third (32%) of the software companies reported annual revenues of more than a million dollars with some reporting more than $5M, another third (36%) between $200K and $1M, and the rest (32%) less than $200K. 6 of the companies had more than 250 employees and another 8 had between 100 and 250 employees. On the whole these 60 companies had experienced an employment growth of about 27.5% and a revenue growth of 37.4% over the last year—pointing at better utilization of excess capacity or value-addition per employee, or both.

A Large number of companies are formed as subsidiaries of foreign companies and many local operations seek to develop front-offices abroad. Around 40% of the companies in our sample are subsidiaries of foreign companies—with majority of them having a parent company in the United States. 55% of the companies had one or more front offices abroad (50% in the US, 11% each in UK and Middle East, and 3% in the Asia Pacific region). There are, however, differences in propensities to seek such arrangements by type of offering (product-service) and target market (exports-domestic) of software companies.

The industry’s market-offering mix is skewed towards export-services and domestic-products, primarily to the private sector. Public-sector or government sales represent small fraction of the total market. Broadly speaking, the sixty companies in our sample derive their revenues from export and domestic markets in a ratio of 60:40. On the exports side, they earn 37% of their revenue from products and 63% from services (22.5% and 38.5% respectively, of the overall revenue). On the domestic side, however, the ratios are somewhat reversed with products and services contributing 58 and 42% respectively (23% and 16.5% of the total). These ratios are further skewed if we looked at specializations of firms. For example, clearly-domestic-focused firms would derive, on average, 68% of their revenues from domestic operations and clearly-export-focused firms may derive, on average, 85-98% of their revenues from export of software. Again, the pattern is skewed towards services for export-revenues and products for domestic-sales. Our conversations with the top leaders of the industry suggest that majority of the product revenues are from customized rather than "shrink-wrapped" products. As much as 85% of the software sales are to private sector and only 15% to the public-sector or government.

There are few clear-cut differentiating patterns in managerial practices of exports- and domestic-focused software operations. There is some suggestive evidence, however, that export-focused software operations are more likely to distribute stocks/ownership among employees, hold employee bonding activities, and benefit from employee-driven innovation while domestic-focused software operations are more likely to share profits with employees, provide additional benefits to female employees, have greater financial discipline, and provide time to employees to work on their own interests. Despite the latter, however, they seem to benefit less from employee-driven innovation and suffer more from a perception of lower delegation quality. Hybrids fall in between the two categories on almost all these measures.

Export-focused operations tend to spend more, on average, on quality assurance while hybrids tend to have a greater propensity for seeking a quality certification. Exports-, hybrid, or domestic-focused software operations are equally likely to have a dedicated quality assurance team. The former, however, spend a much higher percentage of their expenditure on quality assurance function (17% of the employee payroll as against 12% for the other two categories) but are much less likely to seek a quality certification. Only 50% of export-focused software operations have an ISO/CMM certification while 72% of the hybrids have it. The corresponding figure for domestic-focused operations is 36%. The trend probably points towards the privileged business relationships of export-focused operations that do not require a quality certification. Data on cost-structures of export, hybrid, and domestic-focused firms also confirms these findings. Hybrids, on the other hand, tend to operate in a much more competitive environment and hence feel a greater need to signal quality through a certification. Hybrids also display a greater propensity to use standard software engineering methodologies (e.g. waterfall, iterative, prototyping) as against export-focused operations that tend to use more esoteric (“homegrown”) and non-standard approaches that fit best with their clients’ processes.

Companies, across the board, focus on high-contact strategies to seek customers. That “selling software is a highly contact intensive sport” is evident from data on use and perception of successfulness of marketing approaches. All types of organizations identify high-contact methods like 1-to-1 contacts, network and relationships, and word-of-mouth referrals as the most successful (all rated > 3.5 on a scale of 5, on average) of the marketing approaches and low-contact ones like advertising and going to conferences and exhibitions as least successful (rated <2.5 on a scale of 5, on average) of the approaches. The use of alliances and agreements with channel partners seem to fall in between these two extremes—with the important caveat that these don’t seem to work as well for domestic-focused operations as they do for hybrids and export-focused ones. Consequently, in line with the perceptions of successfulness, companies seem to have focused their energies on approaches that appear to work best.

Additionally, the data on cost-structures (i.e. % of total expenditure spent on various heads) of software operations seem to suggest that export-focused companies engage more in “relationship-selling” rather than direct marketing & advertising, while hybrids under-invest in product-development, perhaps, to pay for costlier marketing/advertising and training and certifications. Also export-CEOs operate in a relatively tactical profile—focusing more on day-to-day management and less on product and strategic planning & marketing/advertising.

Classifying the data in other ways (e.g. development center-type operations vs. the rest, products vs. services focus, small vs. large, pre-DotCom vs. post-DotCom etc.) suggest few interesting insights. The dedicated development centers tend to be smaller, more rigorous, from a technical and process standpoint, than the rest of the industry. They, however, seem to experience serious constraints to revenue and employment growth—a fact that can be explained as manifestation of their “mid-life” crisis and/or the recession in the markets of the respective parents. Although there is a trend towards productization in the industry, there are few significant differences between product-focused and services-focused operations. This lack of differentiation (e.g. in the cost structures of services and product-focused operations) is problematic, to say the least. There were also few significant differences between software operations created before and after the DotCom Bubble burst, over and above those that can attributed to relatively younger profile of the latter.

Do aggregate statistics reveal a pattern of “best practices” within the Pakistani software Industry? We use multiple comparison groups (e.g. 40 most prominent companies, top-10 companies, 14 fastest growing companies, 14 companies that describe themselves as falling within the top-quartile, globally) and found mixed results on that account. For example, we find robust evidence to support the fact that better-performing companies tend to adopt a set of employee-friendly management practices (e.g. flexibility, stock ownership, profit-sharing etc.) and have access to high quality managerial talent (e.g. mix of technical and business backgrounds, prior venture experience, financial discipline etc.) than the rest of the industry. All companies, across the board, prefer high-contact marketing approaches over low-contact ones but better-performing companies report higher satisfaction with the former than the rest of the industry. Our results on various measures of technical and process quality were, however, inconclusive, at best. Here, we did not find any clear patterns that differentiate better-performing companies from the rest of the industry. We believe that best practices within technical and process realms are dependent on the type of work performed and a number of project-specific variables. As reported elsewhere, therefore, project-level data might be better suited to identify these differences.

On the whole these findings paint a picture of lack of real focus and specialization within the Pakistani software industry. That product-focused operations are similar to services-focused operations and pre-DotCom operations are not qualitatively very different from post-DotCom operations does not speak well for the maturity of the industry as a whole. The second finding is especially disturbing in the sense that the DotCom Bubble burst in the United States is widely seen as a watershed event in the relatively short history of the country’s software industry and is widely perceived to have brought clarity of thought and business focus to the industry’s entrepreneurs. An alternate, and perhaps the right, way to look at this seemingly discouraging finding is that the more dominant of the effects of the DotCom Bubble burst has been the influencing of the business models of the already established firms. For example, there is a clear trend among export-focused software operations towards diversifying their bets by establishing a significant local presence too.

One can also observe a trend towards the “hybridization” of software development activity in the country and the emergence of some managerial best practices. The hybrid firm has emerged as an important organizational class on its own rather than the average of the two extremes. While the hybrid firm tends to do better than the two extremes on some measures and hence might be seen as a manifestation of the industry’s survival instinct in tough economic times, it is not quite clear if it is the optimal model of organization of software development activity in the long run. Another important organizational observation pertains to the average size of the firm. Scalability has often been cited as a major managerial issue confronting Pakistan’s software industry. It is often believed that the industry, as a whole, suffers from a 200-people barrier. We found that to be true, figuratively if not literally. One positive finding is the adoption of employee-friendly policies and profit-sharing among the relatively more successful companies. The study results seem to suggest that, contrary to the general perception, such employee friendly policies seem to pay off the form of better performance in the long-run. Their importance is also paramount in software industry in particular because of the highly creatively and eccentric workforce that it tends to employ.

Finally, while the software industry has managed to grow at a decent 37% over the last year, the results vary considerably across sub-sectors. To a large extent, domain and domain expertise has emerged as a key determinant of firm success. In the domestic market, for instance, software firms developing products for financial and more recently the telecommunications sub-sector have done much better while those dealing with ERP and industrial automation systems have done much worse than the average. This essentially drives home the fact that the fate of the software industry, in general, and the domestic software industry, in particular, remain largely linked with the growth in relevant sectors of the economy. That software industry in and of itself cannot generate growth in stagnant industrial and economic environment is an important insight for policy makers as well as aspiring entrepreneurs. The situation is only slightly different for the export markets where customers are increasingly demanding a prior track record, domain expertise, and experience in handling large projects as a pre-condition for lucrative foreign contracts thus pushing the industry into a chicken and egg problem. Software companies founded by expatriates, although better prepared to meet the above challenge, have failed to grow beyond a certain size due to multiple reasons, not the least important of which is the depressed demand for software in the US market.

In short, while the industry has had its fair share of challenges and problems, it seems to be on a fast learning curve, has done much better than before, and is expecting even better performance next year. Next week, we would look at the qualitative findings of the study attempt to identify the key strategic challenges faced by the industry.

Athar Osama, the author of the PSEB Study, is a Science and Technology Policy Analyst based in Santa Monica, California. He maybe contacted for comments and queries at Athar.Osama@gmail.com