UNIVERSITY OF PENNSYLVANIA - AFRICAN STUDIES CENTER
Building Software Industries in Africa

Building Software Industries in Africa

[Richard Heeks]

BUILDING SOFTWARE INDUSTRIES IN AFRICA

Richard Heeks

Software production capabilities are building up within every African nation, and one can see formative software industries in many places1. This article investigates past and future directions to this capability build-up, and the lessons that can be learned from other software producers.

Before proceeding, one fundamental question needs to be addressed: does it make sense for African nations to set up their own software industries?

The answer is an unequivocal "Yes". Information technology (IT) overall is one of the most critical technologies affecting economic growth in Africa and, within the overall set of technologies that make up IT, software is vital since other technologies cannot function without it. Software has also been forming an increasing component of overall value within information technologies and "is increasingly becoming a pervasive technology embodied in a vast and highly diversified range of products and services" (Gaio 1989).

"Computer software has become the 'lifeblood' of business, industry, and government." (World Bank 1993)

The development of a local software industry can therefore lead to many positive externalities, and is a necessity if African countries are to adapt software technology to suit particular local needs.

"Software production is nowadays an industry, essential for the growth of the economies of the developing countries; and the launching of programmes to promote strong and indigenous software industries is a priority task." (Fialkowski 1990).

Software production is also the best entry point for Africa into the IT production complex. For example, compared to hardware production, software production has much lower entry barriers because it is less capital-intensive, more labour- intensive, with a lower rate of obsolescence, and (at least for certain types of software) it has far fewer economies of scale. All of these factors work in Africa's favour, and software's labour-intensity of production combined with low African labour costs offer a clear opportunity.

1. STRATEGIES FOR FUTURE GROWTH

Experience shows that there are four strategic positions that can be taken within the software industry, as shown in Figure 1.

			          Software
			          Business
		       	    Services		     Packages
		           ------------------------------------------------
		           |		|		|
		           |		|		|
   	  Export           |	C	|	D	|
		           |		|		|
	Market             |		|		|
	Served	     ------------------------------------------------
		           |		|		|
		           |		|		|
   	 Domestic          |	A	|	B	|
		           |		|		|
		           |		|		|
		           ------------------------------------------------

Figure 1: Strategic Positioning for Software Firms

For many African countries and software firms, the export- oriented strategy seems highly attractive. They look to the example of India as proof that software exports can be hugely successful for developing countries. The figures seem to support this. Software exports from India in 1994/95 were a staggering US$480m - up from less than US$4m in 1980 and growing at 40% or more per year. 18 Indian firms earn more than US$5m annually from software exports, with the largest (Tata Consultancy Services) earning US$90m - more than all merchandise exports from entire African countries such as The Gambia and Burundi.

However, software exports are not a panacea for African IT development nor, therefore, are strategies C or D to be recommended at present for two key reasons, detailed below.

2. THE ILLUSION OF INDIAN EXPORT SUCCESS

Firstly, India's export performance is not all that it might seem. The figures quoted for India give gross earnings. Most of this money leaves the country to pay for travel costs and living allowances of those software developers who work overseas, marketing costs, profits of multinational subsidiaries, and the import costs of related hardware and software. In reality, then, 75-80% of the foreign exchange earned does not end up in India.

Much of the work done is in no way glamorous, but involves the most basic, dull conversion and maintenance tasks that Western programmers are unwilling to undertake. Partly because of the consequent lack of job satisfaction for Indian software workers - combined with their frequent siting in the US, Europe or Australia - the export trade has encouraged a brain drain of talent amounting to around 15% of total industry workforce each and every year.

Finally, the export drive has incurred a large opportunity cost as the finest local minds now produce software to boost the performance of American and European companies rather than using their talents to address domestic developmental needs. Dependence on foreign multinationals is heavy, and this has been associated with a snuffing-out of local innovation.

3. BARRIERS TO AFRICAN SOFTWARE EXPORTS

Even if one still wishes to believe that, "warts and all", India's export experience provides a model for Africa, there is a second, stronger argument against the export-oriented approach: that Africa currently faces far too many barriers to enter this market.

The barriers include:

* Lack of skills. Though software exports are often "low- skilled", they still require at least a graduate with in-depth technical experience. Such workers are in seriously short supply in all African countries.

* Lack of infrastructure. The software export trade increasingly demands a sizeable installed computer base; reliable and pervasive telecommunications links both domestically and internationally; and reliable electricity supply. With hand on heart, few African countries can yet claim to provide this.

* Lack of market information. Exports are based on understanding your export market and having trading contacts in that market. Such information is not readily available within Africa.

In addition, African nations face the difficulties of any late- comer to a market. Countries like India, Singapore and the Philippines arrived on the export scene many years ago. They have already developed the requisite skills, contacts, policies and infrastructure that are so lacking in Africa. As a result, these established players will continually consolidate their position whilst squeezing out potential African newcomers.

There are, of course, a few producers in countries such as Egypt and South Africa who export software. Unfortunately, they have tended to do more harm than good in terms of image because they are the exception, not the rule, and the focus of African software industry attention must be placed elsewhere.

4. "PLAN B" - THE SOFTWARE PACKAGE MARKET

If export-orientation is currently unviable, should African firms, instead, be aiming for the production of domestic software packages? In the general applications market (e.g. word processing software, spreadsheets, databases), they certainly should not.

The Microsofts of this world have that market wrapped up, and entry barriers for African firms remain formidable:

*High development costs. Any low labour cost advantage in development is quickly eroded because of the huge advertising and marketing budget required for a successful package2. Aside from these costs, experience suggests that only about 1-5% of products succeed, thus providing very little return on most package investments.

*Foreign software preference. Irrespective of price, quality and features, African consumers appear to prefer foreign rather than local software.

*Piracy. Rampant package piracy squeezes an already small domestic market to minuscule proportions, thereby rendering unit costs for marketing and distribution even higher.

*Dearth of market information. Information on the domestic market is lacking, making it very hard to plan, design and market a new package

Narrow, vertical package markets do exist in Africa, in public administration, health administration, hotel management, insurance, accounting, etc. African software firms are addressing these markets but one finds that their "packages" are often just a set of menu or window interfaces that are used as a marketing or development platform for further customisation. In addition, the steady globalisation of all markets means that competition from multinational imports (both legal and pirated) increasingly threatens even these local developments.

5. MASTERING STRATEGIC POSITION A

Firms therefore need to aim for strategy A by providing software services for their local markets. This is the easiest market segment to enter and it is here, inevitably, that most African software firms sit. They offer installation, training and customisation based on imported packages (typically dBase, Access, Lotus and Excel) or, more rarely, custom-build software from scratch according to customer requirements.

In the longer-term, African software firms can seek to follow strategic development paths such as A->C or even A->B->C. However, this can only come about once strategic position A has been successfully mastered.

A cold, hard analysis of firms' performance in this strategic sector shows that all is not well and that position A has yet to be mastered. There is still heavy reliance on foreign software developers for most large contracts. Systems integrators have shown themselves more adept at selling and installing equipment than at meeting user needs. Software development staff are often self-taught from the manual and, while they may have some technical skills, lack an understanding of information systems analysis and design techniques, especially an understanding of human and organisational requirements.

The US Software Engineering Institute offers a five-point scale against which firms' software development processes may be classified:

1. Initial: ad hoc processes

2. Repeatable: basic management practices are defined and followed

3. Defined: technical practices are defined and enforced

4. Managed: fully defined process is measured so that performance can be controlled

5. Optimising: measurement results and error prevention activities are fed back to identify areas for improvement

The vast majority of African firms fall into category 1, with only the smallest handful finding their way towards categories 2 and 3. It is little wonder that customers are turning to imported packages, with their guarantees of quality, reliability and longevity, and trying to keep local customisation to a minimum, even if this means that they do not have their needs fully met.

The way forward is clear enough and involves actions such as more staff training, a more professional approach from software company management, and greater customer orientation. Yet such things are far easier to describe than to implement, partly because firms face external constraints which they alone cannot overcome. Higher authorities must therefore become involved.

6. THE NEED FOR GOVERNMENT ACTION

The obvious higher authority to involve is government, but not all commentators support this. Many US-based companies and development organisations, for example, claim that the best development path is that provided by reliance on market forces and "rolling back" of government intervention.

These advocates of the free market approach are suffering selective amnesia. America built its IT industry on government money pumped in during critical early growth years in the 1940s, '50s and '60s. Those preaching market forces today do so only because their industry is now fully-established and because the market-only approach means more sales and less competition for US software products.

The free market will never provide a software industry for Africa and government action is essential. If this is the case, what actions should African governments take?

Before doing anything, they must survey the current status, trajectories and needs of the local software industry. Having done so they are likely to find a menu of constraints which need to be addressed, some of which have already been mentioned. They include:

*Lack of capital, especially venture capital.

*A small, heterogeneous domestic market with high piracy, poor IT awareness, a consumer preference for imports, and a lack of recognition of the need for customisation to meet organisational needs.

*Lack of software development skills and software production management skills, leading to problems with local software development productivity and quality.

*Lack of information about local and foreign software markets, and a lack of marketing skills.

*Lack of relevant infrastructure.

*In some countries only, language barriers, given the predominance of English within the software world.

Some of these are hard to address and can only be attacked slowly and indirectly - for example, those on local consumer preferences. However, in all the other areas, government can - indeed, must - act. For example, government and industry must work together to plan appropriate education and training institutions, courses, curricula, and quality assurance, delivery and financing mechanisms. Here, the experience of other countries provides a valuable guide, from India's training certification scheme to Ireland's tax breaks for training3.

In conclusion, then, one can predict that African countries are unlikely to become software giants for many decades. However, they can become players within the global software market, but only if they learn the lessons of those who have gone before. They must eschew the bright lights of software exports and - ignoring the selfishly-motivated calls of the free marketeers - install a flexible but co-ordinated set of government promotional interventions.

Dr. Richard Heeks is a lecturer in technology and development at the Institute for Development Policy and Management, University of Manchester, UK. Fax: +44-161-273-8829. Email: richard.heeks@man.ac.uk

REFERENCES:

*Economist; (1994) The Third Age: A Survey of the Computer Industry. Economist, 17/9/94, Special supplement

*Fialkowski, K.; (1990) Software industry in the developing countries: the possibilities. Information Technology for Development, 5(2), 187-94

*Gaio, F.J.; (1989) The Development of Computer Software Technological Capabilities in Developing Countries - a case study of Brazil. PhD thesis, University of Sussex, Brighton, UK

*Heeks, R.B. (1996) Promoting software production and export in developing countries. In: Information Technology, Development and Policy, E. Roche & M. Blaine (eds). Avebury, Aldermaston, UK

*World Bank; (1993b) Turkey: Informatics and Economic Modernization. World Bank, Washington, D.C.

FOOTNOTES:

1 The work reported here derives from ongoing research into African and Asian IT industries, including continent-wide surveys plus specific fieldwork in countries such as Kenya, The Gambia, Zimbabwe and India.

2 Many major software multinationals spend 40-50% of annual revenue on sales and marketing, and Microsoft spent roughly US$1.5bn in 1994. Even research and development is becoming costly. The main firms spend 14% of revenue on this and, for example, Microsoft spent US$60m developing its Access database (Economist 1994).

3 Details of these and other successful government interventions are provided in Heeks (1996).

**********************************************************

From: Richard Heeks <mzdrbhs@mail1.mcc.ac.uk> Date: Fri, 16 Aug 1996 08:26:56 BST Subject: Building Software Industries in Africa - Paper ?? Message-ID: <C19F24650BA@mail1.mcc.ac.uk>



Editor: Ali B. Ali-Dinar
Previous Menu Home Page What's New Search Country Specific