Computing and Communications continue to be very expensive for SMEs in emerging markets. The result is low penetration of computer hardware, limited use and/or piracy of software and limited, low-speed connectivity to the Internet. These factors (concomitant with the inherent cost barrier of computers, software and communications) ensure that SMEs continue to be very people- and paper-reliant.
In fact, in India the installed base of PCs is less than 6 million (offices and homes), with the addition of 1.8 million new units annually. This is ridiculously small. In fact, in a shocking state of affairs, PC sales for 2001-2 have been more or less flat compared to the previous year.
Even as the world moves inexorably to ubiquitous connectivity, the lack of appropriate technology deployment within SMEs places them at a significant disadvantage as compared to the bigger companies and their counterparts in the developed markets. This should not be so. SMEs should be empowered with the latest technologies in a cost-effective manner, and as a collective they should be able to wield power, make themselves heard, and garner more business. We want to bring about an epidemic of change in the SME Technology infrastructure.
One point: here, when I am talking of SMEs, I intend to refer to not only the small and medium sized companies, but also those independent decision-making units in the BigCos (especially in the emerging markets) who are cost-sensitive. Many of the ideas described here apply equally well to them because there are few who can afford to spend in dollars per employee (since for most, their revenues are still local currency). This thinking helps grow the market dramatically, and in fact these BigCo SME units could be among the early adopters helping us achieve scale and volumes much faster.
A Connected Computer on every Desktop
My starting premise is that every SME should be able to afford a desktop PC for each employee, regardless of salary. The productivity impact of computers is beyond doubt. Any person in the coming earning more than USD 150 (Rs 7,500) per month should be given a connected computer for no more than 10% of the cost of his salary. Today, PC penetrations in SMEs are perhaps at 5-10% (or less) in countries like India. What we are suggesting is a 5-10x increase in the deployment of PCs, along with the software they need to communicate, collaborate and manage their business.
Our target is to provide a computing and communications platform for SMEs for the equivalent of a one-time, upfront payment of USD 100 (RS 5,000) per person, USD 500 (Rs 25,000) per location, and USD 5 (Rs 250) subscription fee per person per month. For this investment, SMEs should be able to give a desktop computer with connectivity to the Internet to each of their employees and all the software needed from messaging, security, collaboration to the enterprise applications.
Hardware in our case follows the 30-30 Rule: providing 30% of the functionality at 30% of the price. Thus, we dont need USD 300 computers on the desktop but USD 100 machines. Software follows the 10-10-10 Principle: 10 times the applications, 10 times the speed (since it is on the LAN), at one-10th the price (of a single application). Communications is the paradise of the commons, built bottom-up using open spectrum.
Thus, the internal IT architecture of SMEs will be as follows:
Used PCs as Desktops, serving as Thin Clients, running Linux and a Windowing System; without hard disks and CD drives; booting off the network. The used PC configuration can be as low as 486 machines with 32 MB RAM. By simplifying the desktop PC, maintenance becomes very easy: in the event of any problem, just replace it with another PC. In fact, the PCs too are interchangeable since all data is stored on the server.
Current Desktop as an Enterprise Server, at a cost of about USD 500. This will host all the software and content for the desktops, along with the mail, files and preferences. The desktops work as thin clients, with this thick server. All that is required is to use the latest desktop configuration available in the market as the server. Since all the access from the clients is to the server, there is no dependency on a real-time network connection, even though this is likely to be there.
Communications: Use 802.11 technology within the enterprise and for external connectivity. 802.11b works with open spectrum (2.4 Ghz) at 11 Mbps, which puts it in the same range as the 10 Mbps Ethernet we have been familiar with. By using 802.11 within the enterprise, it simplifies the whole networking exercise. No need to lay cables, which again requires maintenance. Put a wireless access point in neighbourhoods (managed by entrepreneurs, akin to how the cable TV revolution began in India in the early 1990s). This ensures that there is high-speed, shared bandwidth available across multiple enterprises. 802.11 technology can also be used in rural and semi-urban areas through the use of directional antennas to support access over greater distances. What is effectively an alternate (adjunct) technology in countries like the US can become a mainstream technology in countries like India.
Software: Software will be available on the Enterprise Server. By making all software available on the Enterprise Server on the LAN, we bypass the low-speed and real-time Net connectivity bottleneck. So, its an ASP model, but the ASP sits on the corporate LAN. The software created, especially the enterprise software, will need to adhere to the various business process standards that are being drafted ebXML and RosettaNet are two examples. My belief is that SMEs will be willing to change some of their business processes they have anyways evolved their processes internally. By standardizing business processes, we can now do three things: create more standardized software for use by increasing numbers of SMEs, standardize the data flow via XML across enterprises and also allow SMEs to work together for bigger projects (since they start resembling an extended, virtual company).
There are four types of software which are needed (the first three are common across industries, while the fourth is customised per industry):
Server Software to enable Messaging (Email, IM, SMS) and Collaboration this is akin to what is available in MS-Exchange and Lotus Notes today. Along with this, we also want to make available software for Website management, Content management (especially, weblogs), software for the creation of Intranets within the company, and Security Software (firewall, proxy, anti-virus and VPN). The Server software will interact with its peers within the company (in the case of multiple locations) to replicate data; this will be done via the Central Servers on the Internet.
Client Applications for Email, IM, Web Browser, Office Productivity (word processor, spreadsheet, presentation with data formats compliant with MS-Office), Acrobat, multimedia (audio, video, image management). These will all run off the server but available via the thin client on the desktop. All storage will be done centrally on the server.
Enterprise Applications (Generic) will be in the form of building blocks which perhaps do 70-80% of what the company and industry needs. One of they requirements for the enterprise applications is integration data should be entered only once, and available across the vertical applications. Enterprise Information cannot reside in silos. This is the notion of an e-business suite which Oracle and others talk about.
Enterprise Applications (Industry-specific) are created by third parties on top of the Web Services platform we will make available. Some of these applications may be priced additionally.
Building an Ecosystem
I dont think we can create every software application that will be needed. What we want to do is to use a Reverse Walmart strategy. In Walmart, as specific products become popular, Walmart looks to replace or complement some of them with cheaper in-store brands, which have higher margins. We will do the reverse. We will start by using in-store software which has been developed by us so that we can seed the entire exercise. At the same time, we will publish the APIs and XML data formats for the modules, so others can develop replacements (or in fact, better versions) for the software. These then can be made available to SMEs for small, additional fees depending on their needs.
This strategy helps us to build an ecosystem around what we are doing. We want to build the platform, and not necessarily all the applications. Over time, this makes us an SME aggregation and distribution system (like Walmart is a mass market customers aggregation and products distribution system).
Used PCs: From Where?
A big question is on the Used PCs. Where are they going to come from? We have to figure out the supply chain of second-hand PCs in the developed markets, especially US and then see how to aggregate them and transport them to India and the other markets, and then test/certify them for use in the SMEs. This is a big challenge and one of the key unknowns as of now. It is also a very critical component in our vision of bringing down the cost dramatically. Unless the cost comes down to the USD 100 level, volumes will not take off. And without the desktops, our software market will be very small (since we are charging per person). The only way, according to me, to bring down the cost to the USD 100 level is by using second hand PCs there is no new PC whatever be the configuration which can be made available at that price point. We do not want to create, we want to aggregate. The PCs are already there, we need to figure out the supply chain to get them to the desktops of SMEs in the emerging markets.