This article compares two manual water pumps operating in the same general geographic region: the Zimbabwe Bush Pump B-Type, and the South African PlayPump ®. The Bush Pump is concentrated mainly in Zimbabwe, though the pump is also manufactured and installed in other Southern African countries. Standardised in 1987, approximately 30,000 B-types had been installed in Zimbabwe by 2010, 12,000 of them since 1998. Over the equivalent period, from the mid-1990s till 2010, about 1,700 PlayPumps had been installed, mainly in South Africa, but also in Swaziland, Lesotho, Zambia, Malawi and Mozambique.

The ‘B-type’ Bush Pump is the latest version of a hand pump invented by a colonial water officer in Rhodesia in the 1930s. First redesigned in the 1960s for the Rhodesian government as the ‘A-type’ Bush Pump, in 1987 it was redesigned and standardized for the Zimbabwe government by engineer Dr. Peter Morgan. The trajectory of design for the Bush Pump over time has been to reduce the number of its working parts, increase its durability, and facilitate user maintenance of the pump. Users are intended to perform basic maintenance of the pump, while local, provincial and national repair crews undertake heavier repairs, funded by the state. The B-type is a conventional hand pump, operated by an adult or child through the up-and-down motion of the pump handle. In 2010, the hardware cost for an average pump installation was about US$1,200.

The PlayPump, first installed in a rural district of KwaZulu Natal in 1993, uses a conventional borehole pump similar to the Bush Pump’s, but with a novel addition: instead of a handle, the pump is driven by a children’s roundabout. The motion of the roundabout pumps water to an elevated water tank. The maintenance system for the PlayPump is also novel: billboards installed on the water tank are rented out for advertising to fund the maintenance of the pump. Maintenance is carried out by the company that manages the advertising on the billboards, Roundabout Outdoor, in response to SMSes and phone calls from pump users. Until 2010, the total cost to a donor for a PlayPump was US$14,000, including installation. The hardware cost in 2008 was approximately US$5,600.

The PlayPump’s two novel additions to the conventional hand pump, which promise work accomplished through play, and a self-sustaining maintenance system, have attracted an extraordinary amount of attention to the project. After winning a World Bank Development Marketplace Award in 2000, and through a series of positive reviews in the international press, the project accelerated rapidly. The UK company Global Ethics brought out a bottled water brand, One Water, to help fund the PlayPump, and the Case Foundation in the US set up the fund-raising organization PlayPumps International. PlayPumps International ran very successful campaigns to gather donations online, organizing major concert tours and celebrity endorsements. In 2006 the Case Foundation partnered with USAID and the President’s Emergency Plan for AIDS Relief to present PlayPumps International with a grant for US$16.4 million, intended as the first installment in a commitment to raise US$60 million to roll out PlayPumps across southern Africa.

But in 2009, support for the project began to decline sharply. Critical press reviews and blog posts, and previously unavailable reports on the PlayPump by UNICEF (2007), and by the Mozambiquan government (2008) identified a number of problems with the project on the ground. Children’s play was not an adequate source of input to the pump, so adults had to turn the roundabout by hand, and they found this much less efficient than a hand pump’s lever. Users were also dissatisfied, particularly outside South Africa, with the amount of time it took for faults to be repaired. In addition, a lack of advertising uptake on the PlayPump was identified, with many billboards blank, indicating that funds were not coming in for maintenance. The Case Foundation withdrew from the project and dissolved PlayPumps International in early 2010.

The dissatisfaction of users of the system was compounded by the fact that, contrary to the impression created by the PlayPump’s publicity, the project did not bring water where there was none before, but were usually placed on existing boreholes, replacing hand pumps. Where users had hand pumps before, they preferred these, and complained about a lack of consultation in their choice of technology. PlayPumps are also much more expensive for funders than conventional handpumps. As a position statement on PlayPumps by the charity WaterAid noted in 2009, at a cost to the donor of US$14,000 for a single PlayPump, four conventional hand pumps could be installed.

One Southern African country in which PlayPumps are not installed is Zimbabwe, because the state there will not enter into an agreement with the PlayPump’s producers to waive import duties on its hardware – and Zimbabwe has its own successful national hand pump, the Bush Pump ‘B-type’, which it encourages water and sanitation organisations working in the country to adopt. The Bush Pump, as noted, is designed to be a simple, robust hand pump, which outperforms many similar hand pumps.

The Bush Pump is the subject of a much-cited paper by science and technology scholars Marieanne de Laet and Annemarie Mol , ‘The Zimbabwe Bush Pump: Mechanics of a Fluid Technology’ (2000). In their paper, de Laet and Mol argue that what makes the Bush Pump an appropriate technology is its ‘fluidity’ – a quality they define in part as the pump’s ability to keep working even in disrepair, and in the pump’s changing configuration over time as it has been progressively redesigned.

The boundaries of the pump are ‘fluid’ too in that they embrace its users: the trajectory of its design has been to make it as open as possible to maintenance and repair by its users, and communities are involved in siting and managing pumps. This makes it more likely that the pump will keep working even when it breaks. Redesign of the pump has been in response to observations of its use – de Laet and Mol describe the Bush Pump’s designers as ‘fluid’ in their letting go of control, in learning from how people use the pump.

Some of the PlayPump’s problems could conversely be attributed to its lack of ‘fluidity’. Maintenance of the PlayPump is not possible by users, and repairs must be undertaken by the pump’s producers or by their contractors in other countries. This has contributed to long delays in the repair of pumps. As opposed to the ‘open’ configuration of the Bush Pump, the PlayPump’s pump is sealed within the roundabout, and users cannot undertake even minor repairs or maintenance. Users do not seem to have been effectively involved in the siting and management of pumps.

In contrast to the Zimbabwe Bush Pump, there has been little change to the design of the PlayPump over time, and little evidence of responsiveness to how the technology is experienced by users. Both UNICEF’s and the Mozambiquan government’s reports on the PlayPump recommended that users be given an informed choice of water technology, and that the system’s maintenance and advertising system be made more locally accountable. Part of the problem with the PlayPump has been that its ‘success’ has been driven largely by its appraisal by bodies other than the users of the technology. There is ample evidence of its success as an image: press reports and comments on social networking sites testify to how compelling audiences removed from the project find its message of work accomplished by children’s play.

The importance of this image, which relies on the roundabout as input to the pump, has arguably ‘frozen’ the development of the technology on the ground. Maintaining the effectiveness of its image has restricted its ability to change in response to what is learned from its use, so reducing its ‘fluidity’ in response to the user. This situation is exacerbated by the highly unequal actors connected by the PlayPump: developing world users and ‘first world’ audiences. The PlayPump facilitates a situation in which users of the technology must perform to a script written for audiences to it – in conforming to the image of play accomplishing work, adults must push a children’s roundabout around in a parody of play.

The story of the PlayPump highlights a risk inherent in the growing reliance on first world audiences to fund and support design for developing world users: that projects which are effective story-tellers and compelling image-makers may be selected for over technologies whose main effectiveness is for users, not audiences. Though it is possible to improve the design of the PlayPump, both in its hardware and in the way it engages with users, a more fundamental question precedes this: is the PlayPump responding to a real need? If the trajectory of its development has been determined from the outset not by its success with users, but with external audiences, then perhaps advancing the development of existing technologies with a proven record of success for users would be a more appropriate course of action.

Article commissioned and edited by Design With Africa, and written by Dr Ralph Borland.