Formula and interpretation

Expected failures/year = pumps × failures per pump-year. Pump-outage days avoided/year = expected failures × max(0, current average outage days − target average outage days). The optional person-day proxy multiplies avoided pump-outage days by users per pump; the optional cost proxy multiplies avoided pump-outage days by a user-supplied operational disruption cost.

The calculator assumes the current and target average outage durations apply to the same failure definition and comparable fleet. If failure frequency changes when monitoring is introduced, model that separately. Averages can hide long-tail outages, so operational programmes should also track median and percentile repair times where enough events exist.

OutputUse it forDo not interpret it as
Expected failures/yearWorkload planning and scenario analysisA guaranteed future failure count
Pump-outage days avoidedOperational exposure if target repair time is achievedMeasured benefit before target is achieved
Person-day proxyScale of service exposure under a simple users/pump assumptionIndividual consumption, health or welfare impact
Cost proxyUser-defined operational scenarioA universal economic value of water service

Use before/after evidence carefully

Published Kenyan handpump studies reviewed by Oxford show how repair times can fall dramatically when automated fault information is paired with an organized response service. One study cited in the review reported average repair time falling from 27 days before the trial to 2 days for pumps with automated data; another reported less than 3 days from a 37-day pretrial average. These are deployment results from specific programmes, not a forecast for another fleet.

To evaluate a programme, define failure and restoration consistently, retain event and ticket timestamps, and compare distributions over comparable periods. Record changes in staffing, spares, transport and reporting practices because monitoring is only one part of the response system.

  • Choose a fixed failure definition before comparing periods.
  • Track detection → acknowledgement → dispatch → verified restoration separately.
  • Report event count and distribution, not only one average.
  • Document censored/open outages at the end of a reporting period.

Method, assumptions and limitations

Version: 2026-08-23. All calculations are deterministic and run in the browser. No values are transmitted or stored by this static site. Currency is intentionally unspecified because the cost proxy is supplied by the user and may represent local transport, temporary service, contractual penalties or another clearly defined operational cost.

This tool does not estimate health outcomes, economic welfare or the value of access to water. Use a specialist evaluation framework for those questions. Its purpose is to make the operational downtime assumption visible and easy to replace with observed fleet data.

Sources and limits

Use these references to verify the underlying guidance. Local regulations, operator coverage and manufacturer instructions can change the correct implementation.