Uduma — In-line Chlorination Pilot (November 2024)

Note: This page summarizes the rationale behind a GiveWell grant to Uduma. Uduma staff reviewed this page prior to publication.

Published: February 2025

1. The organization

Uduma is a French company that manages water infrastructure on behalf of local authorities. It currently operates in Mali, Burkina Faso, Côte d'Ivoire, and Benin.1 Uduma estimates that it currently has around 1.5 million customers, of which around one million are in Benin.2 Uduma is responsible for the maintenance and operation of the infrastructure it manages, as well as for collecting tariffs.3 Uduma’s infrastructure predominantly consists of handpumps, solar-powered water stations, and larger piped networks. Uduma currently tests each water point twice a year and chlorinates the water supply if test results show it to be contaminated, but does not chlorinate as a matter of course.4

2. The intervention

We think that water quality is one of the most promising grantmaking areas for GiveWell, primarily due to the evidence of its effect on reducing under-five child mortality. Our full report on water quality interventions can be found here.

In-line chlorination (ILC) is one of the two chlorination interventions to which GiveWell has directed significant funding to date, alongside chlorine dispensers.5 In-line chlorinators are typically installed on a pipe feeding a water storage tank. As water flows past the chlorinator, it disinfects the water.6

There are a wide variety of commercially available in-line chlorination devices.7 For this pilot, Uduma plans to test tablet erosion chlorinators, similar to those GiveWell has funded in Malawi and India, and mechanical liquid injection chlorinators, which we have not previously funded at scale.8

3. The grant

This grant will fund a pilot of in-line chlorination on 10 of Uduma’s water systems in Mali.9 The pilot will cover a range of Uduma’s infrastructure (both solar water stations and larger, piped water systems).10 In addition, the grant will fund a non-profit research organization, Aquaya,11 to conduct an evaluation, and Mangrove Water,12 a specialist chlorination advisory organization, to provide technical assistance to Uduma on its chlorination processes.

The total cost of the grant is $480,501. This breaks down as follows:

• Uduma: $292,44113
• Aquaya: $162,18514
• Mangrove Water: $25,87515

3.1 Grant activities

Uduma will conduct ten in-line chlorinator installations. It will also use this pilot period to:16

• Conduct community engagement to address potential concerns around the smell and taste of chlorinated water;
• Investigate the chlorine supply chain in Mali;
• Develop its operating procedures, including for water sampling and testing, ongoing operation of the devices, and the handling and storage of chlorine (with support from Mangrove Water);17
• Select field-testing equipment;
• Investigate the feasibility of covering ongoing operational costs of chlorination in the future with carbon credits (Uduma has allocated approximately $40,000 for this).18

The grant covers all of Uduma’s operating and infrastructure costs for the pilot phase. The overall pilot timeline is expected to be 15 months, starting in January 2025.19

Aquaya will conduct data collection on:

• Microbial contamination levels at baseline and endline at the point of consumption (from household storage containers) and at the point of collection (at household taps or public standpipes);20
• Chlorination levels at baseline and endline at the household level and point of collection;21
• An estimate of the number of people using water points where chlorination is installed (Aquaya has suggested using a combination of Uduma’s metering data, household surveys, and geographic data techniques);22
• Data on chlorine acceptance in one of our highest-priority countries, Mali (this will be done through household surveys and focus groups).23

For more details on what we expect to learn during the pilot, see here.

4. The case for the grant

4.1 Chlorination of Uduma’s infrastructure is plausibly cost-effective at scale

Our best guess is that if implemented at scale, the chlorination of the infrastructure that Uduma manages in Mali and Burkina Faso (the two countries for which we would consider providing scale-up funding)24 will be highly cost-effective. We estimate that an at-scale chlorination program in both countries would be 12x as cost-effective as unconditional cash transfers (GiveWell’s benchmark for comparing different programs).25 This represents the impact of installing chlorination on all Uduma-managed infrastructure over a four-year period and maintaining it for six years (10 years total).26

A simplified version of our cost-effectiveness analysis of a potential at-scale program in Mali and Burkina Faso is below.27 You can find the full version here.28

What we are estimating	Mali	Burkina Faso
Estimated grant size
Total GiveWell costs over 10 years	$2.3m	$2m
Child mortality benefits
Annual cost per child under 5 targeted29	$7.80	$7.69
Increase in the proportion of targeted children drinking treated water as a result of the program (percentage points)	49%	57%
Annual cost per additional child drinking treated water as a result of the program	$15.80	$13.58
Annual all-cause mortality rate among children under five	2.1%	1.3%
Effect of chlorination on under-5 all-cause mortality among children treated, from RCTs	24%	24%
Internal validity adjustment	-25%	-25%
Adjustment for proportion of deaths linked to water quality in program vs RCT contexts	-2%	10%
Effect of chlorination on under-5 all-cause mortality, adjusted	18%	20%
Number of under-5 deaths averted	560	394
Moral weight placed on each under-5 death averted, adjusting for age profile of the population (units of value)	114	113
Initial cost-effectiveness estimate from under-5 mortality benefits (in multiples of GiveDirectly’s unconditional cash transfer program)	8x	6x
Percentage of program value from different benefits
Percent of benefits from averted under-5 mortality	56%	55%
Percent of benefits from averted over-5 mortality	27%	28%
Percent of benefits from morbidity reductions	2%	2%
Percent of benefits from long-term income increases	15%	15%
Additional adjustments
Adjustment for grantee-level factors	0%	0%
Adjustment for additional economic benefits from the program	+25%	+25%
Adjustment for diverting other actors' funding into the program (leverage) and out of the program (funging)	-27%	-35%
Overall cost-effectiveness of Uduma’s chlorination program at scale (in multiples of GiveDirectly’s unconditional cash transfer program)	13.4x	9.5x

The key drivers of our at-scale cost-effectiveness estimate are:

• High child mortality burdens in Mali and Burkina Faso. We estimate that the annual under-five all-cause mortality rate is 2.1% in Mali, and 1.3% in Burkina Faso.30 The Mali estimate is substantially higher than other countries where GiveWell has funded chlorination programs to date, while Burkina Faso is in the same range.31
• The high underlying cost-effectiveness of chlorination. We think that chlorination is generally a highly cost-effective intervention. This is because it is inexpensive and there is evidence that it causes significant reductions in child mortality in areas with high burdens of waterborne disease. For more on the evidence for water chlorination’s impact, see our full intervention report. In Mali and Burkina Faso we estimate that the program would reduce child mortality by approximately 20% among children who drink chlorinated water.32
• We expect in-line chlorination to significantly increase water treatment rates. This is based on evidence that in-line chlorination can achieve household-level chlorination rates of around two thirds,33 as well as low baseline water treatment rates in both Mali (11%) and Burkina Faso (3%).34
• We expect Uduma will be able to cover some of its operating costs with other sources of revenue. Uduma believes it can use (a) carbon credits and (b) payments from Uptime Water, a results-based rural drinking water payment provider,35 to fully cover the ongoing operational costs of an at-scale chlorination program.36 Uduma estimates these costs could come to around $460,000 per year at scale in Burkina Faso and Mali. Even given our skeptical assumption that GiveWell will in fact need to cover 15% of these operating costs,37 this implies 33% of the cost of running the program over ten years will be leveraged from other sources in Burkina Faso, and 52% will be leveraged this way in Mali. Our best guess is that the cost-effectiveness of others’ spending is lower than the cost-effectiveness of the marginal GiveWell funding opportunity, and so this increases cost-effectiveness relative to a program where GiveWell would cover 100% of the cost.38 (For more on how we think about leverage, see this blog post.)

Limitations and uncertainties of our analysis include:

• Costs. While Uduma has shared reasonably detailed cost estimates, we see these as uncertain because of the organization’s limited previous experience of chlorination.39 Our expectations for future monitoring and evaluation costs, which are not included in the costs Uduma shared, are also based on rough projections.
• The number of people reached. Uduma currently projects that its user numbers will rise from around 500,000 people today to around 910,000 people in 2028 in the countries where it currently operates (excluding Benin).40 We apply a rough -15% adjustment to account for the possibility that this might be overstated, but this is a rough and uncertain guess.41
• Water contamination rates prior to in-line chlorination.
  • Uduma’s own data on water quality at the point of collection for Mali and Burkina Faso in 2024 finds that 68% of samples in Mali and 79% of samples in Burkina Faso are safe.42 This implies that Uduma-managed water is substantially less contaminated than data from independent surveys in neighboring and other Sahelian countries would suggest.43
  • Uduma attributes this to sanitary inspections it conducts to reduce contamination risks when taking on new infrastructure and measures it takes to mitigate contamination risks, such as fencing to prevent animal access and ensuring toilets are located far from the water point.44
  • We don’t currently include an adjustment to account for these measures, which means that we’re implicitly assuming Uduma-managed water is as contaminated as the water in randomized controlled trials (RCTs) that our overall estimate of the mortality effect of chlorination relies on.45
  • We think this is a reasonable assumption because data we’ve seen from independent surveys in neighboring and other Sahelian countries suggests almost universal contamination at the household level in rural areas,46 and one chlorination expert we spoke to told us that the water is unlikely to be safe at the household level in the contexts where Uduma works.47 However, we’re uncertain about this, and it’s possible that we should be using more optimistic estimates of baseline water quality, which would reduce the cost-effectiveness of this program.
• Lack of plausibility cap. Due to time constraints, we haven’t included a plausibility cap of the effect of chlorination on mortality in our cost-effectiveness analysis, even though this is an important part of our cost-effectiveness analyses for other water chlorination grants.48 However, our best guess is that our existing estimate would fall below the cap, and that therefore there would be no impact on cost-effectiveness, but it’s possible that we are overestimating the impact of the program by not explicitly modeling this.

4.2 Learning value

Learnings about Uduma’s chlorination program

We expect the pilot to teach us about the operational feasibility of in-line chlorination on Uduma’s water systems, and give us some signal about the potential cost-effectiveness of the program:

• Operational feasibility. This pilot will primarily focus on testing how feasible it is for Uduma to adopt in-line chlorination on the water infrastructure that it manages. We think this is important because Uduma doesn’t currently have experience with in-line chlorination in Mali or Burkina Faso,49 and experts we have spoken to have emphasized that in-line chlorination is challenging to implement, particularly because of the wide variety in underlying water infrastructure, requiring adaptations to suit each context.50 We expect the pilot to address operational questions such as:
  • Which specific chlorination devices Uduma will use;51
  • How often it will need to refill these devices;
  • How it will engage in community sensitization around chlorinated water’s smell and taste;
  • How it will source the chlorine.

• Cost-effectiveness. We think that the pilot will address many of our current uncertainties around the cost-effectiveness of this program, and therefore increase the precision of our cost-effectiveness estimate. We expect to receive some signal on:
  • How much the program will cost: Uduma plans to report an itemized breakdown of costs.52
  • How contaminated Uduma’s water is: Aquaya will collect data on contamination levels at both the point of collection and the household.53 This will help to address our uncertainty around underlying water quality in Uduma-managed infrastructure.
  • The number of people reached: Aquaya will collect data on the number of people using the water points where in-line chlorination devices are installed.54 We plan to use this data to improve our estimate of Uduma’s current customer numbers (see above). However, the small scale of the pilot means we expect to remain uncertain about this even after we receive this data, and we don’t expect to learn about the accuracy of its future projections.
  • Chlorination levels and chlorine acceptance: Aquaya plans to collect chlorination data at the point of collection and the household level.55 This will help to address our uncertainty about whether Uduma can achieve high chlorination levels.
  • Consumer acceptance of chlorine. Aquaya plans to learn about attitudes to and acceptance of chlorine (e.g. of its taste and smell) through household surveys and focus groups.56 We are interested in this because we have heard that chlorine acceptance may be an obstacle to scaling chlorination programs in high-burden countries.57
  • The plausibility of Uduma covering its operational costs with other sources of revenue. Approximately $40,000 of Uduma’s overall budget will go towards assessing the potential for using carbon credits to cover its operating expenditures.58 We expect this will provide us with some indication of how plausible it is that Uduma will be able to run an in-line chlorination program without ongoing support from GiveWell, although we think that Uduma will need more time to know for certain whether this can work.

Broader learning value

This would be our first in-line chlorination implementation grant for an organization other than Evidence Action.59 As we have previously written, we are keen to deepen our understanding of the water sector as a whole. We think that working with Uduma will be informative for us, whether or not the pilot leads to a larger scale-up grant in the future. Some things we expect to learn more about through this grant include:

• Whether working with water utilities is a promising approach to fund chlorination programs. Unlike Evidence Action’s Safe Water programs, which add chlorination to water systems managed by other actors, Uduma is responsible for operating the water systems. We see funding an organization like this to do chlorination as an intuitively appealing approach, because it does not distribute operational responsibility among multiple actors. We expect to get some signal from this grant about how promising this approach is in general.
• The availability of chlorine in Mali. We have heard from experts that the availability of chlorine may be a barrier to expanding chlorination programs beyond the countries where Evidence Action already works.60 This grant will directly test this hypothesis in Mali.
• The promisingness of different chlorination devices. Uduma plans to test a mix of tablet erosion and liquid injection chlorinators during the pilot. To date, the grants we have recommended have almost exclusively funded tablet erosion chlorinators.61
• The broader water sector in Mali. Mali is one of GiveWell’s highest priority countries for expanding chlorination programs.62 We think it’s plausible that we’ll learn more about implementation feasibility in Mali (e.g. the types of water infrastructure, key actors, and funding landscape) through making this grant.

4.3 Making this grant could allow us to leverage funding for chlorination from other actors

The grant budget covers all of Uduma’s operational costs for the pilot. However, Uduma told us that, beyond the pilot, it hopes to fund ongoing operational costs at-scale using other sources of revenue, namely Uptime performance-based payments and carbon credits (more above). Approximately $40,000 of the grant budget is also allocated for Uduma staff to investigate the feasibility of using carbon credits for this purpose.63

If this materializes, this grant could thereby have unlocked opportunities to fund the chlorination of Uduma’s infrastructure at scale in an arrangement whereby GiveWell would cover Uduma’s installation costs, and other actors would cover its operational costs. We think this would have two benefits:

• It would increase the cost-effectiveness of our spending. Our best guess is that these other sources of revenue would be used for something less impactful than chlorination in our absence. This increases the impact of our funding relative to a scenario where GiveWell covers 100% of costs, by leading to an increase in the value of others’ spending. (Our current assumptions about leverage are an important part of our cost-effectiveness of this grant, as described above.)
• It would make this program potentially more sustainable than others that GiveWell has funded. Historically, our chlorination grants have tended to rely on ongoing philanthropic funding.64 All else being equal, we think that programs with more diverse sources of ongoing funding are likely to be more sustainable. Our general impression, based on conversations we have held over the past year, is that this grant likely moves us closer towards the recommendations and practices of other funders in the sector.

4.4 Positive qualitative impression of Uduma

This impression is partly based on our extensive interactions with Uduma throughout this grant investigation, where we were impressed with Uduma’s level of rigor and willingness to adapt to our requests (e.g., our requests that they work with Aquaya and Mangrove Water, and that they reduce the initial pilot size). It is also partly based on the broadly positive feedback we’ve received about Uduma from experts in the sector. We haven’t been able to verify these impressions in practice, but we expect the pilot will give us a signal on whether our current positive qualitative impression is accurate.

5. Risks and reservations

5.1 Low estimated cost-effectiveness of the pilot grant and uncertainty about cost-effectiveness of an at-scale program

Our primary cost-effectiveness estimate for the grant itself (as opposed to chlorination of Uduma’s infrastructure at scale) is that it is 7x as cost-effective as direct cash transfers (GiveWell’s benchmark for comparing the cost-effectiveness of different programs).65 At the time of writing this page, this is below GiveWell’s normal funding threshold (currently 10x or more as cost-effective as cash transfers).66

We model the cost-effectiveness of the pilot in terms of the “option value” it may create by unlocking opportunities to fund chlorination of Uduma-managed water infrastructure at a larger scale (by demonstrating this is feasible).67 Because our best guess is that chlorination of Uduma’s infrastructure is more cost-effective than our most likely alternative funding opportunities (~12x as cost-effective as cash transfers vs ~10x),68 we think this pilot could enable GiveWell to redirect funds in a somewhat more impactful direction than we otherwise would have.

We have decided to make the grant for this pilot even though our primary estimate is below our cost-effectiveness threshold because:

• We put significant weight on factors that are not captured by this cost-effectiveness model, e.g. the broader learning value of the grant.
  
• We have an internal disagreement about this cost-effectiveness estimate, with two GiveWell staff members involved in this grant investigation coming up with different estimates (7x vs 18x as cost-effective as direct cash transfers). We put weight on both estimates when considering whether to make this grant.

Our cost-effectiveness analysis for the pilot is available here. See the appendix for a summary and the factors driving the differences between our different estimates.

5.2 Risks associated with this grant

We see this grant as moderately high risk overall. Possible risks include:

• Overall pilot failure. Our best guess is that there is a moderately high (25%–30%) chance that the pilot will fail outright (e.g., Uduma is not able to install and operate chlorination equipment at a reasonable cost and level of effort). This is primarily due to (a) Uduma’s limited experience with chlorination,69 (b) the fact that Uduma is producing its own chlorine solution for its liquid injection chlorinators, and (c) the need to work in a context with high security risks:
  • Limited experience with chlorination: Experts we have spoken to have told us that implementing in-line chlorination is challenging.70 In particular, devices need to be adapted to the specific local infrastructure, and our understanding is that achieving the correct chlorine dosage consistently is very difficult, particularly for organizations that are relatively less experienced in this.
  • Uduma plans for its own staff to produce the chlorine solution they intend to use for their liquid injection chlorinators (rather than procuring it from external suppliers).: This is different to Evidence Action’s model, where it is procured from external suppliers. Daniele Lantagne, an expert on chlorination, told us that she believes this is a reasonable overall approach, but it could increase the risk that Uduma’s chlorine fails to meet quality or dosing standards.71 Our expectation is that Mangrove Water’s involvement in this grant should reduce this risk.
  • Security risks in Mali: Our understanding is that parts of Mali are affected by security issues. Uduma told us that its water systems in Mali are challenging to access, but safe to operate, which reduces our concern.72 However, we think it’s possible that there could be issues if the security situation deteriorates, or if challenges with accessing Uduma’s water points makes it more difficult, for example, for Aquaya to conduct monitoring and evaluation.

• There could be some risk associated with working with a for-profit water utility. We have limited experience making grants to for-profit organizations in general, and it’s possible that this could lead to unanticipated risks. For example, Uduma charges its users a small fee for access to the water it provides. We were initially concerned about possible risks associated with this, e.g., whether Uduma could unilaterally decide to significantly raise its tariffs. However, Uduma has told us that its tariffs are capped in the locations where it works,73 and we have seen contracts to this effect,74 so we are relatively less concerned about this type of risk.75

5.3 Limitations in what we’ll learn from the grant

Although we see learning value as an important part of the case for this grant, there are also some important limitations to what we’ll learn:

• The pilot is small (10 water systems) and will take place in Mali only. We’d expect the learnings about the other contexts where Uduma works to be more limited.
• It’s possible that our learnings about the promisingness of working with a water utility will not generalize beyond Uduma. This is because we’d expect other water utilities to vary substantially in their technical capacity, interest in chlorination, underlying water infrastructure, business model, and other factors.
• Also, while we expect to gather substantial data as part of the pilot, this will not address all the uncertainties in our cost-effectiveness analyses (details in footnote).76

6. Plans for follow up

We plan to keep in close contact with Uduma, Aquaya, and Mangrove Water during the pilot period, and will schedule calls with them around key pilot milestones, such as shortly following the beginning of installations, and after endline data collection.

If the pilot is successful, we will consider making a grant to Uduma to scale-up in-line chlorination in both Mali and Burkina Faso.

7. Internal forecasts

For this grant, we are recording the following forecasts, which represent the average of two GiveWell staff members’ forecasts:

Confidence	Prediction	By time
40%	GiveWell makes or recommends a grant to scale up a chlorination program on Uduma-managed infrastructure in Burkina Faso and Mali (grant size of >$1.5m).	By June 2026
45%	Conditional on Aquaya conducting analysis of free chlorine residual rates at the household level, GiveWell will conclude that Uduma managed to achieve >60% chlorination rates at pilot endline.	By June 2026
55%	Conditional on Aquaya conducting analysis of water contamination rates at the household level during the pilot, GiveWell will conclude that >70% of household level water samples at baseline are at “high” risk of e.coli contamination (using the Demographic and Health Survey/Multiple Indicator Cluster Survey thresholds).	By June 2026

8. Our process

We identified this grant opportunity through our conversations this year with water sector experts, who proactively recommended Uduma to us. We held several conversations with Uduma, Aquaya, and Mangrove Water, as well as outside experts and organizations familiar with Uduma’s model or in-line chlorination, like Paul Gunstensen, Daniele Lantagne, and Uptime Water. We also put a large number of questions to Uduma in writing, which inform much of the case for the grant above. Our cost-effectiveness analysis of a potential at-scale program is here, and our cost-effectiveness analysis of the pilot is here.

Appendix: Breakdown of our pilot cost-effectiveness analysis

We model the cost-effectiveness of funding this pilot specifically, in addition to our cost-effectiveness analysis of Uduma’s program at scale (which we discuss here). Here, the primary benefit we’re accounting for is the value of the information that we expect the pilot grant to provide, which we estimate according to the size and the expected cost-effectiveness of the grantmaking opportunities that would be unlocked in the event that the grant allowed us to become confident enough to fund the program at a larger scale.

As we mention above, we have two internal estimates of the cost-effectiveness of the pilot, one which is above our cost-effectiveness bar, and one below. Simplified versions of these are as follows:

What we are estimating	Scenario 1	Scenario 2
Grant size
Total pilot costs	$0.48m	$0.48m
Grant cost going toward direct benefit	$0	$0.2m
Potential value generated
Expected cost-effectiveness of counterfactual opportunities (x cash)	9.6x	9.6x
Expected annual room for more funding (RFMF) over 10 years if we decide to scale up Uduma’s program	$0.43m	$1m
Best guess cost-effectiveness of reallocated funding (x cash)	11.6x	11.6x
Units of value generated through reallocated funding if we decide to scale up Uduma’s program in target countries	2,863	6,564
Likelihood that funding this program will cause us to reallocate funding
Probability the pilot fails	30%	25%
Probability that we decide to scale up Uduma's program after the pilot (assuming no failure)	57%	72%
Subtotal: Overall probability that we decide to scale up Uduma's program after the pilot	40%	54%
Expected value generated
How long we would fund a scaled-up program (years)	10	10
Expected increase in units of value from funding this program per year	1,237	3,776
Units of value generated by direct benefits during the pilot	0	7,754
Present-discounted expected value of increase in units of value from funding program	10,030	38,383
Subtotal: Cost-effectiveness from funding program (x cash)	6x	24x
Ad-hoc adjustment for additional program benefits and downsides	+5%	-25%
Overall cost-effectiveness	7x	18x

The difference in these two bottom-line estimates is driven by the following:

• Uncertainty around the total room for more funding we expect the grant to unlock: Our best guess (Scenario 1) is that the total funding gap to install chlorination infrastructure in Burkina Faso and Mali is around $4.3m over ten years, including monitoring and evaluation costs. However, we’re much less certain about how much grantmaking potential this pilot could inform beyond these two countries. Uduma has told us that it is currently considering expanding into Togo, Guinea, and Uganda, which could plausibly open up additional room for more funding (RFMF).77 In Scenario 1, we account for this with a 25% supplemental adjustment, but in Scenario 2 we assume that this unlocks significantly more grantmaking opportunities, and raise our estimate of the total funding gap to $10m over ten years (or by $0.57m/year compared to Scenario 1).
• Uncertainty around how likely it is that we decide to support program scale-up: In Scenario 1, we assume that we are only 40% likely to scale up the program in Mali and Burkina Faso following the conclusion of the pilot, while we assume this is 54% likely in Scenario 2. This difference is driven by a lower estimate of the chance of pilot failure, and a higher chance of scale-up if the pilot succeeds (which would be affected by whether or not our at-scale cost-effectiveness estimate falls in the meantime).

Sources

Document	Source
Aquaya, Home	Source (archive)
Daniele Lantagne, email to GiveWell, November 20, 2024	Unpublished
Evidence Action, “ILC Kenya Pilot Results Report,” 2022	Source
GiveWell Blog, “Re-evaluating the Impact of Unconditional Cash Transfers,” November 2024	Source
GiveWell Blog, “Research Strategy: Water,” July 2024	Source
GiveWell Blog, “Revisiting leverage,” February 2018	Source
GiveWell, “2020 moral weights,” 2020	Source
GiveWell, “Approaches to Moral Weights: How GiveWell Compares to Other Actors,” November 2017	Source
GiveWell, “Evidence Action — India In-Line Chlorination Pilot Expansion,” July 2023	Source
GiveWell, “Evidence Action — Scale-Up of In-Line Chlorination in India,” September 2023	Source
GiveWell, “Evidence Action's In-Line Chlorination Program — General Support”, July 2022	Source
GiveWell, “GiveWell's Cost-Effectiveness Analyses,” December 2024	Source
GiveWell, “How We Produce Impact Estimates,” February 2024	Source
GiveWell, “Internal forecasts,” October 2024	Source
GiveWell, “Uduma – running Q&A document for chlorine pilot”	Unpublished
GiveWell, CEA for GiveDirectly's unconditional cash transfers, 2022	Source
GiveWell, Dispensers for Safe Water grant size, costs and room for more funding, 2021 (redacted)	Source
GiveWell, Notes from call with Daniele Lantagne, March 19, 2024	Unpublished
GiveWell, Notes from call with Daniele Lantagne, November 1, 2024	Unpublished
GiveWell, Notes from call with Daniele Lantagne, November 20, 2024	Unpublished
GiveWell, Notes from call with Paul Gunstensen, October 18, 2024	Unpublished
GiveWell, Notes from call with Uptime, November 12, 2024	Unpublished
GiveWell, Uduma in-line chlorination CEA, February 2025	Source
GiveWell, Uduma ILC pilot in Mali – Option value CEA, Feburary 2025	Source
GiveWell, Vouchers Water Quality CEA, February 2023	Source
GiveWell, Water quality CEA (ILC and DSW), 2022	Source
Lindmark et al. 2022	Source
Mangrove Water, Uduma Budget 2024	Unpublished
Pickering et al. 2019	Source
The Aquaya Institute, "Evaluating the effectiveness of chlorination by an innovative commercial rural water supplier in West Africa," November 2024	Source
The DHS Program: Demographic and Health Surveys, Home	Source (archive)
Uduma, "Considerations for Introducing Chlorination to Uduma’s Services in Burkina Faso, Côte d’Ivoire, and Mali," October 2024	Source
Uduma, Budget: Uduma Chlorination Pilot in Mali, 2024	Unpublished
Uduma, email to GiveWell, November 25, 2024	Unpublished
Uduma, Home	Source (archive)
Uduma, "OpEx for all systems in 2028," 2024	Unpublished
Unicef, Multiple Indicator Cluster Surveys (MICS)	Source (archive)
Uptime Water, Home	Source (archive)

• 1

  “UDUMA helps improve access to drinking water in rural areas of sub-Saharan Africa by taking charge of water supply equipment and ensuring its proper operation and maintenance for 15 years.” Uduma, Home

• 2
  • GiveWell: “We have a note from our conversation in March stating that you have around 500k customers in Cote d’Ivoire / Mali / Burkina Faso and ~1m in Benin. Is this estimate still correct?”

  Uduma: “Yes – there are small variances in Burkina Faso due to the insecurity, but generally this is still our best estimate.”

  • GiveWell: “Does this estimate [of 500,000 customers in the countries other than Benin] refer to all 3 countries together, or each country separately?”

  Uduma: “The 500,000 is for all three countries together, Burkina Faso, Côte d’Ivoire, and Mali.”
  GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished)

• 3

  This was confirmed by Uduma during its review of this grant page.

• 4

  “We shock chlorinate once we identify a source with contamination. We will share via email test results from Burkina Faso (146 tests from 2024) and Mali (961 tests from 2019 to 2024).” This quote and the breakdown of water infrastructure are taken from Uduma’s responses in GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished).

• 5

  For more on GiveWell’s approach to chlorination to date, see our blog post: GiveWell, “Research strategy: Water”.

• 6

  See this section of a previous GiveWell grant page for an in-line chlorination program in Malawi, which compares ILC to Evidence Action’s “Dispensers for Safe Water”, for more details. Note that this description is of a tablet erosion chlorinator, the only type of in-line chlorinator GiveWell has funded at scale to date.

• 7

  For more, see Lindmark et al. 2022. Table 2, “Evaluations Conducted on Passive Chlorinators Identified in Peer-Reviewed Literature,” lists many different in-line (“passive”) chlorination devices.

• 8
  • “In Burkina Faso, Mali, and Côte d’Ivoire, we are proposing two types of technologies, erosion chlorinators, which are used for medium-sized systems (solar water stations serving approx. 500 people). These are less accurate at dosing, but more reliable considering the context (simple to operate and maintain). The second technology is mechanical injection technology, from the brand Dosatron. Compared to erosion chlorinators, these are more accurate for larger systems (approx 1,000 - 8,000 people), but harder to operate and maintain (i.e., better trained personnel required).” GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished).
  • Evidence Action previously tested larger chlorination devices on a limited scale during a GiveWell-funded pilot in India. For more on this, see here.

• 9

  Uduma, Budget: Uduma chlorination pilot in Mali (unpublished)

• 10
  • Five of these systems are medium-size solar water stations, where Uduma will fit a tablet erosion chlorinator (similar to those used by Evidence Action in GiveWell-funded programs in Malawi and India);
  • The other five are larger, piped water systems. Uduma plans to install small injection chlorinators on four of these, and a large injection chlorinator on one.

• 11

  “Aquaya is a nonprofit research and consulting organization dedicated to advancing global health through universal access to safe water, sanitation, and hygiene (WASH).” Aquaya, Home.

• 12

  Mangrove Water is a new organization focused on providing support for the implementation of in-line chlorination, including advising on site evaluation, installation, operations and maintenance, and monitoring and evaluation. Mangrove Water was set up by Amy Pickering and Jeremy Lowe at the University of California, Berkeley, and Megan Lindmark at Cova, a Central America-focused water provider that operates in-line chlorination at scale.

• 13

  Uduma, Budget: Uduma chlorination pilot in Mali (unpublished). Uduma’s budget breaks down as follows:

  • Staff costs: $173,214
  • Consultants: $16,464
  • Travel and logistics: $20,080
  • Water treatment systems: $49,924
  • Chlorine: $2,971
  • Water quality testing: $1,214
  • Overheads on non-staff costs: $13,598
  • Contingency costs: $14,977

• 14

  Uduma, Budget: Uduma chlorination pilot in Mali (unpublished)

• 15

  Uduma, Budget: Uduma chlorination pilot in Mali (unpublished)

• 16

  Uduma, "Considerations for Introducing Chlorination to Uduma’s Services in Burkina Faso, Côte d’Ivoire, and Mali," October 2024.

• 17

  Mangrove Water will provide remote technical support for:

  • The design and installation of systems (including installation instructions and training materials and technical recommendations);
  • Operation and maintenance (including standard operating procedures and training materials);
  • Water quality monitoring protocols (including field testing protocols, guidelines for corrective actions, and sampling and testing training materials);
  • Community engagement / sensitization (by creating a summary document);
  • As-needed ongoing technical advice.

  Mangrove Water, Uduma Budget 2024 (unpublished).
  

• 18

  Uduma, email to GiveWell, November 25, 2024 (unpublished).

• 19

  Uduma, Budget: Uduma chlorination pilot in Mali (unpublished)

• 20

  “To examine chlorination effectiveness, we plan to measure water safety at baseline and after the chlorination intervention (RQ2). We will measure water safety via two water quality parameters: free chlorine residual and E. coli, both measured at the consumer water points (i.e., household taps and/or public standpipes) and household drinking water (e.g., from storage containers) (Table 1).” The Aquaya Institute, "Evaluating the effectiveness of chlorination by an innovative commercial rural water supplier in West Africa," November 2024, p. 3.

• 21

  “To examine chlorination effectiveness, we plan to measure water safety at baseline and after the chlorination intervention (RQ2). We will measure water safety via two water quality parameters: free chlorine residual and E. coli, both measured at the consumer water points (i.e., household taps and/or public standpipes) and household drinking water (e.g., from storage containers) (Table 1).” The Aquaya Institute, "Evaluating the effectiveness of chlorination by an innovative commercial rural water supplier in West Africa," November 2024, p. 3.

• 22
  • “Meter data: Uduma has indicated that their water connections are generally metered. Data from water meters will allow us to estimate the number of users based on volumetric consumption; however, this method will also require data on the average amount of water used per person (determined via household surveys and/or observation). If pre-paid meters are used, we may be able to use financial payment records to inform our estimates of water use.” The Aquaya Institute, "Evaluating the effectiveness of chlorination by an innovative commercial rural water supplier in West Africa," p. 4.
  • “Household surveys: During the selection process for the household surveys, we will go to randomly selected GPS points. As part of the household screening, we will ask households about their main water source, alternative sources, and seasonal use patterns. This screening data from the randomly selected GPS points will create a population-representative sample of water source use, which can potentially be extrapolated to the broader community population. (After screening, we will do the full household survey with the piped water users.)” The Aquaya Institute, "Evaluating the effectiveness of chlorination by an innovative commercial rural water supplier in West Africa," p. 4.

• 23

  “Surveys and focus groups will ask households about their experience / perception of chlorinated water and explore what factors (messaging, involvement of leaders, types of services, etc.) resonate with consumers to encourage acceptance. We will also consider including discrete choice experiments into household surveys to measure household preferences for hypothetical scenarios”. Confirmed by Aquaya in their review of this grant page, February 19, 2024.

• 24

  Uduma initially approached us with a proposal to scale up this program in Mali, Burkina Faso, and Côte d’Ivoire, but the latter came in below our funding bar in our cost-effectiveness analysis.

• 25
  • 12x is the weighted average of our at-scale cost-effectiveness figures in multiples of unconditional cash transfers for Mali and Burkina Faso (13.4x and 9.5x respectively).
  • At the time of writing this page, GiveWell’s funding bar is to fund grants that we estimate to be ~10x or more as cost-effective as cash transfers.
  • This benchmark is based on ‘moral weights’, a system we use to quantify the benefits of different impacts (e.g. increased income vs reduced deaths). We benchmark to a value of 1, which we define as the value of doubling someone’s consumption for one year. Our estimate of the value of direct cash transfers is 0.00335 per $. For more on how we use moral weights, see this document. For more on how we use cost-effectiveness estimates in our grant-making, see this page.
  • Note that our estimate of GiveDirectly's value per dollar donated is out of date as of 2024 (more in this blog post). We are continuing to use this outdated estimate for now to preserve our ability to compare across programs, while we reevaluate the benchmark we want to use to measure and communicate cost-effectiveness.

• 26

  Note that the 10-year timeline over which we model benefits is arbitrary. It reflects our rough guess of the period over which it is reasonable to attribute benefits to GiveWell in cases where we believe GiveWell’s funding caused a program to take place but GiveWell doesn’t cover most of the costs over time. The 10-year window we use here is broadly in line with other GiveWell cost-effectiveness analyses in similar scenarios, although there may be some variation case-by-case.

• 27

  These figures are drawn from this sheet of our cost-effectiveness analysis.

• 28

  Our full cost-effectiveness analysis includes an analysis of Côte d’Ivoire, as Côte d’Ivoire was suggested by Uduma as another potential expansion location. However, we ultimately decided to deprioritize Côte d’Ivoire as an expansion location, due to low expected cost-effectiveness.

• 29

  This refers to the annual cost per child under five using an Uduma-managed water point where chlorination is installed.

• 30

  These figures are based on figures from the Institute for Health Metrics and Evaluation, the United Nations, and the Demographic and Health Surveys program. For more detail on this, see this cell note in our cost-effectiveness analysis of Uduma’s program.

• 31

  Compare these figures to 1.1% to 1.3% under-five annual all-cause mortality in Kenya, Uganda, and Malawi, as calculated for our cost-effectiveness analysis of Evidence Action’s Safe Water programs.

• 32

  See this row of our simplified cost-effectiveness analysis.

• 33

  Our estimate of the chlorination rates that Uduma will achieve using in-line chlorination (ILC) is based on two studies: (1) Pickering et al. 2019 (a randomized controlled trial of ILC in Bangladesh), and (2) an Evidence Action pilot of ILC in Kenya in 2021-2022. The former found detectable free chlorine residual (FCR) at the point of collection (at taps) 83% of the time and the latter found detectable FCR in water stored in surveyed households 68% of the time. Note that Pickering et al. 2019 only used low chlorine doses and it's possible that free chlorine would be undetectable at the household level. Since we are interested in chlorination in a different context (which may use higher dosing), we construct an alternative household level estimate by adjusting the Pickering et al. 2019 results based on the Evidence Action pilot, where among households served by water points testing positive for FCR, 82% of households tested positive for FCR. This implies household-level chlorination rates of 66% in Pickering et al. (80% x 82% = 66%).
  We construct a further estimate of 60% as the chlorination rates we think Uduma could achieve as follows:

  • We take a simple average of our FCR estimates for Pickering et al. 2019 (66%) and the Evidence Action pilot (68%).
  • We include a -10% adjustment because we think it’s likely that Uduma will achieve lower chlorination rates at scale. This is because of its relative lack of previous experience in chlorination and the general challenges of implementing a program at high quality at scale. This -10% adjustment is a rough guess.
  • ((66% + 68%) / 2) × 0.9 = 60%

  See this section of our cost-effectiveness analysis for calculations and sources.

• 34

  See this section of our cost-effectiveness analysis for calculations and sources. These estimates put 75% weight on the latest Demographic and Health Survey (DHS), adjusting for self-report bias, and 25% weight on our best guess of counterfactual water treatment rates of Evidence Action’s Dispenser for Safe Water (DSW) program in Kenya, Malawi, and Uganda. We put some weight on the estimates for Evidence Action’s program because they include some data on objectively measured chlorination rates (whereas the DHS estimates are all self-reported); however, we put limited weight on this data because it is from different countries to the countries in which Uduma works.

• 35

  “Results-based funding accelerates water service development and ensures that water keeps flowing. Water users pay for these services, but often not enough to cover the full service cost in rural and remote contexts. Results-based contracts help to address this funding gap. Over time, these rural water services can either transition out of requiring a subsidy or begin to access public funding locally. Interim catalytic funding is critical to service development in early stages when user payments do not fully cover operating costs.” Uptime, Our Solution.

• 36
  • “Yes, we believe we can cover the full OpEx costs with Uptime and carbon credits funding.”
  • “We believe we need at least 3 years of experience before we can conclude that [we] could fully cover our chlorination OpEx costs through carbon credits, which is why we are continuing to explore other results-based financing solutions in parallel.”

  GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished)

• 37

  We make this skeptical assumption because Uduma’s belief that it can cover all of these costs is based on uncertain projections, and WASH sector experts we’ve spoken to told us that carbon revenue (one Uduma's core revenue streams at scale) can be modest.

• 38

  We roughly assume that other philanthropic actors’ spending is about two times as cost-effective as cash transfers, against our funding bar of ‘10x cash’. This is a rough guess, benchmarked to the estimate we used when assessing Evidence Action's Dispensers for Safe Water program.

• 39

  Uduma, Budget: Uduma chlorination pilot in Mali (unpublished)

• 40

  GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished).

• 41

  Factors driving our -15% adjustment are:

  • Uduma has told us that its current customer number figures are based on assumptions of the number of people served by each type of infrastructure, rather than more precise methods (GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished)).
  • Uduma’s estimates of user growth between 2025 and 2028 are based on projections. In general, we believe projections like this are more likely to be overly optimistic than overly conservative.
  • Uduma told us that its estimates assume 400 users per hand pump on average (GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished)). This is significantly higher than Evidence Action's estimates of households served by each of the dispensers in its Dispensers for Safe Water Program (119 to 260, varying by country), many of which are installed next to handpumps. We put limited weight on this because Evidence Action operates in different countries to Uduma and there might be differences in the population structure and water infrastructure. However, we think this lends some weight to the possibility that Uduma’s estimates might be too high.
  • One countervailing factor is that Uduma also has estimates of “indirect” customers, i.e., people who use the supplied water but do not directly interact with the Uduma-managed water systems, based on population data estimates for the communities it serves. It thinks that these estimates might be more accurate because “in the majority of communities we serve, the entire population of that village uses our water” (GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished)).
  • In its 2028 projections, Uduma estimates the total number of users is >50% higher than the number of direct users (~1.5m vs ~900k across Mali, Burkina Faso and Côte d’Ivoire). If in some cases whole communities use Uduma-managed water systems, this could mean our adjustment is too conservative (Uduma, “OpEx for all systems in 2028,” 2024 (unpublished)).

• 42

  ‘Safe’ here is as defined by Uduma. GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished).

• 43

  Data from nationally representative household surveys, i.e. Multiple Indicator Cluster Surveys (MICS) and Demographic and Health Surveys (DHS), indicate the following contamination rates:

  • In Chad, 87% of the rural population was at high risk from E. coli at the water source, and 100% at high risk at home (MICS 2019, pp. 414-416, Tables WS 1.6-1.7).
  • In Côte d'Ivoire, 78% of the rural population was at high risk from E. coli at the water source, and 94% at high risk at home (DHS 2021, pp. 456-457, Tables 19.6.1-19.6.2).

• 44

  GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished).

• 45

  For the RCTs included in this analysis, see this sheet of our cost-effectiveness analysis. For more background on these RCTs, see this page. For our calculations of under-5 and over-5 mortality effects of chlorinating drinking water, see this section of our cost-effectiveness analysis.

• 46

  As above, from nationally representative household surveys, i.e. Multiple Indicator Cluster Surveys (MICS) and Demographic and Health Surveys (DHS), indicate the following contamination rates:

  • In Chad, 87% of the rural population was at high risk from E. coli at the water source, and 100% at high risk at home (MICS 2019, pp. 414-416, Tables WS 1.6-1.7).
  • In Côte d'Ivoire, 78% of the rural population was at high risk from E. coli at the water source, and 94% at high risk at home (DHS 2021, pp. 456-457, Tables 19.6.1-19.6.2).

• 47

  “It will be contaminated at the household level (even if it’s not contaminated at the source).” Notes from GiveWell call with Daniele Lantagne, November 1st, 2024 (unpublished).

• 48

  More context on the importance of mortality reduction to our cost-effectiveness analysis of water quality interventions is available here: “Based on the mechanisms we believe are most likely to mediate the relationship between water quality improvement and mortality reduction, we generate an estimate of mortality reduction that is the maximum we believe is biologically plausible in the beneficiary context. We use this to constrain mortality reduction estimates in our model.”

• 49Uduma does have some experience with chlorination in Benin, where it has recently started a program as part of a larger consortium where some of the infrastructure it manages will be chlorinated. However, we think there are limitations to how informative this experience will be for chlorination in Mali or Burkina Faso, as the infrastructure, and Uduma’s responsibilities, are different in Benin.
  For example, Uduma told us: “Many, if not most, of the systems in Benin already had chlorination when we took them over. The project is focused on rehabilitating, upgrading, and densifying existing systems rather than building new ones … As there has been a history of these systems operating, and as it was not required by the project, we have not undertaken any community engagement in Benin.” GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished).

• 50

  For example, we have heard from Daniele Lantagne, an expert on chlorination, about the need for modifications according to the water infrastructure.

• 51

  For this pilot, Uduma plans to test tablet erosion chlorinators, similar to those GiveWell has funded in Malawi and India, and mechanical liquid injection chlorinators, which we have not previously funded at scale (more here). However, it’s still to be determined which specific brands or models of these chlorinators Uduma will use.

• 52

  GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished).

• 53

  The Aquaya Institute, "Evaluating the effectiveness of chlorination by an innovative commercial rural water supplier in West Africa," November 2024, p. 3.

• 54

  The Aquaya Institute, "Evaluating the effectiveness of chlorination by an innovative commercial rural water supplier in West Africa," November 2024, p. 4.

• 55

  The Aquaya Institute, "Evaluating the effectiveness of chlorination by an innovative commercial rural water supplier in West Africa," November 2024, p. 3.

• 56

  The Aquaya Institute, "Evaluating the effectiveness of chlorination by an innovative commercial rural water supplier in West Africa," November 2024, p. 4.

• 57

  For example, chlorination expert Daniele Lantagne told us that she would expect there to be resistance to the taste of chlorine in Mali, in part because Population Services International (PSI) has not done significant work marketing its household-level chlorination products (e.g. WaterGuard) in Mali, as it has in a number of other countries in sub-Saharan Africa (GiveWell, notes from call with Daniele Lantagne, November 1, 2024 (unpublished)).

• 58

  Uduma, email to GiveWell, November 25, 2024 (unpublished).

• 59

  We have made other in-line chlorination grants, but only for research purposes (see our grant to the Development Innovation Lab at the University of Chicago to conduct preparations for an RCT of water quality interventions) or for scoping (such as this grant to the Abdul Latif Jameel Poverty Action Lab, for which the grant page publication is forthcoming).

• 60

  For example, Daniele Lantagne mentioned this to us during a conversation on March 19, 2024 (unpublished). Evidence Action currently implements chlorination programs at scale in Kenya, Uganda, and Malawi.

• 61

  Evidence Action did test some liquid injection chlorinators as part of their multi-village in-line chlorination pilot in India, but deprioritized further work on them because the large scaling-up grant that was to follow involved only smaller, single-village schemes, to which tablet-based dosers are better suited, and because of operational issues with injection chlorinators.

• 62

  For a ranked list of the promisingness of water chlorination by country (Mali is 5th), which formed part of our cost-effectiveness for chlorine vouchers, see here.

• 63

  Uduma, email to GiveWell, November 25, 2024 (unpublished).

• 64

  This means that if we were to decide to stop providing funding, the programs we’ve supported would either require other sources of philanthropic support, or they would cease to operate.

• 65

  Another GiveWell staff member who was part of this investigation estimates that the cost-effectiveness of this pilot is 18x, which is above our bar (see our breakdown in the Appendix).

• 66

  This benchmark is based on ‘moral weights’, a system we use to quantify the benefits of different impacts (e.g. increased income vs reduced deaths). We benchmark to a value of 1, which we define as the value of doubling someone’s consumption for one year. Our estimate of the “units of value” generated by direct cash transfers is 0.00335 per $. For more on how we use moral weights, see this page.

• 67

  See this section of our cost-effectiveness analysis for the specific adjustments that model the information value of this pilot.

• 68

  See this row in our cost-effectiveness analysis.

• 69

  As we discuss above, Uduma does not have prior experience of chlorinating water infrastructure in Mali or Burkina Faso, although it does have limited experience with chlorination in Benin (for more on Uduma’s work in Benin, see here.)

• 70

  For example, we have heard this from Amy Pickering and Daniele Lantagne, both of whom we regard as in-line chlorination experts.

• 71

  Daniele Lantagne, email to GiveWell, November 20, 2024 (unpublished).

• 72

  Uduma told us that it categorizes all of its systems in Mali in the “medium” security risk category, meaning the systems are “challenging to access, especially by international staff, but safe to operate”. GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished).

• 73

  “National government sets a maximum threshold for tariffs (CFA 500 per m³ [of water], or USD 0.81), and then Uduma bids on or negotiates a tariff with local government (Burkina Faso and Mali) or national government (Côte d’Ivoire) for each contract which is lower than this threshold. As you would expect, most are very close to that CFA 500 threshold. We do not have the option to increase the tariff unilaterally.” GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished).

• 74

  For example, we’ve seen local contracts for Defina in Mali, and Karangasso Vigue in Burkina Faso (unpublished).

• 75

  Another factor that helped assuage our concerns about funding a for-profit utility company was that Uduma has told us they “have engaged with funders who have put in place specific requirements within our grant agreements that would avoid Uduma making a profit from these grants”, such as “funding from US and other international foundations, with financing agreements that included terms stipulating specifically what the funding could be used for, and ensuring shareholders cannot profit from that grant funding”. GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished).

• 76

  Limitations to what we’ll learn:

  • We will get some signal about the possibility of future carbon credit financing, but Uduma expects this to take three years: “we believe we need at least 3 years of experience before we can conclude that [we] could fully cover our chlorination OpEx costs through carbon credits.” (GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished)). For more on our uncertainty around this, and on why we still think Uduma will be able to cover some of its operating costs with sources of revenue beyond GiveWell, see above.
  • We will only learn about pilot costs, which might not be an accurate guide to at-scale costs.
  • Our assumptions about leverage will remain guesses.
  • We won't learn how much future M&E will cost. Our projected M&E costs for the at-scale intervention are lower per year than for the 15-month pilot; this is based on the assumption that the research necessary for the pilot will be more exhaustive than the lighter-touch model that will likely suffice thereafter, and that we will work with a local, less expensive M&E partner in the future. Both are guesses which the pilot is unlikely to help us learn more about.
  • We're not yet sure what method Aquaya will use to corroborate Uduma's customer numbers, nor how accurate it will be. Aquaya does not yet know what methodology it will use to estimate this, but has suggested using a combination of Uduma’s metering data, household surveys based on random sampling, or geographic data techniques (see Aquaya’s concept note for this grant).

• 77

  The total number of people Uduma is expecting to reach through these projects is potentially high (~1.5m per year in total), but we expect this will overestimate RFMF because:

  • We don’t know if chlorination in any of these contexts would meet our cost-effectiveness bar for funding. On average, we would guess that cost-effectiveness in these countries would be lower than in Burkina Faso or Mali, which we estimate are two of the highest-impact locations for chlorination programs based on rates of under-5 mortality and percentage of mortality from enteric infection (for more, see out estimates of the impactfulness of chlorination programs in different locations here).
  • Some of the water Uduma will be managing in these additional countries will already be chlorinated (for example, it estimates that this will be 50% of the water for Togo).
  • It is not guaranteed that all these projects will go ahead.

  GiveWell, Uduma – running Q&A document for chlorine pilot (unpublished).