Discovery questions

[olafveerman]

Hi @katrinskaya and @mphillipsea . The discovery is well under way and we have a couple of questions around the content of the dashboards for your team.

1. From what date should we build the charts? The first dispensers were installed in 2008, but the installations only seem to take off somewhere in late 2010.
   Eliminating 2008 and 2009 will make for better charts.
2. How many liters of water can one chlorine refill purify? Is ‘total litres of water purified’ an interesting indicator?
3. How do you determine when to refill a dispenser? Is there a fixed schedule, is it always triggered by a request from a promoter? Do local promoters keep a stock?
4. We’d like to explore the links between promoter engagement and adoption rates. Do you think there's there's any significance in this? Can it be measured?

[bchen10]

@katrinskaya or @mphillipsea When you get a chance, could you respond to these questions as well? Thanks!

[katrinskaya]

1. Evidence Action took over the project officially (and was formed as an organization) in 2014. We should actually start the dashboard July 1, 2014 which is when you start seeing the growth as well. That is the official 'take over' date from IPA, the previous org that started Dispensers, to Evidence Action.
2. Yes, it may be an interesting indicator even though it's not in our KPIs per se. 3 ml of chlorine solution purifies 20 l of water. One chlorine bottle contains 5 litres.
3. Local promoters keep a small stock of about 4-5 chlorine bottles. They are on a delivery schedule but also call when the stick is low. Deliveries are made once a month or upon request.
4. Yes, our MLIS team actually has those numbers - though they may not be in Progmis. The team ran some randomized controlled trials AFAIK - which showed that with more engaged promoters adoption rates are up. Previously we said:

Adoption Booster #1: Promoters

One of the primary components of Dispensers for Safe Water is the election of a local community ‘promoter’ who is responsible for marketing the dispenser in the village, reporting any problems, and refilling the dispenser with chlorine. In return, the promoter gets a small stipend for communication purposes and a team T-shirt. Promoters are critical lynchpins in a community actually using dispensers, it turns out. Whether or not the promoter uses the chlorine dispenser in his or her own household is one of the single biggest predictors of adoption in a community, resulting in as much as a 17% increase in adoption over communities whose promoter does not use chlorine.

WHAT MAKES FOR A SUCCESSFUL PROMOTER?

We looked closely at a number of characteristics relevant to choosing a promoter, including - but not limited to - gender, age, distance to water points, education, and household characteristics. According to adoption rates, promoters aged 18-50 years are correlated with a 9% increase in community adoption relative to those aged 50 years and above. For every additional year in age, adoption decreases by 0.3%. In addition, our most successful promoters live near the dispenser,and have at least one child under age five at home. Where promoters lived more than 15 minutes round trip from the water source, communities were 10% less likely to test positive for chlorine.

Who communities choose as promoters is one thing, but having high-performing promoters another. We know that high quality initial training on the steps involved in using dispensers increases adoption (by up to 16%). We would love to know more about how to increase adoption at later stages. To that end, we have conducted a number of small randomized control trials looking at, for example, offering incentives and prizes to promoters and regularly calling to remind them of their role. These small “nudges” don’t seem to drive up adoption in qualitatively important ways.

We have noticed an increase in promoter fatigue. Currently we have only qualitative information on this - so this is to be taken with lots of caveats until we have more data. However, we may consider having new election of promoters, refreshing their supplies, or other strategies to rekindle enthusiasm as our program matures.

[katrinskaya]

@lndiku - do you have more info on whether we have more recent insights into this? It probably does get a bit too geeky for the average user of the dashboard :)

[lndiku]

Hi @katrinskaya no I do not have any more information as regards the link between promoter engagement and adoption rates. I do know that we carry out continuous evaluation of promoters and use the data to determine refill rates, promoter adoption, Promoter water collection practices, knowledge retention, Issues with the dispenser as reported by them but there is no current analysis that we share frequently as yet.

[katrinskaya]

Great, thank you @lndiku. @olafveerman, while this is super interesting (and we are interested and testing lots of things, as Leah said) for the purposes of the dashboard might be too geeky and not specific enough. We need easily understandable KPIs - this one is a few levels down in terms of understandability for a layperson.

[olafveerman]

Thanks for the background info @katrinskaya and @lndiku We'll leave the topic of promoter engagement and adoption rate for another phase.

[katrinskaya]

Just to be clear - adoption rate is a primary KPI so that is important to reflect. The relationship between it and promoter engagement is TBD even if we know a lot. Adoption rate is simply the rate of chlorination in household drinking water as measured brougham randomized testing of household water every month. It is a critical measure for us. Happy to explain further if that is not clear.

[katrinskaya]

Measured through randomized testing. Mobile typing is hard.

[bchen10]

Hi @lndiku

We have another question that needs clarifying here. How are you guys getting "Total People Served" numbers here? We're trying to crunch the numbers and seem to be getting numbers slightly (off by a little) different the total people served column. We're wondering if your'e factoring something else we didn't see here?

[bchen10]

@lndiku Nevermind. Disregard the question up top. We got it. Thanks!

[bchen10]

Discovery doc. comments followup questions:

@katrinskaya and @lndiku , we have some followup questions for you regarding some of the comments in the docs.

Missing entirely is our carbon generation. I know that it’s not tracked in Progmis so that is probably why is got dropped but it’s a key KPI. We are tracking it in excel so could be easily a fusion table.

• We would like to include some of the carbon generation dataset in this round of wireframes, would you be able to provide that for us by Thursday?
• Also, could you clarify what areas from the carbon generation dataset you would advise us to use?