Project Everest

Adopted Experiment

[Experiment Adopted] Buka Stove Sales Fiji December 2018

Lean Phase: Solution

Assumption: Given our current risk prevention methods, the default rate on buka stove loans over 4 weeks is around 15%. This default rate must be monitored in order to understand how to further tighten our security requirements.   

Time Box: 4 weeks.

Success Metric:

This experiment is measuring the percentage of borrowers who default on payments. This will be tested with 20 sales of the buka stove in conjunction with the Fuel team in December.

Part of this experiment that is also crucial is setting up the systems to make this data collection usable and easy (see experiment build for this).

Green Light- If assumption is proven, next move is to understand why the 15% defaulted, and how to prevent this default in future times.

Success point - Success if 15% or lower default on loan.

Orange Light- To optimise, look at sales pitch and ways to analyse information.

Failure Point - Failure if greater than 40% of participants default on loan.

Red Light- Failure Protocol is to follow the risk protocol to understand why certain individuals defaulted, and thus understand how to better the current risk prevention methods.

Experiment build:

  1. Set up a CRM that can track repayment data effectively. 

  2. Establish a risk protocol in order to have a plan in place if someone were to default on a loan.

  3. Establish current risk prevention methods to use for this experiment. These can include;

    1. a. Using a guarantor. 

    2. b. Using an appropriate loan structure

    3. c. Using an appropriate method of payment technology

  4. Set up MoU’s/contracts for loan provision.

  5. Provide at least 20 individuals with credit for a buka stove, in conjunction with the fuel team.

  6. When making sales, take down the relevant data as per the CRM, and analyse this data over 4 weeks.

  7. After the 4 week repayment, determine if success or failure.


Tagged users
edited on 29th November 2018, 22:11 by Alex Staples

Alex Staples 8 months ago

Status label added: Experiment adopted

Reply 0

Alexander Teicher 8 months ago

Love the idea of continuing with using the Buka sales as a data point. Could it be worthwhile to look at AB testing 4 week repayments compared to 8 weeks as over summer that is a possibility. This might help to understand whether the timeframe of 4 weeks is too short due to variable income, or if time isn't the biggest issue and maybe money management is.

I feel like there is the opportunity to get more information out of this experiment, perhaps changing interest rates by adding increased upfront or end payment. At the same time though you obviously want to be careful to not change too many variables.

Keen to see how these loans are going (and how the stoves are going too).

Reply 0

Rose Martin 8 months ago

Hello teich!!!

This is a really good point. The December teams are also running a 'loan structures' experiment alongside this, which looks not only at the timeframe of the loan, but other elements of the way it is structured in order to provide the best option for Buka stove customers. The info from this experiment will hopefully be enough to understand whether 4 week, 8 week, 12 week, or however many week payments are best for Buka stove customers. If you want to keep up with this, here is the experiment for it .

Hope u are well and enjoying Timor :) remember, no bad days

Reply 1

Alexander Teicher 8 months ago

Amazing, just read through the word doc in that post and you have thought of everything! Left a comment on that post with a couple thoughts.

Timor is brilliant, and all good days someone once told me xxx

Reply 1

View all replies (2)