KPI Framework – Technical Guide

The KPI Framework seeks to improve information flow between investors and
companies in the PAYG industry by establishing standardized definitions and
reporting standards. The KPIs are not designed to replace an appropriate appraisal
process and an investment officer, however they will allow for a more structured
assessment and comparable values. However, given the highly variable – and
continuously changing business models of PAYG companies, the KPI framework is not
meant to be a scorecard of a company’s operational performance, but offer a more
structured way to continuously assess the performance of different parts of the company
and its customer base.

This Technical Guide seeks to provide an overview of each KPI along with a more
detailed explanation of each definition. With a handful of exceptions, each of the KPIs
is defined across a portfolio of PAYG solar assets, generally a cohort in time, contained
within a single country or region. Indicators are defined at a point in time, and measuring
the KPIs on a periodic basis (monthly or quarterly) will help to monitor performance over
time. It is also recommended that, if possible, distributions be provided for the high-
priority KPI/ratios of KPI (such as Average Revenue Per User (ARPU), Portfolio at Risk
(PAR) and unit cost), in addition to the summary statistics specified in the definitions.

Since digital technologies and sensors are intrinsically a part of PAYG business
models, the industry naturally lends itself to the generation and use of large,
heterogeneous, and fast- moving streams of information, otherwise known as “big
data.” Examples include customer consumption and payment data, customer credit
histories, and web server logs, as well as cell phone records and satellite imagery from
partners and third-party vendors. As concluded in the 'State of the Data' report, these data
sets are not necessarily captured in a structured way by PAYG companies today, due to
lack of priority, capacity, or feasibility. As discussed in a recent World Bank publication,
the tools for dealing with such data – from distributed computing to machine learning and
data mining—have evolved over the last decade to the point where they are readily usable
in commercial settings. Potential uses in the PAYGO industry include customer targeting
and marketing, credit assessment, collections, and market research. By establishing
standardized definitions and reporting standards, the KPI framework could complement
the broader vision of increasing the sector's sophistication in data and analytics.

Box 1: Example on how to compute KPIs

With the exception of EBITDA Breakeven, all of the shortlisted KPIs can be
computed from a simple ledger-style data extract, which includes all scheduled and
completed transactions for each customer, plus a table of product characteristics.
The ledger columns should include:

• Date
• Customer ID
• Receipt ID
• Amount
• Currency
• Expected/Completed (completed only for pay-per-use models)
• Type (deposit, return, monthly payment, repair, etc)
• Product type

For example, a new customer purchase of a financed product would result in a Completed
deposit entry and N expected monthly payment entries. Each month (ideally), a new
completed payment with the same Receipt ID would be entered into the ledger. This
structure makes it easy to calculate the KPIs across a wide range of business models.

Key Performance Indicators

1. Average Monthly Revenue per User (ARPU)

1.1. Definition: Total payments (including interest) received over the last 30 days divided
by portfolio size.

1.2. Formula: (Total revenue over most recent 30 days [Local Currency])/(Total # of
active [Units]).

1.3. Explanation of Definition: This definition is inclusive of all revenue, including

deposits and maintenance. Revenue is denominated in the relevant local currency; in the
rare case, where a portfolio includes revenue denominated in multiple currencies, the
dominant one should be used for reporting, with the other(s) converted using the present
day exchange rate.

1.4. Commentary: In this calculation and in all that follow, active users are defined as
those who have made at least some payment in the preceding 90 days.

2. Average Unit Cost

2.1. Definition: Mean of cost of active units, inclusive of hardware, international and in-
country transportation and installation, import taxes and stock insurance, but exclusive of
customer acquisition and maintenance

2.2. Formula: Sum of (Cost of hardware plus cost of transportation and installation on
active units [Local Currency]) / (# of active [Units]).
2.3. Explanation of Definition: This indicator quantifies the average unit cost to the
company, not to the consumer.

2.4. Commentary: Customer acquisition and maintenance costs were excluded of the
KPI composition due to the wide variability in marketing and service arrangements
among distributed energy service companies (DESCOs), which might lead to meaningful
inconsistencies in accounting.

3. Average Credit Period

3.1. Definition: Average nominal number of months between system acquisition and
expected final payment.

3.2. Formula: Sum of (Expected final payment date - system acquisition date [Months]) /
(Total # of active [Units]).

3.3. Explanation of Definition: Expected final payment is defined in the nominal sense,
under the assumption that no payments are skipped or postponed – even if such flexibility
is explicitly allowed in the contract.

3.4. Commentary: In the case of true pay-as-you-go business models, where there are no
“expected” obligatory payments per se, the contract end date should be used as the
expected final payment date.

4. Average Customer Deposit as a Proportion of Unit Cost

4.1. Definition: Average customer deposit as a fraction of total unit cost.

4.2. Formula: Mean over active units of (Deposit received [Local Currency])/(Cost of
hardware plus cost of transportation and installation [Local Currency].

4.3. Explanation of Definition: As in Average Unit Cost, the cost here is defined
inclusive of hardware, international and in-country transportation, and installation, but
exclusive of customer acquisition and maintenance. Deposit, or down payment, is defined
as the total amount paid at time of unit acquisition, exclusive of payment per use.

4.4. Commentary: A case where a portion of the payment remitted at time of unit
acquisition might be excluded from this definition of “deposit” includes, for example, a
purchase of some number of days or a payment of the first regular billing cycle.

5. Portfolio At Risk (PAR)

5.1. Definition: Total amount owed on units with any balance billed in the last 90 days
which is overdue by 30 or more days, divided by total amount owed by customers
(inclusive of interest and exclusive of any potential late payments).

5.2. Formula: Sum of (Total amount owed by customers with any balance billed in the
last 90 days which is overdue by 30 or more days) / Sum of (Total amount owed on
active units).

5.3. Explanation of Definition: This is a rolling definition of Portfolio At Risk, which is

“forgiving” of missed payments. 5.4. Commentary: A stricter definition of PAR, in
which any 30+-day overdue balance designates an asset as at-risk, would cause many
portfolios to appear >50% at-risk due to occasional missed payments. This definition
treats a customer who missed a payment six months ago but is now back on track as “not
at risk,” which many industry actors regard as appropriate.

6. Portfolio Size

6.1. Definition: Total number of customers in portfolio.

6.2. Formula: # of active [Units].

6.3. Explanation of Definition: As in the above definitions, customers who have made
no payments in the preceding 90 days are deemed to be inactive and are not included in
the portfolio size.

6.4. Commentary: Larger portfolios are expected to have more stable performance.

7. Churn

7.1. Definition: Fraction of units that have gone inactive over the previous 90 days.

7.2. Formula: Sum of (# of units on which no payment was made in the preceding 90
days) / (# of active [Units] as calculated 90 days ago).

7.3. Explanation of Definition: “Churn” here is used to denote customers dropping out
of the portfolio.

7.4. Commentary: This indicator was chosen in place of percent write-off due to
widespread variability in standards and practices surrounding write-offs.

8. Standard Deviation of Amount Ahead/Behind on Payments

8.1. Definition: Total revenue received minus total amount expected up to present date,
in local currency, divided by number of active units.

8.2. Formula: Mean/Standard Deviation over active units of (Total revenue received
from unit acquisition to present, exclusive of unscheduled maintenance - Total revenue
expected from unit acquisition to present [Local Currency]).

8.3. Explanation of Definition: For true pay-as-you-go business models, a customer who
has paid for credits but not yet used them would count as “ahead,” while a customer who
has used days for which (s)he has not paid would count as “behind.” For asset finance
business models, a customer who has prepaid their regular payments would count as
“ahead” and a customer who has missed or is late on a payment would count as “behind.”
8.4. Commentary: Due to the important differences between hardware and business
models, these KPI are primarily useful for monitoring a portfolio’s performance over
time or for cross-comparing the portfolios of two DESCOs with similar business models.
They are extremely valuable for these purposes, but should be used carefully – if at all –
for comparing performance across a wide range of portfolios.

9. Average Total Expected Revenue

9.1. Definition: Mean total anticipated payments, including deposit but excluding
unscheduled maintenance and late fees, in local currency.

9.2. Formula: Sum of (Total expected payments from active units, excluding any
unscheduled maintenance [Local Currency]) / (# of active [Units]).

9.3. Explanation of Definition: Taken together with Average Unit Cost, this KPI bears
on the portfolio’s anticipated profit margin.

9.4. Commentary: Does not apply to true pay-as-you-go business models.

10. EBITDA Breakeven

10.1. Definition: EBITDA breakeven yes/no for the company as a whole.

10.2. Explanation of Definition: Explicitly, for companies providing services in addition

to distributed solar generation, EBITDA breakeven is still evaluated for the company as a
whole rather than for the distributed solar division.

10.3. Commentary: This KPI was highlighted as important for a variety of reasons.
Among them is portfolio maintenance, due to the risk of unprofitable DESCOs eventually
becoming unable to effectively service their equipment and/or collect their payments.

11. Average Maintenance Cost

11.1. Definition: Total payments received in local currency for scheduled and
unscheduled maintenance, over the last 30 days.

11.2. Formula: (Total maintenance payments received over most recent 30 days [Local
Currency])/(Total # of active [Units]).

11.3. Explanation of Definition: Maintenance includes both scheduled and unscheduled

service, whether provided by the DESCO itself or by a contractor. 11.4. Commentary:
This indicator was identified as an important proxy for system reliability and technology
disruption risk.

12. Standards Compliance

12.1. Definition: Percentage of systems complying with Lighting Global quality

standards.
12.2. Formula: (Total # of active units complying with LG standards)/(Total # of active
units).

12.3.Explanation of Definition: Standards are specified here:

https://www.lightingglobal.org/qa/standards/.

12.4. Commentary: This indicator was highlighted as an important indicator for both
operator and technology disruption risk, as well as for regulatory risk.

13. FX Exposure: Net Open Position as a Percentage of Equity

13.1. Definition: Net open position divided by equity, calculated on absolute value
basis.

13.2. Formula: (Assets - (Liabilities + Equity))/(Equity in Local Currency).

13.3. Explanation of Definition: In the rare case where a portfolio includes assets paid in
multiple local currencies, this KPI should be calculated separately for each currency.

13.4. Commentary: During consultations, stakeholders highlighted currency risks as an

important factor to consider. This is the most widely-used metric for FX exposure in
microfinance, and has also been highlighted by several IFIs as a critical indicator. The
team behind ‘The Currency Exchange Fund’ has more insights on how to manage FX
exposure.