Uganda has the third largest off-grid solar market in the world, and the last two years realized a 25% growth rate in the off-grid solar industry for Uganda. Off-grid solar solutions clearly contribute to improving access to power, but how can it be recognized and supported as a mainstream solution to improving electrification, especially in the new Energy Policy?
Beginners Guide to TikTok for Search - Rachel Pearson - We are Tilt __ Bright...
Insights from the Energy Ladder Research in Uganda
1. Insights from the
The Energy Ladder Research, Uganda
Richa Goyal
Schatz Energy Research Center
Closing the energy access gap through off-grid solutions
Solar 4 Uganda Thinkshop
April 26, 2018
2. Contents
1. Ways in which end-users adopt higher levels of
solar energy access
2. Motivations for purchase – In particular role of
flexible financing in increasing system
affordability
3. User preferences for subsequent purchases
3. Level 1: Solar portable lamps(SPL) owners
(226 respondents)
Level 2:
Small solar home system
kit owners (166
respondents)
Level 3: Large
solar home
system owners
(162
respondents)
Constructing the notional solar energy ladder in the sampling plan
Who were the respondents
4. Key insights
1. We reject the solar energy ladder hypothesis.
2. We observe energy stacking behavior.
3. Adoption rate:
• For subsequent purchases of solar energy products, impulse buying is low and planned buying is
high.
• Existing solar users that profess to purchase solar products constitute a promising upsell
opportunities pool.
• Lead conversion rate for subsequent purchases is high.
4. Fulfilling lighting energy demand by stacking solar lighting products, and buying solar components such
as panels and batteries to expand existing system capacity are popular solar product purchases.
5. Direct marketing, referrals and demonstration effect, play a big role in driving sales.
6. Flexible financing on solar products together with avoided energy spending on fuels do not lead to
rapid economic payback on solar home systems. Quality of life improvements are a greater driver than
economic payback for adoption of SHSs.
7. Users prefer to avoid flexible financing for subsequent purchases. Dealer stickiness for subsequent
purchases is significant.
6. Solar Energy Ladder?
Biomass based
fuels E.g. cow
dung,
firewood
Traditional fuels:
E.g. kerosene,
coal, charcoal
Modern energy:
E.g. LPG,
electricity
Original Energy Ladder
Energy stacking?
Diagram showing the conceptual solar energy ladder
The energy ladder hypothesis was
adapted into a hypothetical
construct of a ‘solar’ energy
ladder
7. Respondents that bought
their first product in 2015
Total
respondents
%
Level 1 221 226 98%
Level 2 163 166 98%
Level 3 150 162 93%
Solar energy ladder hypothesis does not hold!
Hypothesis violation 1: Users do not need to purchase smaller systems before buying big ones
Percentage of respondents that purchased a solar
product for the first time in the year 2015
Annual income in USD of respondents across
different levels of solar energy access
8. Within-solar product stacking – who bought what
Hypothesis violation 2: Users stack solar products and not substitute smaller ones with bigger ones
Level of solar
energy access
Overall actual
purchasers
Systems with
lights
Overall actual
adoption rate
Light adoption
rate
Components bought
to expand current
solar system
System expansion
rate
Level 1 44 40 22% 91% 0 0%
Level 2 26 15 18% 58% 6 23%
Level 3 22 9 15% 41% 3 14%
Total 92 64 18% 70% 9 10%
Popular solar energy products bought by subsequent purchasers (respondents that bought at least one other solar
energy product between baseline (May-Jun, 2015) and endline (Feb, 2017) surveys.
9. Individual trends lines for daily energy use in Wh across all solar systems owned by respondents
Note: Graph has been zoomed to optimize graphing area. Some Level 3 respondents
have daily energy service exceeding 600 Wh.
Solar off-grid energy adoption trends over time
10. Solar off-grid
technology
stacking behavior
Solar and non-solar technology
and fuel stacking
Deliberative stacking
This is often observed in the case of
cooking fuels. We do not observe
stacking of solar with non-solar
technologies as a result of a free
choice of user.
Constrained stacking
Users are constrained to use
technologies such as dry cell
torches, kerosene lamps and
candles albeit in reduced quantities
for energy security purposes post
purchase of solar systems, leading
to a limited ‘stacking’ behavior.
Within-solar stacking
This type of stacking is observed
as availability of a diverse range
of commercial off-grid solar
products allows users to fulfil
their energy demand in parts
Breaking down stacking solar and non-solar technologies and fuels among users
11. Solar with non-solar stacking
Respondents’ use of non-solar energy fuels and technologies
13. Key prior experiences or information that influenced users to purchase each of the solar products they own
Factors that influenced end-consumer purchase decisions
Note: Respondents could report more than one influencing factor.
14. Motivations for purchasing a solar home system as the first solar product
Reasons why Level 2 and Level 3 respondents bought a solar home system in their first product purchase
Note: Graph reports within level percentages. E.g. 66% Level 3 respondents bought a big system because they had demand for many
lights. Respondents could report more than one motivation.
15. 1
33
1
Graph shows monthly savings for typical customers of each level of solar energy access
Is motivation for purchase an economic one? - Net present value (NPV) analysis for typical customers
Level 1 typical users
Level 2 typical
users
Level 3 typical
users
16. Solar Product
Adopted in 2015
Level of solar
energy access
Solar scenario description
NPV over
two years in
2015 money
Break-even
period if less
than 2 years
CommentsPayment method for solar
product
Median monthly
savings due to
avoided spending
on status-quo fuels
Simple lantern
Level 1
Complete upfront payment of
$10
$2.76 $ 56.84 4 months
Almost 5.7x benefit
over two years
Single light with
mobile charging
Complete upfront payment of
$24.2
$2.76 $ 40.62 10 months
Almost 1.6x benefit
over two years
Small 3-light kit
(10 W solar
module)
Level 2
Complete upfront payment of
~$156
$3.61 -$102.16 N/A
There is net cash
outflow during the
analysis period (2
years).
PAYG payment plan: Down
payment: $18; Monthly
payment: ~$10; Repayment
period: 18 months
$3.61 -$127.85 N/A
Large 3-light SHS
with TV (50 W
solar module)
Level 3
Complete upfront payment of
~$546
$5.25 -$541.43 N/A
There is substantial
net cash outflow
during the analysis
period (2 years).
Micro-credit payment plan:
Down payment: ~$117;
Monthly payment: ~$28.50;
Repayment period: 24 months
$5.25 -$735.39 N/A
Results of a net present value analysis
Is motivation for purchase an economic one? - Net present value (NPV) analysis for typical customers cont.
18. Adoption rates
*Baseline pool: Respondents selected for analysis at baseline
†Planned purchasers: Respondents that planned to purchase solar product(s) at baseline
‡Planned unreachable at endline: Respondents that were unreachable at endline from within the pool that planned to purchase solar products(s) at baseline
⁰Planned-actual purchasers: Respondents within the pool that planned to purchase a product at baseline, that purchased a solar product between baseline and endline surveys
ꬹPlanned-actual adoption rate: Adoption rate from within the pool that said they planned to purchase at the time of baseline
ⱡTotal unreachable at endline: Total respondents that were unreachable at endline
꙳Overall actual purchasers: Total respondents that purchased solar product(s) between baseline and endline
Table and the figure represent how many respondents planned and actually purchased solar products between baseline (May-June, 2016) and endline surveys
(Feb, 2017)
Baseline pool*
Planned
purchasers†
Planned
adoption rate
Planned
unreachable at
endline‡
Respondents
at endline
Planned-actual
purchasers⁰
Planned-actual
adoption rateꬹ
Total unreachable
at endlineⱡ
Overall actual
purchasers꙳
Overall actual
adoption rate
Syste
wit
ligh
Level 1 226 169 75% 21 201 34 23% 25 44 22% 40
Level 2 166 150 90% 16 148 25 19% 18 26 18% 21
Level 3 162 123 76% 6 150 20 17% 12 22 15% 12
Total 554 442 80% 43 499 79 20% 55 92 18%
19. Users’ preferences for subsequent purchase of energy product(s): Summary
Key metrics Level 1 Level 2 Level 3
Preference for upfront payment using personal savings 98% 97% 95%
Dealer stickiness 24% 74% 73%
Preference to purchase from a different 'known’ dealer 9% 2% 4%
Preference for PAYG as financing method 0% 0% 0%
Preference to purchase non-solar energy products
or take utility grid connection
<1% <1% 0%
This Table summarizes some of the key highlights from the ‘subsequent purchase analysis’
20. Dealer stickiness
Level 1 Count %
Not sure 112 66%
SunnyMoney 41 24%
Unknown solar shop 5 3%
Electronics shop 4 2%
Fenix 3 2%
Solar Now 2 1%
Solar Solutions 1 1%
Pallisa Teacher's SACCO 1 1%
Utility grid 1 1%
Total 170 100%
Level 2 Count %
Fenix 111 74%
Not sure 36 24%
Unknown solar shop 1 1%
Solar Now 1 1%
Utility grid 1 1%
Total 150 100%
Level 3 Count %
SolarNow 90 73%
Not sure 28 23%
Soltec 1 1%
M-Kopa 1 1%
Sunshine Solar 1 1%
Construct solar systems 1 1%
Solar Solutions 1 1%
Total 123 100%
The tables shows the dealer with whom respondents that planned to make subsequent purchases at the time of baseline survey
(May-Jun, 2015)
21. Acknowledgements
• The research benefitted from the following organizations and individuals
• UNCDF CleanStart Team: Hee Sung Kim, Robin Gravesteijn & Vincent
Wierda
• Research advisory board: Kat Harrison, Acumen & Michael Nique, GSMA
• Fieldwork implementing agency: CIRCODU
24. Research methods and timeline
March,
2016
Stakeholder
feedback
process
Solar off-
grid product
distribution
chain
mapping
Baseline
surveys:
Phone
interviews
Baseline
surveys:
Face to face
interviews
Endline
surveys:
Phone
interviews
April,
2016
May-Jun,
2016
Jun-Jul,
2016
Feb,
2017
25. Retail footprint in districts where the study was implemented
Solar shops interviewed in Luwero district
Solar shops interviewed in Pallisa district
The red dots indicate the GPS location of solar retailers that were interviewed as part of distribution mapping for this study
26. Solar off-grid opportunity in Uganda
1. The grid electricity is erratic and of
poor quality
Collapsed electricity pole in farmland along the
Kampala-Tirinyi highway that crosses Pallisa
district
End-user perception on electricity tariff rates in
rural Uganda
3. Respondents showed a keen
understanding of the high cost of fuel-
based lighting
End-user savings after investing in solar energy
product(s)
2. High cost of grid electricity
27. Sample design: Original sample design strategy
The table details original plan for sampling - Total 700 respondents including 100 respondents to account for
attrition between baseline and endline surveys
Data partner Type of product
Level of solar
energy
access
Payment method
Baseline surveys -
Phone interviews
Baseline surveys –
Surveys for addressing
attrition between
baseline and endline
surveys
Baseline surveys –
Face to face
interviews‡
SunnyMoney
and
GreenLight Planet
Single light products without
mobile charging
Level 1 Cash or credit
150
70†
25
Single light product with mobile
charging
150 25
Fenix
International
Small plug and pay solar home
system kits
Level 2
Mobile money
based
pay-as-you-go
(PAYG)
150 15 20
SolarNow
Advanced component based
solar home systems
Level 3 Micro-credit 150 15 20
Total respondents 600 100 100
‡ Face to face interviews were conducted with a subset of phone interviewees within each respondent group.
†SunnyMoney sales records were not digitized. Further, since a significant portion of their sales occurred via solar agents, the sales records were
expected to be less accurate. Therefore, a attrition rate higher than other respondent groups was factored in for SunnyMoney respondents.
Endline phone surveys were planned to be conducted with the baseline respondents selected for analysis.
28. Sample design: Usable surveys for analysis
Respondents with duplicate records and other data entry errors were eliminated as follows.
During baseline surveys –
• 615 phone interviews were conducted, 15 of which had to be discarded.
• 117 face to face interviews were conducted, 3 of which had to be discarded.
During the endline surveys, 543 phone interviews were conducted, 2 of which had to be discarded.
The final number of usable surveys are as follows.
Data partner
Largest product respondents owned
at the time of baseline surveys
Baseline surveys -
Phone interviews
Baseline surveys –
Face to face interviews*
Endline surveys -
Phone interviews
SunnyMoney and
GreenLight Planet
Single light system without mobile
charging
134 21 119
Single light system with mobile
charging
92 31 81
Multi-light kit 37 7 37
Solar home system with lights and at
least one other appliance
8 2 5
Total SunnyMoney respondents 271 61 242
Fenix International
Multi-light kit with/without other
appliances
166 30 148
Solar home system greater than or
equal to 50W in panel size
1 - 1
Total Fenix respondents 167 30 149
SolarNow
Solar product was a solar home
system greater than or equal to 50W
162 23 150
Total SolarNow respondents 162 23 150
Total respondents 600 114 541
Level 1
Level 2
Level 3
29. Sample design: Final sample size selected for analysis
In the final surveys selected for analysis, respondents within SunnyMoney/GreenLight Planet group that had a
system with more than one light and the respondent within the Fenix group that had a product with panel size
>=50W were removed. This was done to achieve 3 respondent groups as follows:
Data partner
Largest product
respondents owned at
the time of baseline
surveys
Level of solar energy
access at the time of
baseline surveys
Payment method
Baseline surveys
- Phone
interviews
Baseline surveys
- Face to face
interviews*
Endline surveys -
Phone interviews
SunnyMoney
and GreenLight
Planet
Entry level single light
systems
Level 1 Cash or credit 226 52 200
Fenix
International
Plug and pay small solar
home system kits
Level 2
Mobile money
based
pay-as-you-go
(PAYG)
166 30 148
SolarNow
Advanced solar systems
with larger PV panels
(>=50 Wp)
Level 3 Micro-credit 162 23 150
Total respondents 554 105 498
30. Welch’s one-way analysis of means (not assuming equal variances)
Data: Annual_income_USD and Solar_level
F = 10.045, num df = 2.00, denom df = 220.38, p-value = 6.685e-05
Tukey multiple comparisons of means at 95% family-wise confidence level
Data: Annual_income_USD and Solar_level
Test results for differences in income between levels of solar energy access
The table shows test results for Welch’s one-way ANOVA and Tukey multiple comparisons of mean
incomes between the three solar levels
Welch’s ANOVA test for unequal variances shows that incomes are significantly different between the three levels. Further, the
Tukey multiple pairwise-comparison shows that difference between income means of level 1 and 2 is statistically insignificant,
while difference between income means of levels 1 and 3 and that of levels 2 and 3 is statistically significant.
31. Level
Median daily
energy service
Wh 2013
Median daily
energy service
Wh 2014
Median daily
energy service
Wh 2015
Median daily
energy service
Wh 2016
Median daily
energy service
Wh 2017
% change in daily
energy service Wh
from 2015 to 2016
Level 1 0 0 1.86 4.08 4.08 119%
Level 2 0 0 43.35 43.35 43.35 0%
Level 3 0 0 127.5 127.5 127.5 0%
Income quartile Level
Sample size
(n)
Median daily
energy
service Wh
2013
Median daily
energy
service Wh
2014
Median daily
energy
service Wh
2015
Median daily
energy
service Wh
2016
Median daily
energy
service Wh
2017
% change in daily
energy service Wh from
2015 to 2016
Income quartile 1
Level 1 48 0 0 1.86 4.08 4.08 119%
Level 2 40 0 0 25.5 31.08 31.08 22%
Level 3 27 0 0 127.5 127.5 127.5 0%
Income quartile 2
Level 1 58 0 0 1.86 1.86 1.86 0%
Level 2 29 0 0 25.5 26.47 26.47 4%
Level 3 23 0 0 127.5 127.5 127.5 0%
Income quartile 3
Level 1 46 0 0 1.86 5.01 5.01 169%
Level 2 30 0 0 43.35 43.35 43.35 0%
Level 3 31 0 0 127.5 127.5 127.5 0%
Income quartile 4
Level 1 37 0 0 1.86 4.08 4.08 119%
Level 2 24 0 0 43.35 43.35 43.35 0%
Level 3 55 0 0 127.5 127.5 127.5 0%
The table shows daily energy use in Wh for respondents across each level of solar energy access
Solar off-grid energy adoption trends over time – Available daily energy use measured in Watt Hours (Wh) cont.
The table shows median daily energy use in Wh for respondents across each level of solar energy access
and income quartiles
Key takeaways -
- Level 1 has more than 100% increase in
median daily energy service from year 2015
to 2016. This is because adoption rate is
higher for Level 1 respondents than Level 2
and 3. Level 2 and 3 respondents fulfil their
lighting demand efficiently with minimum
number of systems by investing in multi-light
kits. Level 1 respondents ‘stack’ solar lighting
systems to fulfil their lighting demand. See
section ‘Adoption rates’.
- In Levels 1 and 2, there is a noticeable
increase in daily energy service levels in
income quartiles 3 and 4 when compared
with income quartiles 1 and 2.
- The daily energy service levels for Level 2
and 3 are fairly consistent across the years.
Note: 106/554 respondents did not report income. These tables reflect median daily energy service levels only for those
respondents that reported income.
32. Savings in energy spending in status-quo fuels by shifting to solar off-grid energy products
The graphs show the savings in energy spending on status-quo fuels across respondents in the 3 levels of solar energy access
Key takeaways
An average rural Ugandan anywhere between 2-8 litres of kerosene a month.
Consumption of candles and dry cells is relatively high – 75% respondents use up to 80 candles and up to 16 dry cells a month.
33. Net present value (NPV) analysis for typical customers across all levels - Description of customer profile chosen for
analysis
We identified typical users of solar products within each solar energy level to perform a NPV analysis of the solar scenario as
opposed to the scenario where the user does not use any solar products. These users are described as follows -
Level 1
1.1 Lantern user - ~43% Level 1 users purchased a 0.5W lantern in that meets Lighting Global standards in 2015, and did not purchase a solar
product after.
1.2 Single light with mobile charging user - The median energy service derived from various solar products owned by Level 1 users
corresponds to the use of a 1.5W single light and mobile charging. This product meets Lighting Global standards. The users purchased the
light in 2015 and did not any other solar product after.
Level 2
2.1 Small 4-light kit user - ~36% Level 2 users purchased a 17W product with 4 lights, mobile charging and a radio with a built-in torch. They
purchased the product in 2015 and did not purchase any other solar product after.
2.2 Small 3-light kit user - ~22% Level 2 users purchased a 10W product with 3 lights, mobile charging and a radio with a built-in torch. They
purchased the product in 2015 and did not purchase any other solar product after.
Level 3
3.1 Large 3-light kit user - ~37% Level 3 users purchased a 50W product with 3 lights, mobile charging and a radio.
3.2 Large 3-light kit with TV user - ~31.5% Level 3 users purchased a 50W product with 3 lights, mobile charging, radio and an 18.5-inch TV.
In the table below, the median daily energy service (DES) in Wh available from the solar off-
grid products owned by each user are listed by year.
Year
Level 1 Level 2 Level 3
Lantern user
Single light with mobile
charging user
Small 4-light kit
user
Small 3-light kit
user
Large 3-light kit
user
Large 3-light kit
with TV user
DES until 2013 0 0 0 0 0 0
DES in 2014 0 0 0 0 0 0
DES in 2015 0.96 3.46 41.07 38.07 65.57 185.57
DES in 2016 0.96 3.46 41.07 38.07 65.57 185.57
DES in 2017 0.96 3.46 41.07 38.07 65.57 185.57
34. Note: Graph has been zoomed to optimize graphing area. Some outliers have been excluded. Some respondents have monthly
savings going up to $50.
Does ‘stacking’ single light systems have a better pay-off than purchasing a multi-light kit?
Key takeaways
₋ Savings from purchasing a single product
with many lights can be similar to savings
from ‘stacking’ several single light systems,
but, the cost of multi-light kits is
substantially higher. Therefore, motivation
for purchasing a multi-light kit is not an
economic decision, but has a more
aspirational character.
₋ Median monthly savings of ‘Single light
system stackers’ is 63% more than other
Level 1 users that use single light Level 1,
and is between 21-25% more than small
multi-light kit users at Level 2.
Graph shows monthly savings for typical customers of each level of solar energy access alongside users that stacked single light systems.
Single light system stackers are defined as Level 1 users that purchased multiple single light systems with or without mobile charging,
instead of purchasing a single product with more than one light.
35. Some users stack solar products
Tables shows solar product adoption rates and who bought what?
Table and the figure represent how many respondents planned and actually purchased solar products between baseline
(May-June, 2016) and endline surveys (Feb, 2017)
Baseline pool*
Planned
purchasers†
Planned
adoption rate
Planned
unreachable at
endline‡
Respondents
at endline
Planned-actual
purchasers⁰
Planned-actual
adoption rateꬹ
Total unreachable
at endlineⱡ
Overall actual
purchasers꙳
Overall actual
adoption rate
Syste
wit
ligh
Level 1 226 169 75% 21 201 34 23% 25 44 22% 40
Level 2 166 150 90% 16 148 25 19% 18 26 18% 21
Level 3 162 123 76% 6 150 20 17% 12 22 15% 12
Total 554 442 80% 43 499 79 20% 55 92 18%
Level of solar
energy access
Overall actual
purchasers*
Systems with
lights†
Overall actual
adoption rate
Light adoption
rate
Components bought
to expand current
solar system‡
System expansion
rate
Level 1 44 40 22% 91% 0 0%
Level 2 26 15 18% 58% 6 23%
Level 3 22 9 15% 41% 3 14%
Total 92 64 18% 70% 9 10%
The table shows the popular solar energy products bought by subsequent purchasers (respondents that bought at
least one other solar energy product between baseline (May-Jun, 2015) and endline (Feb, 2017) surveys.
36. Dealer stickiness
Level 1 Count %
Not sure 112 66%
SunnyMoney 41 24%
Unknown solar shop 5 3%
Electronics shop 4 2%
Fenix 3 2%
Solar Now 2 1%
Solar Solutions 1 1%
Pallisa Teacher's SACCO 1 1%
Utility grid 1 1%
Total 170 100%
Level 2 Count %
Fenix 111 74%
Not sure 36 24%
Unknown solar shop 1 1%
Solar Now 1 1%
Utility grid 1 1%
Total 150 100%
Level 3 Count %
SolarNow 90 73%
Not sure 28 23%
Soltec 1 1%
M-Kopa 1 1%
Sunshine Solar 1 1%
Construct solar systems 1 1%
Solar Solutions 1 1%
Total 123 100%
The tables shows the dealer with whom respondents that planned to make subsequent purchases at the time of baseline survey (May-Jun, 2015)
Key takeaways –
- Dealer stickiness is significant, >70%, for Level 2 and 3 respondents, but not because of the credit financing option these dealers provide.
97% of total respondents prefer to make subsequent purchases by paying for the full cost of system upfront using personal savings.
- Dealer stickiness rate for Fenix is 75%; 90% of Level 2 respondents planned to make a subsequent purchase at the time of baseline
survey – none preferred to use PAYG financing.
- Successful direct marketing has a role to play in high dealer stickiness: Dealer stickiness is high in part because of low proximity to solar
product dealers. Most dealers are located in central towns or highways, and dealers seek out customers through direct marketing.
- Significant upsell opportunities exist for dealers within the existing customer base.
37. Users' plans for subsequent purchase of energy product(s) for Level 1 respondents
Preferred dealer
Preferred financing method
Available sales method of preferred dealer
Preferences for subsequent planned purchases of respondents that planned on making further energy product purchases at
the time of baseline (May-Jun, 2016)
38. Users' plans for subsequent purchase of energy product(s) for Level 2 respondents
Preferences for subsequent planned purchases of respondents that planned on making further energy product purchases at the time of
baseline (May-Jun, 2016)
39. Users' plans for subsequent purchase of energy product(s) for Level 3 respondents
Preferences for subsequent planned purchases of respondents that planned on making further energy product purchases at
the time of baseline (May-Jun, 2016)
43. Relationship between mobile money transactions, income and level of solar energy access
The graphs show total transaction value (TXVA) by level of solar energy access and income quartiles in year 2015 and 2016
The above graph shows that mobile money
transactions increase with the level of solar energy
access. The graph on the right splits respondents by
income quartiles and year, thereby providing
additional detail about the relationship between
mobile money transactions, energy access levels, and
income.
Note: Graphs have been zoomed to optimize graphing area. Some outliers have been excluded.
Some respondents have TXVA exceeding $60,000.
44. Relationship between mobile money transactions, income, and level of solar energy access cont.
Graphs show Median Total transaction value (TXVA) by level of solar energy access and income quartiles in year 2015 and 2016