According to an African Mini-grid Developers Association (AMDA) report, 40/42 mini-grid developers aren’t operationally profitable. But that’s not to say we don’t believe in mini-grids - we do - we just think they have a specific niche that they’re ideal for. In my role as Customer Success, my main mandate is to support Okra’s customers in designing, implementing and operating profitable Okra mesh-grids.
The AMDA report cited Average Revenue Per User (ARPU) and end-user energy demand as the most prominent barriers to operational profitability. ARPU in mini-grids sampled has shrunk by 51% from 2016 to 2020 to an average of $4.44 per month. Initial energy demand from end-users was much lower than originally anticipated. You can put two and two together: lower initial demand and reducing ARPU almost definitely spells commercially unviable deployments. It can also be expected that depreciating local currencies have factored into this lower ARPU. Due to high capex costs, it’s become clear that mini-grids are more suited to large, peri-urban communities with higher commercial activities than the majority of last-mile communities.
With mesh-grid technology we have strived to keep cost per connection lower than mini-grids since we launched the technology in 2020. But last-mile energy access is challenging: even with standard Okra Kits, we found that our customers were struggling to achieve project break-evens of less than 10 years - which is a rough target for attracting traditional investors.
As we observed the data from 8 of Okra’s customers in 3 countries, we found that similar to mini-grid projects, many initial mesh-grids were oversized for community needs. By analyzing the consumption, load and payment data, it became clear to us and our partners that:
- Most revenues, even in small communities, are driven by productive users using power for revenue-generating activities (1,000Wh - 3,000Wh/day) and;
- Other residential users typically only use energy for around 4 lights and a fan and grow their energy usage over time (100-300Wh/day).
The chart above shows aggregate data from our projects globally with more than 3,000 households in the dataset. The top 5% of users are accounting for more revenue than the highest group (69%) of users. It’s clear that a large chunk of the revenue is coming from a small number of consumers. These findings aren’t revolutionary - most mini-grid developers have known this for a long time! The difficulty is in knowing where and how many productive users you will have because that’s what determines sizing and network design.
Luckily, at Okra our philosophy has been to design mesh-grids for this exact scenario, unpredictable community needs. In this blog, we will be taking a look at the high-level design along with a deep dive into the impact of lower upfront cost and increased utilization on off-grid energy project economics and payback periods.
What and how?
The findings in the previous section are based on the following typical last-mile community that we’ve seen in Nigeria:
- Almost 100% of end-users are farmers of some kind (even if it’s subsistence farming, and as a result energy usage can fluctuate seasonally
- There is commercial activity in communities - in the form of stores, community milling, salons, small restaurants and tailors. Roughly 20 - 30% of users fall into this category and are “productive”
- A number of users already have diesel generators before the community becomes energized, roughly 5 - 10% of households. They have an existing spend between $20 - $40 USD / mo for fuel.
This is based on the first-hand experience of our service delivery teams on site, and also recently with surveys we completed for a project we are doing with MECS - more on this later.
Our sensitivity analysis has shown that the electricity sale price (tariff per kWh) is rather price inelastic - we ran a trial experiment where the tariff was dropped by 50% in a community and it resulted in less than a 5% increase in energy consumption.
Therefore, in order to increase the utilization and thus revenue generation of a single system we had to get creative on targeting the high-consuming users, and sizing assets appropriately to meet their needs.
Introducing Hub & Spoke
Increasing the utilization of a given set of energy assets was core to our thinking of improving Mesh-grid project economics. This is why we leverage the Hub & Spoke model, using a single Okra L90 Kit capable of producing an average of 3kWh per day and using it to power a cluster of households.
The approach was designed around productive use. The L90 Kit installed at a prosumer location prioritizes energy availability, thereby allowing customers to use appliances such as fridges and electric cooking.
The remaining energy would be split between nearby consumer locations. This ensures prosumers who are the most trustworthy payers to be serviced first, and allows nearby consumer households to use basic services such as lights, fans and phone charging at a low monthly ARPU, optimizing total project economics.
Prosumers, with Hubs, could then be connected with one another to share power in both directions, depending on each one’s generation and load profile.
Over time we then expect that all Spokes will become prosumer-level users, adding solar and battery capacity at each household and becoming Hubs, turning the network into an all-Hub mesh-grid.
The Hub & Spoke Financial Model
In order to properly understand this approach, rather than energizing an entire community from the get-go, the Hub & Spoke model enables productive users to be connected first along with other residential users near them. Demand can be built up over time with unconnected users encouraged to ditch their diesel generators or request appliances on finance - both of which would encourage more clusters to be built to bring them into the mesh-grid over time.
The above tables represent a simple payback model of a Hub & Spoke project with a 40% subsidy applied (similar to the UEF Subsidy in Nigeria). Along with a no-subsidy case, based on our latest supply chain prices, the deployed cost per connection using hub & spoke is $576.00 (including 4 X lights and a fan per connection).
A few key things to note regarding the financials:
1. A monthly ARPU of $7.5 is assumed. This is an average number for this project lifecycle and higher ARPUs are attainable by starting with productive users, then dynamically connecting residential households.
2. The capex cost per connection at $576.00 is at a 38.17% reduction vs. mini-grids according to the AMDA report, and these clusters can be deployed virtually anywhere as opposed to just large sites where mini-grids are most suitable for.
3. Running as an "energy-as-a-service" model, this has a simple payback period under a no-subsidy scenario of 8.73 years, which is roughly the time when batteries would need to be replaced. Under a subsidy scenario the payback period can be brought down to 5.24 years.
This model can be further optimized by using tools like Renewable Energy Credits, Key-Maker Model techniques or by providing additional services like the Internet-as-a-service.
Next Steps and Testing:
After the success of this model in Haiti, we have been rolling out this mesh-grid configuration with a number of our partners in Nigeria. We’ll start with small pilots with Creeds Energy, Engie and Haven Hill. For these projects, we will be with roughly the parameters on the model presented in the previous section:
Our engineering team is working on a hardware innovation which will further reduce the cost per cluster based on this model. It will put us well on the way to hitting sub-$500 per connection in the next two years.
Conclusion:
Our goal is for us to achieve 100% electrification by 2030. To reach this daring goal we need to rapidly ensure off-grid energy projects are commercially viable, hence the sub-$500 cost per connection target. By using this model for the mesh-grid, we’ve managed to get the forecast breakeven down to sub-9 years (with no subsidy) and a very attractive model when including subsidies - this is a great step for “energy as a service”. Taken together, this positions the mesh-grids as the go-to solution for the majority of communities, allowing us to mobilize private sector capital at scale.
We’re always on the lookout for further optimisations and will be publishing more findings in less than 6 months. If you find that anything is missing from this blog, or would like to discuss anything related to Hub & Spoke please shoot me an email at markus@okrasolar.com I’d love to talk to you more about this.
Alternatively, you can fill out the EOI form here.
26/03/24: Updated with small wording and diagram changes.
References:
https://drive.google.com/file/d/12h1FjykPePftQRLVOihtHQeowaPhToP1/view
https://africamda.org/wp-content/uploads/2022/06/Benchmarking-Africa-Minigrids-Report-2022-Key-Findings.pdf
https://africamda.org/wp-content/uploads/2023/02/AMDAs-Benchmarking-Africas-Minigrids-Report_2022.pdf
