Where BESS models usually go wrong
A storage financial model has maybe two hundred inputs and perhaps eight that decide the outcome. Renewable energy teams coming from solar and wind are well equipped for site, interconnection and construction risk, and less equipped for the ones that are specific to electrochemistry.
The recurring gaps look like this.
Degradation is treated as a straight line. Cell ageing is rarely linear, and the knee, where fade accelerates, tends to sit inside a twenty-year model horizon. If the guaranteed capacity curve and the degradation assumption in the model came from different places, one of them is wrong.
Augmentation appears as a single number in year eight. In practice it's a schedule, and it depends on the augmentation strategy, cell availability at that future date, enclosure headroom, and whether the original supplier still exists. A model with augmentation capex but no augmentation plan is not a model.
Round trip efficiency is quoted without a boundary. DC-to-DC, AC-to-AC and metered-at-the-point-of-interconnection produce very different numbers. Auxiliary loads, mainly thermal management, are real and seasonal, and they eat into arbitrage spread every hour of every day.
Cycle limits sit in the warranty, not the model. Most warranties cap throughput, or cycles per year, or both, alongside temperature and state of charge operating windows. An optimiser dispatching for maximum revenue can void the warranty it was supposed to protect.
The revenue stack assumes today's prices hold. Ancillary service markets saturate. Frequency response revenues in several European markets fell sharply as capacity arrived. A project underwritten on a single high-value service is a merchant bet.
The assessment checklist
| Area | The question worth asking | What a weak answer looks like |
|---|---|---|
| Technology | Which cell, which chemistry, which format, and what field data exists for it? | "Tier 1 supplier", no cell model named |
| Degradation | What curve, from what test conditions, at what temperature and duty cycle? | A single percentage per year |
| Warranty | What are the exclusions, throughput caps, operating windows and measurement method? | The headline "10-year, 70% capacity" with no detail |
| Augmentation | What is the schedule, the space allowance and the cell availability assumption? | A lump sum in a later year |
| Efficiency | At which boundary is round trip efficiency measured, and are auxiliary loads included? | A percentage with no boundary stated |
| Availability | How is availability defined, what liquidated damages apply, and what are the caps? | Guarantee with uncapped optimism |
| Revenue | Which services, over what term, with what contracted versus merchant split? | Ancillary revenue held flat for ten years |
| Safety and permitting | Which standards, what UL 9540A test data, what does the local authority require? | "Compliant" with no test report |
| Counterparty | Who carries the performance risk if the integrator fails, and is it bankable? | Warranty from a company younger than the warranty |
| Grid | Interconnection status, queue position, curtailment exposure, export limits | "In the queue" |
The four questions that move IRR most
Ask these before anything else.
What happens to the return if the degradation curve is 20% worse than modelled?
If the answer changes the investment decision, the degradation assumption needs third-party validation rather than a supplier datasheet.
What does the warranty actually guarantee, and under which operating envelope?
Read the exclusions before the headline. Capacity guarantees measured under conditions your dispatch strategy will never hit are worth less than they look.
If the highest-value revenue stream halves, does the project still clear its hurdle rate?
Storage revenue stacks change faster than storage assets depreciate.
Who is standing behind the performance guarantee in year twelve?
Cell supplier, integrator, EPC and O&M provider carry different pieces, and the chain is only as strong as its shortest-lived member.
Why technical fluency changes the negotiation
The difference between a team that can and can't interrogate these points shows up in the term sheet, not the technical report.
A team that understands why a C-rate limit exists can negotiate a wider operating window and gain dispatch flexibility. A team that understands how throughput caps interact with an arbitrage strategy can price that constraint rather than accepting it. A team that knows what UL 9540A large-scale fire testing does and doesn't demonstrate can hold a supplier to the right evidence and answer the local fire authority without a consultant.
None of this requires an electrochemistry degree. It requires enough working knowledge to ask a second question after the first answer.
Building the capability in-house
Consultants are the right answer for one-off transactions. For a pipeline, the economics reverse: a single project's advisory fee often exceeds the cost of training the deal team.
The practical route for most renewable energy firms is a mixed group, development, technical, commercial and finance, going through the same programme together so they leave with a shared vocabulary. The value is as much in the alignment as in the content. Deals slow down when the technical lead and the finance lead mean different things by "degradation" or "availability".
How BatteryMBA fits
BatteryMBA is a 12-week CPD-accredited live online programme run by Battery Associates. It covers the battery value chain end to end: raw materials and cell chemistry, manufacturing, integration and BESS, EV applications, second life and recycling, plus markets, policy, investment and business models.
For project assessment work, the relevant parts are cell technology at working depth, degradation and lifetime, BESS integration and safety, and the commercial and regulatory context that determines what a project can earn. Lectures are taught by practitioners, and past lecturers have included specialists in BESS advisory, power conversion, and battery finance.
The format is built for people with delivery deadlines: 10 live lectures, weekly office hours across multiple time zones, recordings for what you miss, and 2 to 3 hours a week of commitment, or 4 to 5 with the optional case study track.
C18 runs 14 September to 5 December 2026. Tuition is €2,900 per person, with group and custom options for teams of five or more, including dedicated company cohorts.
Enrol in C18 · Team and company options · Book a 15-minute overview call
Informational and educational content only. Not professional, financial, legal, or engineering advice.