How to Run a Solar Tender: Writing an RFP That Gets Comparable Bids

A solar tender that works produces bids that sit reasonably close together and differ for reasons you can evaluate: one bidder chose a different inverter arrangement, another proposed more storage. A tender that does not work produces a spread where the lowest and highest bids are not pricing the same system, and the evaluation committee ends up comparing guesses.

We read solar tenders from the other side of the table, as a bidder, and the spread is mostly created by the document rather than by the bidders. Bidders price what they are told and price uncertainty as risk, so every gap in a tender becomes a contingency in one bid or an omission in another.

This article sets out how to structure a solar RFP so that bids come back comparable: the procurement model to settle first, what the documents need to contain, the mistakes that scatter prices, and how to evaluate what comes back.

Choose the Procurement Model First

The first decision is what you are buying, because it decides who carries which risk and therefore what bidders have to price.

Supply only. You buy equipment against a specification and arrange installation yourself. It is cheapest on paper, and every interface risk stays with you: if the modules, inverters and batteries do not work together, nobody else is responsible. It suits owners with a capable installation team of their own.

Supply and install. The contractor supplies and installs to your design. Design responsibility stays with you or your consultant, so if the design undersizes the battery, the contractor has built exactly what was asked for.

EPC turnkey. The contractor designs, supplies, installs and commissions against your performance requirements, and carries the design risk. Most commercial solar procurement ends up here, and it is where comparable bids depend most on the quality of the requirements. The FIDIC EPC/Turnkey conditions, known as the Silver Book, are the widely used international form for this allocation of risk. FIDIC itself says they are not suitable where tenderers have too little time to scrutinise the employer's requirements, or where the employer intends to supervise the work closely, and points those projects to the Yellow Book for plant and design-build instead.

EPC plus O&M. The same, with an operations and maintenance term attached, usually with an availability or performance guarantee. It ties the builder's incentives to long-term output, and it makes bids harder to compare unless the O&M scope is specified as tightly as the build. Our O&M contract guide sets out what that scope should contain.

PPA or lease. You buy electricity or rent the system instead of buying an asset, and the developer carries the capital, performance and usually the maintenance risk. The tender becomes a tariff competition, and the terms behind the tariff matter as much as the tariff itself. Our guide to solar PPAs covers what to check.

Mixing models inside one tender is a common source of incomparable bids: asking for a turnkey price with full design responsibility against a detailed design the bidder is also told to follow, for example. Pick one model and write every document to match it.

The five procurement models and the risk each one moves. Pick one and write every tender document to match it. Source: MIMAH engineering practice; EPC/Turnkey risk allocation after the FIDIC Silver Book.

What a Good Solar Tender Contains

A tender that produces comparable bids usually contains eight things.

Instructions to bidders. Submission deadline and format, bid validity, currency, required forms, the clarification route and its deadline, and the site visit arrangements. It should also say which deviations make a bid non-compliant, so bidders know what is fatal and what is merely scored down.

A technical specification that states performance, not brands. Required standards, minimum module and inverter efficiency, battery usable capacity and cycle life at a stated depth of discharge, operating temperature range, dust and ingress protection, and monitoring requirements. If a brand has to be named for compatibility with existing equipment, say why, and add "or equivalent" with the criteria for equivalence.

A price schedule that matches the specification. A bill of quantities where the design is fixed, or a performance-based price schedule where bidders design. Either way, one structure that puts the same item on the same line in every bid. A performance-based schedule should still ask for a breakdown into modules, inverters, batteries, mounting, balance of system, installation, commissioning and O&M, so the evaluator can see where two prices actually differ.

Site data. Load profile, drawings, roof or ground survey, and the grid situation. This is the section that decides the spread, and it gets its own section below.

Evaluation criteria and weighting. Published in advance: what is pass or fail, what is scored, and how price is combined with technical merit.

Contract conditions. The conditions bidders will actually sign, including payment terms, securities, retention, liquidated damages and the defects period. Bidders price the conditions, and a tender that promises them later gets priced for the worst case.

Warranty and O&M requirements. Terms for modules, inverters, batteries and workmanship, and whether labour is included. Our article on solar PV warranties explains why these are four separate documents.

Commissioning and acceptance tests. Which tests, to which standard, who witnesses them, and what performance must be demonstrated before handover. If acceptance is not defined, every bidder prices a different test. Our solar commissioning and testing guide covers what a proper test programme includes.

The eight things a solar tender needs before bids can be compared. Each gap becomes a contingency in one bid or an omission in another. Source: MIMAH engineering practice.

Site Data Decides the Spread

Load profile. The single most useful document in a solar tender. Interval data from the main meter or a logger over a representative period, generator run hours and fuel records, and for a new facility an equipment list with operating hours. Without it, each bidder assumes a load and the bids size different systems. A bid with a larger battery is not necessarily overpriced; its author may simply have assumed more night-time load.

Drawings and surveys. Roof plans with structure and orientation, or a ground plot with boundaries and soil information. The structural capacity of a roof is expensive for each bidder to establish separately and cheap for the owner to establish once.

Grid situation. Outage pattern, voltage quality, whether export is permitted, and any utility connection requirements. If the solar system must work alongside existing generators, give their ratings and controller types.

A site visit. Offered to all bidders, with minutes and questions circulated to everyone afterwards. Where travel is difficult, a photograph set with dimensions goes a long way. What a bidder cannot see gets priced as contingency.

Mistakes That Produce Incomparable or Inflated Bids

Most of these come from time pressure on the people writing the tender rather than from carelessness, and all of them are cheaper to fix before issue than after.

No load data. Covered above, and the largest single source of spread.

Brand-locked specifications. Naming one manufacturer's model without an equivalence route narrows competition to that brand's distributors and removes the price pressure on it.

BoQ quantities that do not match the specification. The specification asks for one battery capacity and the bill of quantities lists another, or cable lengths bear no relation to the drawings. Some bidders price the specification, some price the BoQ, and the evaluation compares two different systems.

Unrealistic deadlines. A tender period too short for bidders to get supplier quotations and check the design produces bids padded for what could not be checked, or no bid at all from firms that decline to guess.

No site visit. Unknown access, roof condition and cable routes are priced as risk, and different bidders price different amounts of it.

Lowest-price-only evaluation. For a fully designed supply contract this can work. For EPC work, where each bidder designs, it tends to reward the bid that left the most out.

No clarification process. Without a route for questions, ambiguities are resolved by each bidder privately and differently.

Long validity with no price adjustment. Bidders asked to hold a fixed price for months, in a volatile local currency, for equipment bought in another currency, add a premium for that exposure. A realistic validity period, or a stated adjustment mechanism for exchange rates, removes it.

Unassigned scope. Permits, utility connection, civil works, security during construction, customs clearance and import duties. Anything not assigned is priced as a contingency by some bidders and left out by others.

Evaluating What Comes Back

The cleanest structure is two stages: technical compliance as pass or fail first, and price compared only among compliant bids. That keeps price from colouring the technical judgement. For larger EPC work, scored technical criteria alongside price are often more appropriate. The World Bank's Procurement Framework is built around value for money and includes guidance on rated criteria for goods, works and non-consulting services.

Check yield assumptions. Ask every bidder for their energy yield simulation and its inputs: irradiance data source, temperature and soiling losses, degradation and availability. Two bids promising the same energy from different array sizes are making different assumptions, and those assumptions need checking against your site.

Compare batteries in usable energy. Nameplate kilowatt hours at different depths of discharge are not comparable. Convert every battery offer to usable energy and cycle life at the depth of discharge proposed.

Compare life-cycle cost. Capital cost plus inverter and battery replacements over the system life, O&M, and for hybrid systems the generator fuel that remains. The cheapest capital bid can carry the most expensive replacement schedule.

Normalise before comparing. Where a bid omits an item the others include, a common approach is to add a priced allowance for it, based on the other bids, before ranking, and to record the adjustment. Comparing an incomplete bid at face value against complete ones rewards the omission.

Query abnormally low bids. Ask the bidder to explain the price in writing before accepting or rejecting it. Sometimes there is a genuine reason. Often something is missing, and it is better found now than at the first variation claim.

The bidder's own capability is the other half of the evaluation: experience, references, local presence and financial standing. Our guide to choosing a solar EPC contractor sets out that due diligence.

A sequence for evaluating solar bids so that price is compared only between bids that answer the same question. Source: MIMAH engineering practice.

Clarifications, Handled Fairly

A good clarification process is simple. One written channel. A deadline for questions set far enough before closing that answers can still be used. Answers issued to all bidders at the same time, without identifying who asked. Numbered addenda, acknowledged in each bid. And an extension of the closing date whenever an answer changes the scope.

Answers that correct the tender are normal, and bidders read them as a sign of a careful issuer. What undoes comparability is answering one bidder privately, or letting a site visit conversation change the scope without writing it down for everyone.

MIMAH's renewable energy team delivers solar EPC, hybrid and storage projects and prepares feasibility studies for commercial and institutional owners. If you are scoping a solar project and want the load data, specification and acceptance tests pinned down before you go to market, get in touch.