Project insight

Before You Sign: A 7-Step Energy Procurement Checklist (Tesla Batteries, Solar Roofs, and Wind Projects)

A procurement lead's checklist for commercial battery storage, solar, and wind projects — including the battery health test baseline, solar cost traps, and inverter sizing mistakes that cost me $180,000.

Who this checklist is for

If you're the person signing off on commercial battery storage, solar, or wind projects — and you've never done it before — this is for you. I've been handling commercial energy procurement for eight years. I've personally made (and documented) 11 significant mistakes, totaling roughly $180,000 in wasted budget. Now I keep the checklist below to stop other people from repeating them.

Seven steps. None of them are glamorous. All of them have saved me at least once.

Step 1: Build a total-cost model before you look at a single quote

My first mistake, back in 2017, was comparing storage quotes on price per kilowatt-hour. The cheapest option won. It looked great on the spreadsheet. Then we got hit with commissioning delays, an upgraded switchgear requirement nobody flagged, and a service contract that cost 40% more than the equipment itself over five years.

I still kick myself for that one. The $30,000 I 'saved' on the initial quote turned into a $75,000 problem over the first three years.

Here's what I do now: before I open any vendor PDF, I build a TCO model with five columns — equipment, installation, interconnection/permitting, O&M (years 1–10), and end-of-life/disposal. If a vendor can't help me fill in the O&M column with real data, that's not a red flag. That's a disqualification.

People think expensive vendors deliver better quality. Actually, vendors who can prove quality — with operating history, documented uptime, third-party monitoring — are the ones who can charge more. The causation runs the other way.

Step 2: Run a battery health test baseline before you sign

This is the step everyone skips. Ask any battery vendor — Tesla included — for the state-of-health (SOH) baseline on the specific cells or modules going into your system, and demand that the same test gets repeated at year one.

Why? Because battery degradation is invisible from the outside. A system can look fine on the monitoring dashboard while losing 3% capacity per year on the cells that matter. Without a baseline, you can't prove anything at year five when you're arguing about warranty terms.

For fleet vehicles running Tesla hardware, the battery health test in the service menu gives you a real number — not a dashboard guess. I now require the same discipline on stationary storage: annual SOH reports, same test protocol, same conditions. If the vendor resists, ask why. The honest ones don't.

The assumption is that batteries only show problems when they fail. The reality is they fail slowly, and you need the baseline to see it.

Step 3: Tear apart solar quotes into dollars per kilowatt-hour, not dollars per panel

The Tesla solar roof cost question comes up constantly, and I get why — it's a fundamentally different cost model than traditional panels. Solar Roof is priced by roof square footage and roof complexity, not by panel count. That means the same square footage can cost wildly different amounts depending on pitch, dormers, and tear-off requirements.

So I stopped comparing quotes on panel price. Instead I ask each vendor for a single number: estimated lifetime kWh produced, divided by total installed cost (including roof work, structural upgrades, and inverter replacement at year 10–15). That's your real cost per kWh. Anything else is noise.

And when I say 'including inverter replacement' — I mean it. More on that in Step 5.

Step 4: Ground your resource assessment in actual turbine data

If your project touches wind — either as a PPA or as a direct asset — you need to know what's actually deployed in your region and what scale the equipment runs at.

Concrete example: Indiana had just over 1,000 utility-scale wind turbines as of 2024, spread across roughly 30 wind farms. That's the installed base. If a developer hands you a production forecast, you can sanity-check it against what similar turbines are actually producing in the same wind regime.

And know the equipment. A modern onshore turbine blade runs 50–60 meters. Offshore blades have pushed past 115 meters. That matters because size of wind turbine blades determines cut-in wind speed, capacity factor, and — critically — the logistics cost of getting the thing to your site. I once watched a project get delayed six months because nobody checked the bridge clearances on the route.

The wind turbines in Indiana are mostly 1.5–2.5 MW machines. If your forecast assumes 4 MW performance from 2 MW hardware, the model is fiction. Ask for the make and model. Then look it up.

Step 5: Size your power electronics for the load, not the brochure

This one sounds obvious until you're the person who got it wrong. The same logic that applies to a power inverter for a computer — match the VA rating to actual load, leave headroom for startup surge, don't buy the biggest unit on the shelf — applies to commercial-scale inverters and PCS (power conversion systems). The scale changes. The physics doesn't.

My mistake in 2021: I signed off on an oversized inverter 'for future expansion.' The future expansion never came, and the inverter ran at 20% load for three years, which is below its efficient operating band. What I thought was headroom was actually a permanent efficiency penalty.

Step 5's checkpoint: I now require the vendor to provide a load profile simulation at 25%, 50%, 75%, and 100% of nameplate — and I ask what happens at 20% and 110%. If they can't answer, they haven't modeled your site.

Step 6: Write the monitoring and verification clause before you sign

This is the one I regret most. In 2019, I signed a storage contract without a data-access clause. Our vendor owned the monitoring platform. When we wanted raw data to verify performance claims, we had to file a formal request. It took six weeks. Six weeks of a system we'd paid for, running blind.

Now every contract has a monitoring clause with three requirements: (1) we get API access or a data export on demand, (2) monitoring data is jointly owned, and (3) the vendor provides a quarterly performance report against the modeled baseline. No exceptions. No 'we'll provide a dashboard' handshake.

Step 7: Build your 'do not do' list

My last step isn't a step. It's a page I keep in the front of the project binder. These are the things I've seen go wrong, and the ones I refuse to repeat:

  • Don't sign a contract with an unverified O&M cost estimate. Make them show you the last three years of service tickets on comparable systems.
  • Don't accept a 'typical' degradation curve without a written year-one SOH target and a warranty remedy if it's missed.
  • Don't let the interconnection study happen after the equipment order. I've seen this delay commissioning by nine months.
  • Don't compare kWh pricing across vendors without also comparing warranty term, degradation guarantee, and data ownership.
  • Don't assume bigger is safer. Oversized power electronics run inefficiently and age faster.

A few things I still get wrong

I don't want to pretend this checklist solves everything. Last year I underbudgeted the permitting line on a solar project by about $14,000 because I used a 2021 cost estimate instead of pulling current numbers. My gut said the number looked low. The spreadsheet said it was fine. I went with the spreadsheet. I should've listened to my gut — and I've updated the workbook to pull fresh numbers every quarter.

That's the real lesson. The checklist isn't there because I'm careful. It's there because I'm not, and I've got the receipts to prove it. If it saves you one $3,200 mistake, it's done its job.