How to evaluate a solar ERP vendor: a demo script, a scorecard and red flags
Make every vendor run your scenarios on your data, score them on weights you set in advance, and ask the questions that expose a team that has never done solar.
- Same scenarios, your data, live demo. No slides.
- Fix scoring weights before you see any product.
- Weight project costing and field usability highest; they decide success.
- Call three references. Ask what broke first.
- Get data export, ownership and exit terms in writing.
Quick answer
Give every vendor the same ten scenarios drawn from your own projects and make them run them live, with your data, in the demo. Score each result against weights you fixed before the first call, giving most weight to project costing and field adoption. Ask for three references you can phone, get data-export and ownership terms in writing, and walk away from anyone who shows slides instead of your scenarios.
We are a software builder and may bid for work like this ourselves, so apply this script to us too.
What weights should the scorecard use?
Set them before you see any product. A suggestion, which is our opinion and should be adjusted to your business:
Suggested scoring weights (points out of 100, our opinion)
Source: Our suggestion; adjust to your business. Weights sum to 100.
View data
| Weight | |
|---|---|
| Project costing and live margin | 20 pts |
| Field usability and adoption | 20 pts |
| Approvals and subsidy tracking | 15 pts |
| Billing, retention and receivables | 15 pts |
| Procurement, serials and warranty | 10 pts |
| Data export and integrations | 10 pts |
| Implementation and support | 10 pts |
Which ten scenarios should every vendor run live?
Use your own data. Give each vendor the same inputs.
| # | Scenario | Pass looks like |
|---|---|---|
| 1 | Turn an accepted quote into a project with its BOM as the budget | No retyping; budget equals quote |
| 2 | Raise a purchase order that differs from the BOM | The system flags the variance with a price difference |
| 3 | Show estimated, committed and actual cost and forecast margin for a live project | Numbers appear from the data you entered, not a canned screen |
| 4 | Set up a milestone schedule with triggers and evidence, then bill one | Billing blocked until evidence is attached, or clearly flagged |
| 5 | Record retention and a bank guarantee with expiry alerts | Alerts configurable; linked to the project |
| 6 | Create a rooftop project and move it through DISCOM feasibility, inspection and meter stages | Per-DISCOM workflow; days waiting visible |
| 7 | Scan serials at goods receipt and again at installation | Match shown; warranty dates calculated |
| 8 | Do a site survey and installation checklist on a phone in airplane mode | Works offline; syncs after |
| 9 | Raise an O&M ticket from an alert and attach it to the unit’s history | Serial, warranty and AMC status visible |
| 10 | Export all project, cost and invoice data | Open format; no fee; shown on screen |
Count how many they pass without a custom development promise.
What questions expose a team that has never done solar?
- “Show me forecast margin for a project mid-way.” If they show a final profit report, they do not have it.
- “How do you handle a different document set for each DISCOM?” If the answer is a custom field, ask how many.
- “How do you capture 200 module serials without typing?”
- “What do your solar customers complain about?” Honest vendors have an answer.
- “Which of your solar customers can I phone?” You want three, and you want to ask what broke first.
What are the red flags?
- A demo on slides or a canned dataset.
- “That can be customised” for more than two of your ten scenarios.
- No clear answer on exporting your data.
- Pricing that depends on user count in ways that punish site staff access.
- Reluctance to name references.
- A fixed price with a vague scope.
What should the contract say?
Have a lawyer review the final text. At minimum: you own your data; you can export it in open formats at any time without fees; what happens on exit and how long data remains accessible; where it is hosted and backed up; who owns any customisation you pay for; support response times; and what triggers extra charges.
What does a decision look like?
Score each vendor out of 100 using your weights. Check the top two with references and a small paid pilot on one project type. Do not buy on the demo alone.
Where does this not apply?
If you are considering only a single low-cost tool for one narrow job, such as quoting, a lighter evaluation is fine. This script is for a system that will hold your project money.
What to do this week
- Write your ten scenarios from real projects. Remove names where needed.
- Fix your weights and share them with your partners before the first demo.
- Draft the three reference questions.
Part of the guide Running a solar EPC business on connected systems.
The scenarios, weights and red flags are our own model, not an industry standard. No vendor was assessed. Weights sum to 100 (20+20+15+15+10+10+10). Contract points are general and not legal advice.
More by GarunaCDX
Let's talk about your build.
Tell us what you are building. A founding engineer replies within one working day with how we would approach it.
