Residential Job Costing for the Solar Industry - PVBid

Residential Job Costing by Connor EnglishI have been talking a lot about the intricacies of estimating on a commercial level and felt it is time to step back and look at a more residential-centric view. In past articles, I have talked about how much I dislike the price-per-watt metric. I see that the residential industry, as a whole, still focuses heavily on price-per-watt. So I want to talk about ways to tackle residential job costing with less reliance on price-per-watt while maintaining efficiency in pricing.

When a company analyzes their costs, you must first identify big factors and remove them from the equation. This basically boils down to the 80/20 rule. That is to say, solve 80% of the problem using 20% of the work. The low hanging fruit (or big factors) for residential are typically inverters and modules. Reviewing past costs start with subtracting out the module and inverter material costs, which are basically known commodities and will continue to be so on future bids. That leaves us with a dollar amount that encompasses the labor costs and the rest of the material costs.

For the purposes this article, I will call this the labor BOS and the other BOS, respectively. The term Other is due to indirect costs being included.  Why should we take out the big factors?  Well, module and inverter price fluctuation can add noise to a system so the more known numbers you can remove the more effectively you can predict cost.

Residential Costing GraphThe Formula for Residential Job Costing

If your company doesn’t track labor separately from everything else, you should start. Perhaps you can just get out on a roof and time your crew to get a feel for how long it takes to install a system. As always, a bare minimum data set has three points. Put your labor BOS and other BOS data into a spreadsheet so you can start mucking around with the data. From here, you can plot a graph of your cost as it relates to several different metrics, with watts on the x-axis. See the figure for an example. I recommend starting with cost/watt, cost/module, and cost/square foot for your y-axis values. Due to varying module densities in the world today, cost/module or cost/square foot will likely be a more accurate predictor than cost/watt.

You will also see that the graph tends to have a high cost per module for smaller systems. Partly, this happens when there is a time cost for getting a crew on the roof and travelling to site.  Additionally, wire sizing is practically the only thing in this BOS that actually behaves by a cost per watt. Those behave more accurately by a cost/watt/foot. Look at the behavior of the costs, so you can determine cost ranges. For residential job costing, it makes sense to break the cost per module out in groups of modules to match the observed trend.

Don’t forget to add in your profit and margin when you reassemble your costs. I have found this method works well for systems under 10 kW. Once you go over that, you will be more prone to losing money through inaccuracies.

Keep in Touch

I will attend Solar Power Northeast in February and the NABCEP conference in March. Let me know if you’d like to chat at one of those events. As always, I would love to hear from you! Tell me how you approach residential job costing in the comments or by email!