In order to bring down sky-high acquisition costs for residential solar in the U.S., SolarCity and Solar Universe undertook two very different tactics in September. Market leader SolarCity spent $120 million to acquire Paramount Solar, an originator known in the industry as a “closing machine” while working the phones. Solar Universe, meanwhile, bought Gen110 for an undisclosed amount. The company prides itself on being especially skilled at door-to-door sales.

Home security specialist Vivint showed the industry how effective it can be to let sales personnel swarm entire neighborhoods to offer solar leases, climbing to the number-two spot in the national residential installer ranking within two years.

But Solar Universe might have taken an expensive route.

Estimates from GTM Research's new report U.S. Residential Solar PV Customer Acquisition reveal that in the second quarter of 2013, it cost $0.49 per watt for an average 6-kilowatt residential solar PV system to gain a single new customer. To put that into perspective, in the same time period, modules could be bought for $0.68 per watt, so the companies spent roughly 72 percent as much on customer acquisition as they spent on modules. And to compare it on a global scale, in Germany, customer acquisition costs are just $0.07 per watt, according to a study from the Lawrence Berkeley National Laboratory.

Among the six most common lead-generation strategies, door-to-door sales is one of the most expensive, with costs per customer of between $1,000 and $2,000. The closing rate for door-to-door sales is better than those of bidding platforms or direct-response marketing, but not as good as referrals, which are also cheaper.

One of the reasons why door-to-door has worked well so far for Vivint is that the company is going after the low-hanging fruit. Strategically selecting neighborhoods based on factors such as electricity rates and lack of competition has lead to a much narrower geographical distribution than SolarCity's installations, as a map of residential solar systems built in Massachusetts from January 2012 to June 2013 shows.

From GTM Research's U.S. Residential Solar PV Customer Acquisition report

The downside of Vivint's approach is that the point of solar saturation for a neighborhood is reached a lot faster. However, evidence indicates that it will most likely take years before states in which installers are selling door-to-door run out of areas to target.