For all the buzz among entrepreneurs about social media as a tool to gain business, the evidence of its importance is largely anecdotal. When talking of any type of direct marketing--sorry, buzz-riders, but that old-fashioned term is what this is--you need to understand the math behind the campaign.

There are two major numbers underlying any direct marketing campaign: how much customers are worth over their lifetime of doing business with you and how profitable they are. The former is the granddaddy, though, because it drives profitability. The more incremental business someone does with you, the more you can amortize the often expensive cost of acquiring the person in the first place.

That's why a study from predictive analytics vendor Custora is interesting because it purports to measure the relative pull, and quality of customer, of a number of online marketing tools, including Google ads, Facebook ads, banner ads, affiliate sales, Twitter, cost-per-click ads, cost-per-impression ads, referrals, email, and organic search.

The Findings

The study used "data spanning 72 million customers from 86 U.S. retailers across 14 industries." No further details and no profitability numbers (too hard to calculate across so many organizations), but the lessons about customer lifetime value (CLV) were intriguing.

According to this analysis, the top source for high CLV customers was organic search. These people were worth 54.25 percent more than the average. Next up was cost-per-click, where acquired customers were worth 36.9 percent more. Referrals were third at 26.1 percent. At the bottom were social networks. Customers that came from Twitter--wait for it--had CLVs 23.36 percent less than average. Second lowest, coming in just below banner ads (1.34 percent more than average) was Facebook at 1.31 percent.

Custora attributed the low value of Twitter to the volume of discounts offered in promotional tweets. Still, Facebook also shows a relatively low value, suggesting that either the nature of social media or the way companies use them could put a real damper on the relative marketing value of them.

And at the top, you might understand why referrals generally do well. People often talk to like people, so a company is more likely to reach the proper prospect profile. But why organic search and cost-per-click? I'd argue that the mechanism of people actively pursuing and choosing a company naturally makes them more likely to establish a strong relationship.

Also of interest is the growth of various channels in acquiring customers in the first place. Cost-per-click and email have surged ahead since 2009. Lagging far behind again are social media channels.

I don't think that this data rules out using social media by any means, particularly as it was the last recognized channel that would be awarded the credit for acquisition and for CLV levels. Perhaps social networks are useful as a communications step and not necessarily as a way of closing business. But the information should send marketers back to both their spreadsheets and drawing boards.