Facebook claims that in a trial study of 50 brands, the participating companiesall sellers of packaged goodssaw up to a 40 percent increase in ROI if they stopped repeating ads to the same users. Instead, they redistributed ads to appear to users who had seen them less often, distributing repeat impressions more equally.

In the long term, Facebook wants to provide data about the ideal number of impressions for an ad campaign, and it hopes to break the information down to specific campaign types. A new product or service will need more impressions to raise awareness than a well-established one, for example, and the number of impressions needed will likely change based on product type.

What does all this mean for small business owners? In the short term, not too much, since the data mining project is still in its infancy. In the long term, though, it could prove to be a treasure trove of data about ad targeting, especially for companies that sell products at retail.

Datalogix pulls in sales information from a staggering number of purchases. Assuming that Facebook launches the feature in full force down the line (and especially if it includes regional filters), that data should come in handy for fine-tuning your social ad strategyand deciding whether a is worth the cost of admission.

Thats assuming, of course, that Facebook's theory about the value of impressions holds true under scrutiny.