This week, as part of Social Media Week 2014, I participated in a panel presentation titled “The ROI of Social Marketing, Online and Off.” The discussion focused on a question that sounds simple, but is actually one of the complex issues facing marketers today: can you accurately measure the ROI of social marketing?
And if you can, how good is it?
Fortunately, some of the industry’s leading minds are on the case, and there’s a growing body of research showing how to measure and optimize the success of your social campaigns.
Right now, there’s no better way to measure the sales effects of a social-marketing campaign than a matched-panel analysis (MPA). Usually executed with the help of external partners, like Nielsen or IRI, an MPA requires establishing test markets (where a campaign will take place) and control markets (where it won’t); after the campaign has finished and the resulting word of mouth (WOM) has had time to spread, the differences in product sales between the test and control markets can reveal the impact of the campaign, including its ROI.
We’ve used MPAs extensively at House Party. For a recent 3,000-party campaign, for instance, we selected hosts all around the country, but we saturated certain zip codes around Grand Rapids and Buffalo and excluded others:
You can see the result here — a 6.3% sales lift in the test zip codes, which (when projected out to all parties) meant hundreds of thousands of dollars in incremental sales:
On average, we’ve found that a moderately sized House Party campaign — 2,500 parties — results in an ROI of $2.50 after only three months and $5.00 after six months. Larger parties pack an even bigger punch: an ROI of $3.09 after three months and $6.18 after six months.
If you want to dive even deeper into your campaign’s results, you can also conduct a shopper study through a firm like Datalogix. As we explained in a blog post last month:
While matched-market studies can determine overall ROI, shopper-data analyses color in the details. Shopper-data analyses compare shopper-level purchase information (from a provider such as Datalogix or Nielsen Catalina Solutions) of consumers the marketer directly reached with that of a look-alike control group. After the WOM has spread, the marketer compares such metrics as dollar sales, unit sales, dollars per trip, frequency of trips, basket ring and loyalty (the extent to which a consumer switched to or from the brand). Shopper-data analyses can’t determine the total sales lift and ROI of a campaign, since they only include the consumers the marketer reached directly, and not those reached via consumer advocates. But because they reveal individual behavior (as opposed to MPAs, which reveal market behavior), they help marketers understand what’s behind the campaign’s sales lift and ROI.
And more and more, marketing mix modeling (MMM), which analyzes a brand’s full marketing spend to see how well each element worked, can accurately “read” the impact of WOM (which wasn’t always the case). Part of the difficulty results from the fact that marketing components don’t exist in isolation — all of them, especially WOM, impact and amplify the others. As Nancy Smith, of Analytic Partners, explained during our Social Media Week presentation, “WOM not only drives sales directly, but through Paid, Owned and Earned media.” Hearing about a product from a friend can drive a consumer to search for more information about it, or visit a company’s website or a third-party review site. The best modelers are developing ways to account for this, though, and as MMM continues to evolve, it’ll become even easier to properly credit social campaigns.
Marketers no longer have to work in the dark, relying on intuition and qualitative data to guide their social-marketing planning. Robust measurement tools are out there, and — best of all — they’re revealing what we’ve long suspected: that word-of-mouth marketing, when properly planned and well-integrated with the entire marketing plan, is the most powerful tool at a brand’s disposal.