Social networking’s limits
I am still trying to figure out what the primaries are telling us about social networking. I think some of this will be important for CRM and as I have noodled on what it all means, I have been surprised myself. Before I go on, this is not a discussion of who won or my candidate preferences, just musings on how it went down.
First, I think there is something overlooked about the differences between Iowa and New Hampshire — and if it has not been overlooked, it certainly has been underreported. The big difference is in how the votes are cast — and it’s important. In Iowa, as everyone knows, the general description is caucuses and the format is open voting; in New Hampshire, they have a formal election and use a secret ballot. Caucuses are a form of social networking designed to form consensus. In a traditional secret ballot, the consensus is made evident only once the votes are counted.
It might be surprising, but the secret ballot is a relatively recent introduction to politics. The ancient Greeks and Romans used a form of secret ballot, but when democracy made a comeback after a roughly 17-century hiatus, a lot of people couldn’t read and open balloting was the form.
In the 1850s, Australia introduced the concept of a secret ballot, and in the U.S. in 1888, Massachusetts was the first state to do so. Kentucky was the last adopter in 1891, which means that Grover Cleveland in 1892 was the first U.S. president to be elected solely by secret ballot.
Consensus is a good thing, but open balloting is open to abuse and intimidation in voting, and that’s one reason secret balloting came into being. Intimidation can be a subtle thing. It doesn’t require a bunch of big guys with narrow foreheads to affect a vote. For some people, it can take a lot of courage just to stand in a small group to be counted for a particular candidate, or anything else for that matter.
I wonder whether Hillary would have won in New Hampshire if they had not used secret ballots. Since many people apparently made up their minds late (according to news reports), some of them may not have had their reasons fully worked out in their own minds.
In other words, their choice might have been more gestalt than conviction, and they may have not been able to articulate their reasons. Without the ability to clearly articulate their reasons in a caucus format, some people might have folded and selected another candidate. We will never know, but we certainly have proof of a big discrepancy between the information the pollsters collected and the final results.
It’s quite a different thing in the marketing world, though, and one of the valuable attributes of social networking outside of politics is its ability to capture the thought process as an idea percolates and matures — and consensus is reached — in a population. Since people make actual purchase decisions in the privacy of their own minds, the intimidation factor is either not operative or greatly reduced.
That brings us to a related issue, the sample population. Diane Hessan, CEO of Communispace, a company that develops and manages customer communities on behalf of corporate clients, gave me a few insights about populations last week. One of the overlooked issues there is the quality of the population as measured by individual activity levels.
As with any population, there is a bell curve for participation — some people participate a lot, some visit once to sign up, look around and never return. Within those extremes there are gradations — people who visit a site but only read, some who read and post new content, people who show up once in a blue moon and others who are addicts.
When a social networking site’s raw population numbers are quoted, beware. We also need demographics, metrics and filters to understand the meaning of the data that gushes from these services.
According to Hessan, a site’s total membership cannot realistically be expected to be engaged in any single issue, so understanding who is engaged — and how many of them there are — can mean a lot if you are expecting that population to act as a surrogate for the market.
Marketers with high expectations for their data and low thresholds for interpretation can be surprised when reality does not live up to the numeric expectations generated by a community. Furthermore, if a company’s production apparatus takes the raw data and runs with it, complications can be nasty.
All this goes to show that we’re still finding our way. The technology available is powerful, important and valuable for politics as well as conventional marketing but we need to learn how to use it effectively. We need controls, standards and insights to truly understand the data that comes to us in torrents from new tools.
At this point, we’re still watching for ultimate results to see how it correlates with the data we’re collecting. Nothing wrong with that, it’s just the adoption process at work.