What do airport security and social media monitoring have in common? More than you would think (especially since most of you probably answered “nothing.”) Or at least that’s what I’m going to try and make a connection between…
Unless you’ve been living under a rock, the TSA is having some image issues, due to the recent rollout of new security screenings that leave passengers with the unenviable choice of having a picture taken that you probably don’t want to post on your Facebook page or receiving a–shall we say “thorough”–pat-down. This is being done, of course, to reduce the possibility of someone slipping through security to wreak havoc on a plane. What some are now questioning is the cost (in time, privacy, and currency) versus the probable risk. This boils down to a standard question of risk/benefit analysis: to what lengths are we willing to go in terms of cost (time, privacy, currency) to mitigate the risk of a terror attack?
Today’s #MeasurePR chat, hosted by Shonali Burke, featured a discussion with guest Seth Duncan of Beyond Analytics. Seth wrote a great post for Shonali’s blog, Waxing UnLyrical, about automation in PR measurement, and this was also the topic of today’s #MeasurePR chat. The concerns of participants in the chat were what I had expected them to be: accuracy and cost. This too can be viewed as a risk/benefit analysis.
Is there a risk in basing decisions on inaccurate data? Yes. Does that risk outweigh the benefit (using automation to get it done quickly, but with less accuracy)? That will vary from client to client, and will largely be based on the objectives set out.Will you miss key insights if you over-rely on automation? Probably. Will that matter to your client? Again, that depends on what the program is and why you are running it.
There are, of course, hybrid approaches–use the technology (automation) to reduce the load on the human analyst. (That’s what we do here at CustomScoop.) But some were still concerned that including humans in the mix would make the cost too high. Again, this is going to depend on what is being monitored, why, and to what objective.
There is no one-size fits all answer here.
Ultimately, I’m reminded of the Project Triangle that you can pick two of three: Fast, Cheap, Good. If you want your measurement to be fast and good, it will not be cheap. If you want it cheap and good, it will not be fast. And if you want it cheap and fast, well, you will sacrifice some accuracy.
Humans will be involved at some stage in any reasonably reliable analysis of content. Some quotes from the chat today:
Finding potential “influencers” can be automated; but determining their influence still needs human touch.
finding influencers is more of an art than an algorithm
We don’t need to ask automation vs. humans. I think the theme today is that we need to do a hybrid #measurepr. Q is what proportion?
Since humans do need to be involved at some point (we aren’t yet welcoming our Robot Overlords) a key consideration is, at what point are they introduced? This will depend on a variety of factors including the overall volume, what the objective of the monitoring and analysis is, and yes, what the budget is. If your objective is to search for intelligence–that “needle in a haystack”–humans need to be involved early on. Intelligence doesn’t necessarily come from “influencers,” it can sometimes be found on a small blog with few readers. If your objective is a broad understanding of how consumers generally view a new product, that’s a completely different scenario.
The risk/benefit question must be considered when assessing options for clients, and in the years that I have been doing this work and reading case studies, I haven’t seen two exactly alike. The risk–and benefit–will be different for each program.
The good news is, no full body scanners in PR!