The Dark Underbelly of Random Chance Targeting

A few years ago I worked on a campaign for a client whose primary marketing goal was to be perceived as a technology leader. We ended up creating a surprisingly robust rich media ad that incorporated a chat room into the ad banner space. As the ad was served, the back end would connect the two most recent ad recipients so they could actually chat with one another. But because this was such a novel concept and because interactive media was still not widespread, one of the tools we added to the chat room was a robot designed to help motivate the chatters by adding elements to the conversation now and again.

While the campaign was a technological breakthrough, it wasn’t much of a marketing success. One of the biggest hurdles was that total strangers don’t often have anything meaningful to say to one another when put on the spot to suddenly start a dialogue. Second, the robot, with all good intentions to help spur on conversation, usually added poorly timed and totally irrelevant non sequiturs.

Person one: This is pretty cool, a chat room in an ad banner!

Person two: I agree. I wonder how they do this?

Robot: I just love the products that [advertiser name] makes, don’t you?

The robot’s contributions to conversation sounded staged and phony. The client — concerned that people might actually say bad things about the brand during the course of their conversations — requested we add a language-parsing filter that could remove any potentially offensive terms or words, along with any mention of competitors’ products, from any conversation. (This move resulted in me receiving the most interesting e-mail of my career, containing over 300 massively offensive words that we needed to filter out). The result was often a conversation filled with gaps where the potentially offensive or competitive words were removed. All in all, this model relied so heavily on random chance to work that it was likely doomed from the start.

In this particular example, the focus was on getting a message for the greatest number of people and to use technology to filter out how those people interacted with the brand after the fact. Today, we can use behavioral targeting to identify the best recipients for a message before we send it out and thus can better reduce the need to rely on random chance to get the marketing work done.

I mention this because the rise of social media has brought with it a clearer understanding of the parameters that marketers need to consider when engaging and targeting online audiences. Not only does the content being provided to consumers need to be relevant to keep the conversation going, but it also needs to feel real to its recipients.

While social media has only had a passing relationship with behavioral targeting in the past, we’re seeing many of the same strategies that behavioral marketers have used to focus on audience segments can also be applied to social media campaigns in certain cases.

Last week, a SocialMediaToday post took a close look at how British furniture store Habitat’s recent attempts to get its marketing tweets in front of consumers backfired horribly.

Sadly, it appears that the marketing team at HabitatUK were far less interested in reaching people who might be most interested in new furniture and might schedule a future a visit to one of their stores, and instead opted to reach as many people as possible by spamming them with Twitter messages that were linked to popular and largely irrelevant hashtags (define), such as #iPhone, #Apple, and even popular news items covering the recent unrest in Iran, such as #MOUSAVI. Due to the nature of social media, and Twitter in particular, it didn’t take long before message recipients let HabitatUK know just how poorly its marketing attempts went.

With the possible exception of some branding campaigns, using reach and frequency techniques as a way of getting a message out to an audience is a nonstarter today. While a greater number of people reached can result in greater response, taking a few minutes to consider who you are really trying to reach, what they want, and how to start a meaningful conversation with them can not only achieve better campaign results but also decrease the media costs used to connect with that audience. At this point in the game, trying to reach everybody as a target audience relies too heavily on random chance probability and is most likely going to result in a lot of disconnects and annoyed potential future customers.

You’d almost be better off having a robot ping them with irrelevant and nonsensical statements from time to time.