I saw an interesting question this morning on the Twitter from Shiv Singh:
What’s the value of a tweet sent by a person with a million followers? What’s the Cost Per Tweet Impression?
I’m going to marinate on this for a few days, because it’s a thornier problem than one might initially surmise. Here’s where the problem gets deep. First of all, unlike, say, a banner ad, there is no guarantee that I will be “served” the impression of a tweet, if I miss it in my timeline. In contrast, if a website says it served a million impressions of an ad, then it was served to roughly two million eyeballs. Of course, many of those eyeballs may have blocked, ignored, or quickly scrolled away from that ad, but I can at least quantify what those of us who research the Out-Of-Home media space call “opportunity to see.”
With a tweet, it’s a little less straightforward, since one million followers don’t equate to one million impressions. If you follow over a thousand people on Twitter, your timeline is cluttered to the point that you’d likely have to actively seek out tweets from an individual Twitter user in order to guarantee that you had “seen” a given tweet from that user.
Now, there are people on Twitter that I do make a point of catching up with by going to their Twitter page periodically and reading tweets I might have missed. You probably do the same. Generally, these are people that I have a connection with – friends and colleagues with whom I have engaged in conversation with, either on Twitter or elsewhere. “Engaged” is the key word here. If I’ve been away from Twitter for a while, I might catch up with Tim Hayden, because I know he has worthwhile things to say and we’ve had some quality interactions on Twitter. Same with Matt Ridings, who I’ve yet to meet in real life, but is someone with whom I’ve shared some very thought-provoking exchanges.
So, if Tim, or Matt, or Jason Falls, or Amber Naslund or any other Twitter user I find engaging tweets something, I’m likely to see that tweet, even if I missed it in my timeline when it went “live.” Their tweets have value to me, and, it can fairly be said, influence me, because I see their tweets as conversations. Even if they are promoting a link to one of their posts, I have an expectation of engagement. If I ask them a question about that link, or engage them about one of their posts, I know I’m not wasting my time – they’ll probably answer me back.
Contrast that to someone on Twitter who, to return to Shiv’s original question, has one million followers. Let’s take Tony Robbins, who currently has 1.8 million followers. Tony is what can fairly be called a Twitter “broadcaster,” and I make no value judgement whatsoever about this. If he tweets something, I suppose I might reply, but my Twitter affections shall go unrequited – I have no expectation of a reply from Tony. This isn’t to disparage him; merely an acknowledgement that a million followers doesn’t scale. I don’t expect a reply, because I know that he has too many followers, too many possible conversations to engage in to have even seen my tweet, and I generally don’t like talking to myself (or to someone “managing” the Tony Robbins Twitter account not named Tony Robbins.)
At some point, then, a Power-Twitterer stops engaging people and becomes a broadcaster, because they have no other choice. If LeBron James actually does answer someone’s tweet, it’s the Twitter equivalent of answering one piece of fan mail – an “example” of engagement that doesn’t prove the rule. If someone with 1,000 followers tweets a question, I think those followers expect that this is an entree to a conversation. If someone with a million followers tweets a question, I dare say the vast majority of those followers realize the question is rhetorical. (I admit there is also the confounding variable of the retweet to deal with here, but engagement also comes into play in retweet behavior.)
Perhaps, therefore, the number of followers a given Twitter user has becomes a sort of barometer of engagement expectations. South of some magic number, a tweet is an invitation to connect. North of that number, a tweet is an inefficient broadcast advertisement. In this sense, that number is like a “Dunbar’s Number” for a new age of asymmetrical, asynchronous conversation. And that actual number, the “elusive statistic” referred to in the title of this post, is less important than what the number is perceived to be by the followers of a given Twitterer.
So, back to the original question: what’s the value of a tweet sent by a person with a million followers? On a per tweet basis, possibly more than a million traditional mass media impressions, given some level of self-selected targeting. Probably less, however, on a per follower basis, than a tweet sent by someone with 10,000 followers. How much less is, to quote the poet Rumsfeld, an unknowable unknown. The perceived “Dunbar’s Number” for Twitter, then, neatly demarcates the boundary between conversational marketing and broadcast advertising.
OK, so I admit that reach and engagement aren’t mutually exclusive. Clearly, more questions than answers here. If this perceived number doesexist, however, what would you estimate it to be? Or, rather, what is it foryou?
Or have I over-thought this to the point of utter hilarity? 
COMMENTARY: Twitter presents a special challenge for advertisers, because it is nothing like print publications, television or radio, where readership, audiences and demographics are regularly captured and measured. Twitter is one channel that runs 24/7, 365 days a year, and its audience has many different attributes, the time frames and length of time online vary considerably, making measurements very difficult. I am not even sure that Twitter know how to do this type of media campaign analysis. They are barely keeping the ship afloat trying to minimize the technical difficulties and capacity overloads that occur far too often.
I believe that Tom Webster is on the right track with his assumptions of Twitter members and their followers. Once you become a so-called broadcaster, then it becomes one-way engagement, with the exception of a few chosen individuals that Twitter member choses to follow and actually engage with.
So what is a tweet worth? That would depend on the probably that a certain Twitter member can be served an ad, and includes many factors, including, but not limited to the following: 1) how often a follower logs into Twitter 2) how often a follower is active in seeking out the member's tweets, 3) how interesting is the tweet?, 4) the timeline (what time frame) the Twitter member online, 5) how long a Twitter member is logged in, 6) how often that Twitter member reads a tweet and 7) the number of individuals who follow and are being followed by the Twitter member.
A Twitter advertiser using Preferred Tweets would obviously want to serve their ads to Twitter members who log in often, spend a significant amount of time on Twitter, follow brands that are similar or complimentary to their brand, actually read tweets regularly, and so forth.
A smart brand marketer would expect that Twitter have the ability to segment, sort and filter data to suit each advertiser. That data should provide suitable key words and member demographics that can be mine them from the trillions of bytes of data created by Twitter members. I would also expect Twitter to provide advertisers an analysis of each campaign concerning the number of ads served up to individual meeting each demographic and the probability that they saw the ad based on the attributes of each individual member. How you do this with a large homogenous target audience is beyond me, and this is what makes calculating the cost of a tweet impression very difficult, if not impossible.
Courtesy of an article dated September 13, 2010 appearing in brandsavant
Comments