This week our strategy team got up in arms around a question from a partner agency that focuses on traditional (i.e., non-digital) marketing, and I felt it was worth sharing and discussing here. The agency was working on a project for one of their clients and asked the question: “What percentage of tweets are seen?” The data team at this agency was sitting down to build an algorithm to model tweet impressions and was looking for our digital/social expert opinion. An interesting question, indeed, and an example of how much baggage we need to overcome to move to the next evolution of marketing.

First, to address the question directly, our own Jonathan Richman provided some insights on the challenges of measuring how many people see a given Twitter message. He brought up the points that tweet readership varies by the time of day, how many people retweet a message, how many followers they have, the number of lists people might be on, the use of hashtags (#), and the types of Twitter API readers that people are using. There are challenges such as the fact that the more people you follow and more people who follow you, there are more “impression opportunities” but the ability to pay attention to any one of the individual tweets goes down. Richman’s answer as to how you can calculate all of these impacts: You can’t.

I added my own two cents to his response: You shouldn’t.

I didn’t have to ask my partner agency for an explanation to see what they were trying to do for their client. In the traditional marketing world that still dominates, clients want to measure marketing in common terms. For years this least common denominator has been the “impression”; brands have bought TV, print, and radio ads in the cost-per-thousand-impressions format, which allows them to compare spending across any form of interruptive media. In theory, this also helps marketers decide where to focus their budgets and time. Our industry’s most-frequent response to new media is to try and stuff it into the box of old media, so that dollars can flow from one to the other with confidence. So the question: How many impressions does each tweet receive?

So lots of very smart people are now spending their time modeling impressions per tweet, just because it’s the model we’re used to. The very obvious problem is that this is the wrong way to measure new media and new marketing that tools such as Twitter are bringing to brands. If we want to win in a world of exploding social change and killer competition, we must invent new measurement models rather than forcing ourselves through something that means less and less.

Last fall I wrote about how we marketers must abandon the common yet meaningless measure of impressions and instead begin to measure engagement—a key step on the path to Marketing with Meaning. Engagement to most of us in the industry occurs when a customer chooses to spend time interacting with marketing. It’s actually something that can be measured across all media as well. You can count the number of people who, say, choose to watch your YouTube video, subscribe to your email list, or become a fan of your brand on Facebook. Sorry, you do have to do a little more modeling to gauge the value of these different types of engagement—but this is how we marketers must earn our salaries, rather than just turning our jobs over to algorithms and up-fronts.

So instead of trying to count how many people view a branded tweet so that we can compare impressions to TV and print, how about we count something related to engagement? On Twitter, this would be the number of people who sign up for a brand’s Twitter feed, click on a brand-related URL through Twitter, mention the brand in Twitter posts, or retweet something about a brand. These are all examples of customers choosing to engage with a brand and share it with their friends. These activities (note the root “active” versus “impression”) show times when someone is consciously, choicefully dialed into your brand.

And, of course, we could develop similar metrics for traditional advertising. We could count the number of times people subscribe to your commercials on their TV sets, or how many people bring in print ads and hand them to their friends. Wait a minute: You can’t do that. No one does that. Which is exactly the point.