I love Twitter. I really do. Even when I have a beef with it, I recognize that it is an amazing tool connecting people and businesses and cultures and organizations and thoughts and ideas and revolutionaries. And it was built by a small conclave of guys who had no idea what it would grow up to be.
But there is a gap in our measurements, and it’s one that Twitter might make some real money while fixing the most bogus metric of all.
Blind Man’s Bluff
When you send out a batch of marketing emails, you know an awful lot about your audience. You know how many addresses were on that list — how many bounced — how many went unopened — how many converted, etc. You have a clear picture of what works and what doesn’t, and can do nice A/B testing to figure out some ironclad behaviorally-based guidelines. (Many of those crappy-looking webpages you come across at the end of a sales pitch look exactly that way because they work. Pretty doesn’t pay the bills.)
With Twitter, you have something called “Potential Reach.” The Potential Reach of a Tweet is defined as the theoretical maximum number of users who might have encountered it within their timelines. (Let’s ignore aggregate and unique instances for the moment.)
A Tweet from an account with 50,000 followers, in theory, would get 100-times the exposure of an account with 500. There are still other variables at play, like how focused that audience is on your topic, how many accounts those people are all following, time of day… it gets rather messy.
And it doesn’t have to.
You see, all of the metrics that marketing folks have at their disposal are dependent upon assuming apples equal apples in Potential Reach. You compare engagement numbers for various messages. You use link shorteners with Google Analytics campaign codes to slice and dice every last bit of divination as to the effectiveness. You Tweet at 4pm on a Monday, and 4pm on a Tuesday, and you see a five-percent difference, so one day is obviously better, right?
Well, it is if you think Potential Reach is a constant.
What Twitter Isn’t Selling You Yet
The dirty little secret that Twitter isn’t sharing is just how many people did “see” it. The current menu of Promoted offerings will measure engagement: Clicks, Retweets, Follows. But it doesn’t tell the truly data-driven marketer what she wants to know:
How many of my followers accessed the API to receive my Tweet?
Think of the feature that Facebook offers in Groups, that allows you to see how many people have seen a post. For Twitter, that might be an expensive proposition for coding and architecture and bandwidth, but from a data standpoint, it’s not something that has to be delivered in real time. (Facebook’s analytics for its pages are notoriously slow.)
Let’s suppose a particular Tweet – sent to my potential audience of 50,000 – was only accessed by 325 users. (I might get all frowny about the seemingly low number.) I could compare a number of Tweets over time and learn when my audience is “awake.” I could frame questions of effectiveness with regard to percentages instead of a flat metric.
Solving Two Problems
It’s all about the Clicks, you say? I totally agree. But maybe a lot of what you think drives those clicks is just wrong? What if an average message got the fluky benefit of a high spike in Real Reach, and tested better than a superior message? What if you tweaked your writing based on inaccurate assumptions?
It’s a classic statistical dilemma, trying to solve for two factors at the same time. And it is unnecessary, if Twitter gets smart.
So, Twitter: Sell us the analytics to truly know what got seen, and what didn’t. Charge a very pretty penny for it. Charge by the month, and make it steep, because firms will use that month to get a lot better about how they use Twitter for marketing. They will know when their users are really plugged in, and know which messages better resonate. That will be better for them, better for the userbase, and better for your own Promoted Products.
How much would you pay for Real Reach numbers?