A month ago we posted a thoughtful piece on the “Big Data” meme-of-the-month currently Pac-Manning its way across the marketing (and IT) landscape. Tons of bits and bytes, looking for a problem to solve, as we said then.
To point out how overengineered Big Data is, we also said this: “Most marketers can’t even tell you how many website visitors they had yesterday.”
Today we back that up with a true story.
Background: Lairig Marketing helps a client rethink and execute a customer e-newsletter. It is emailed to tens of thousands of customers, with a headline and two-sentence abstract describing each of three articles. For some reason, a digital person decides to put social media links (Facebook, Twitter, et al.) in the email alongside each article’s abstract.
Expectations are that customers will link through to read one or more of the full articles on the client’s website, at a rate higher than prior newsletter issues.
We track the email data after a few days, eager to check the newsletter’s performance. Counts of “clicks” show that three-hundred or so people have linked through to article #1, nearly four-hundred to article #2, and 150 to article #3.
Then this: nearly 2200 readers clicked on the “Tweet This” link for article #1, 2150 for #2, and 2250 for #3.
Dumbstruck that more than one reader would have retweeted a 50-word article abstract FROM AN EMAIL, we asked the email vendor to recheck the data (to protect the vendor, we’ll only say it is a four-letter name beginning with the letter “Z”).
The data come back with the “tweet” numbers essentially divided by 2. Some sort of “mistake” we are told. And, by the way, if we want an updated set of numbers it will cost us additional out-of-pocket.
Net net: we are being held hostage. By sh&t data. We know the “tweet” numbers should be zero. So what faith should we have in the rest of the data?
A simple, but illustrative example of what will happen in about a million businesses today across the U.S., and it promises to cut Big Data at its knees. Imagine if we responded in “real time” to these 3000 phantom clicks. It’s why Big Data will have us firing nuclear missiles on Canada soon.
Until Little Data is clean, Big Data will be a flop.