When more data isn’t the answer for marketers
Just a generation ago, most of the really serious data-driven marketing work was still in the hands of elite brands and companies with a dedicated direct sales model. Now, even the smallest niche marketer has access to vast amounts of behavioral and intent data, and bears responsibility for mining that raw data ore into bottom-line riches.
But more data and all the advanced, enhanced visualization tools in the world can’t paper over creative gaps, strategic blind spots, or a simple failure to recognize changing value propositions and customer priorities. More data isn’t always the answer. The next time you’re tempted to throw more data at a problem, consider these strategic alternatives.
More context can beat more data
The difference between context and data can be seen on any good consumer credit score report. The score itself is just a number. Even a graph showing how that number has changed over time is still, again, just numbers. The real value comes from the narrative portions which explain why the score is what it is: because you’ve recently taken out a loan, or have a good payment history, or haven’t established a track record with a certain type of debt.
These narratives and explanatory signals are often missing from marketing insights because nobody has thought to ask for them. Models produce figures, and campaign automation tools send out offers based on score thresholds, and that’s that. But marketers can design much more interesting and nuanced campaigns if they know not just a raw score, but how the score came to be.
“Marketing professionals and agencies usually have very little visibility into what the models are doing, and just get scores or a bucket of categories from the data,” says Dean Abbott, co-founder and chief data scientist at analytics platform SmarterHQ. “But the real value is not just in knowing that someone is likely to make a purchase in seven days, but why they are likely to purchase.”
Solve interesting problems
Patrick O’Hara, chief strategic officer at Gyro Interactive, says that data-driven organizations are more likely to fall short on their ability to develop interesting hypotheses than they are to actually lack enough data to meet their goals.
“You need to develop a forensic approach to help you find things your competitors aren’t finding,” he says. “That’s a much stronger approach to marketing data.”
Don’t confuse the confidence a data scientist has in their model with a worthwhile model. “It’s almost always the case that a data scientist can build an accurate model,” Abbott says. “Can you fit that model into something that moves the needle for the company? You can build incredibly accurate, but completely irrelevant, prediction models very easily.”
Stop focusing on narrow possibilities
Data-driven marketing shines such a bright light on recent behaviors and signals that marketers have become fixated on developing hyper-accurate responses to a narrow range of behaviors and opportunities. “It’s a feedback loop. You optimize what you can do about a few behaviors, and it makes things slightly better, but you aren’t putting yourself in a situation to surprise and delight them,” O’Hara says.
Shoppers are increasingly aware that their browsing behavior can prompt brands to send a timely discount offer or incentive. It’s a move that overcomes an objection, but only in a transactional way.
“It’s not putting you in a position to surprise and delight them, to tell people not just what they want to hear, but to be serendipitous,” O’Hara says.
Don’t expect data scientists to be unicorns
Data scientists can’t all be machine learning geniuses with decades of experience interfacing both with customers and with IT organizations. Business analysts and IT professionals need to be part of the data-driven world in order to ensure that the right data is available to answer the right questions.
Going after data professionals with a head for business is important, but Abbott warns that an even bigger hurdle for the growing pool of fresh data scientists is the less-than-ideal state of real-world data. “People coming out of university programs have rarely worked with real data, which is far worse than they ever see in school,” he says.
Dream of what could be, not what is
Is your data keeping your creative vision rooted in the world as it is observed today, rather than the world it could be tomorrow?
“It’s a great moment for ad agencies and creative companies to find new ways to do things they are already good at,” O’Hara says. “It’s also a moment to take a page from Steve Jobs and stop asking so much what your customers want, because customers don’t understand what’s possible.”