Chances are you’ve received some form of targeted marketing today. An ad popped up on a site based on the cookies in your browsing history. You received an email suggesting other products you may like based on previous purchases with an online store. Even in the brick and mortar world, that customer loyalty card of yours is building up a detailed profile of your buying patterns.
There’s more data collected today than ever before, and the quantity of information out there is staggering. While marketers struggle to keep up with the volume of information available on their customers, there’s a real hunger to capture it and use it to create smarter, ever more targeted marketing campaigns.
In its 2015 Data Driven Marketing and Advertising report, GDMA interviewed close to 3,000 marketers and industry specialists from around the world. Among other insights, the research revealed widespread interest in the use of data among professionals — and this interest is growing.
Source: GDMA 2015
Data is clearly a major, if not the key factor in marketers' strategies. But are they wise to emphasize data so highly?
Many of us will remember the cautionary tale of Target.
The New York Times ran an article in 2012 about data-driven marketing at the company. Target had collected large amounts of data on its customers, and was able to identify when customers were in their second trimester of pregnancy based on their purchases. They then began direct marketing campaigns with suggestions for nursery items, diapers, lotions and the rest of the paraphernalia related to new-borns to those mothers-to-be their data had identified.
This led to the unfortunate case of a father in Minneapolis visiting his local Target brandishing a catalogue sent to his high school age daughter — advertising maternity clothes and other products for babies. The father was annoyed that Target seemed to be encouraging his daughter to get pregnant, yet was embarrassed to learn later that the store had in fact accurately guessed her pregnancy before he had, based on her buying patterns.
So with all of the data currently available, surely we can make our marketing more targeted than even, er, Target?
Not so fast. As Tim Harford points out in this Financial Times article, the truth is rather more mundane. Although Target got it right with the Minneapolis teen, this is a case of false positives. What we don’t hear about are all the thousands of women who had bought certain products, and received the targeted catalogue but who weren’t in fact pregnant after all. Perhaps they just liked lotion A, or they were buying diapers for a relative, or whatever else. Target in fact knew they had to mix in non-pregnancy related products into those catalogues because of the high chance they would, in fact, get it wrong.
While data can help identify larger patterns we might not have spotted, it’s got its limits. Let’s have a look at some of these, before exploring how we can go beyond our data dependence.
1. Data can’t tell you why
Data tells you that 15 year old boys in California like skateboarding. Data tells you that women aged between 28 and 35 visit your mortgage advice website. Data tells you that rural dwellers buy more pesticide than city dwellers. Data can tell you a lot of things, but it can’t tell you why. Of course, explanations are often pretty common sense (which is, in itself, a drawback of data). The point is, statistics are only good for finding correlations between certain factors, yet they don’t ever tell you about the complex reasons why certain events occur.
In fact, to really understand why, say, 15 year old boys in California like skateboarding, data is pretty mute. It can’t tell you anything about the history of the sport, how tastes and interests were influenced by a whole range of factors from pop music, clothes manufacturers to video games and beyond. Understanding these, and using them to drive sales, is a whole different trick.
2. Data can’t deal with non-numerical problems
If we based all our choices on data, the world would be a very cold place. In a New York Times article, David Brooks tells the story of the CEO of a bank with a presence in Italy. The country’s economy was doing badly, and all the data was saying that it would be wise to withdraw the bank’s operations from Italy. While the CEO took those statistics into account, he finally decided to stick it out; he didn’t want the Italians to perceive his bank as a "fair weather friend."
Data is simplistic; it can’t tell you about trust, values or relationships — the things which marketers want most with their customers.
3. Data creates more noise
We have data upon data upon data these days. We can know ever more about our customers, their age, their origin, ethnicity, tastes, social class, address, means of transport, even the kinds of music they listen to and websites they visit. But very often, this just creates more noise.
What we as marketers really want to know is what customers want and how to provide it to them. Having vast quantities of data helps, but can feel like a waste of energy. Unless you’re going to create targeted adverts for literally every customer you provide for, surely a handful of well-designed personas is enough?
4. Sampling error
A major problem with big data in general is that it presumes that N=all. That’s to say for example, it presumes that all the people who use your loyalty card are all the people. This is clearly not the case — you will have far more customers who don’t have your loyalty card, or who don’t use it every time they shop with you.
If you base all your marketing decisions on data, you’ll miss the many customers who slip through the net, all the people who aren’t yet customers and so on.
5. It’s only useful if there are actual statistics
Data works fine if you’ve got the figures. However, so many of your customers’ choices simply aren’t available statistically: their impressions of your brand, their feelings when they are in your shop, their reactions to the layout of your store or website. Data might tell you that X number of people visited your landing page and followed through to your website, but it won’t tell you what they liked (or didn’t like) about it, what would have made it better and so on.
6. It doesn’t understand complex behaviors
If you analyzed the amount of time you spent with different people throughout the week, the obvious conclusion would be that your work colleagues are more important than your family, friends and loved ones. You spend more time with them, so clearly they’re the most important to you, right? Obviously not.
In the same way, data really can’t tell you much about the complex emotions, feelings and impressions your customers have about your brand. If you based your marketing on pure statistics, not on factors such as values, emotion and other hard-to-quantify dynamics, you’d definitely miss the point of marketing.
What To Do?
Data clearly has its limits, yet it’s certainly not time to fire your statistician. Statistics should certainly be part of the toolkit, but need to be combined with other more traditional research methods too.
Use interviews and focus groups to understand why customers behave the way they do. Use eye-tracking methods to learn how people view your publicity. Simple observation can tell you a lot more about how people visit your store or website than any data, and might well reveal a lot more about customer journeys than customers themselves are aware of.
As with any new technology that appears revolutionary, big data has often been heralded as the future, the be all and end all. It would sweep aside all traditional methods for understanding customers and solve all our problems. This isn’t the case. While useful, it, like any other source of information, has its limits.
Data definitely has its place, but its usefulness should not be overestimated. When it comes to marketing, data is a little like online dating. You might be able to find a bunch of people of the right gender, age, location and anything else. But nothing beats actually speaking them to find out if they’re the right one for you.