Here’s an example of how even the biggest Big Data can let you down, fresh from last week’s New Scientist magazine.
Obama and Romney’s recent election campaigns took targeting into account, using clever algorithms to slice the electorate into segments then sending each segment a specially targeted message.
But counter intuitively, research by a team at Yale and the University of Massachusetts proved that the micro-targeting did more harm than good.
- Targeted messages offered little or no advantage over general appeals
- When a voter got the wrong message they were much more likely than average to ditch the candidate in question
85% of people don’t want to see targeted adverts
At the same time a piece of research by the University of Pennsylvania found that a whopping 85% of people don’t want too see tailored adverts, full stop. Logic says they should. So what’s going on?
Who knows. It’s probably because humans don’t act logically. When faced with a decision we’re usually the diametric opposite of logical: unfocused and contrary, driven by emotions, hangovers, the weather, the common cold, our favourite team’s performance, the state of our relationships, bad hair days…
Big Data? Hm
Talk is easy. Results speak louder. There’s a huge amount of fuss being made about big data right now. But in a marketing context it makes sense to take it with a nice, big pinch of salt!
When you’re neck deep in your industry sector it’s easy to assume you know instinctively what key terms prospects are using to track down products and services like yours. But guessing key phrases can send you down the wrong track.
Here’s an example. As a freelance copywriter I could assume that loads of people a month are searching Google using terms like freelance copywriting and freelance copywriter. And right enough, a decent number of them are.
But a spot of key term research reveals tens of thousands more searching for a freelance writer, presumably because while the word ‘copywriter’ is used extensively in the creative industry, it’s relatively rare in everyday life.
Why does it matter? If, for argument’s sake, 500 people a month search for a freelance copywriter but 25,000 search for a freelance writer, it’s obvious which term has the most potential. ‘Writer’ wins hands down on volume.
On the other hand a little subtlety is required here. ‘Freelance writer’ is a much broader term. People searching Google for a freelance writer might want a novelist, journalist, biographer or ghost writer. Whereas people searching for a freelance copywriter want exactly that. They’re more tightly targeted, more likely to click through to my site and more likely to buy my services.
I decided long ago to concentrate my on-site SEO on the less popular terms first because they’re low hanging fruit, an easier and faster SEO win. Then chase after the term ‘freelance writer’, taking it slowly and treating it as a challenge.
Ranking well for such a popular term is going to be an uphill struggle without resorting to jiggery pokery. And I’m hopeless at link building! It might not happen. But it’s worth a punt, so here’s what I’m going to do:
- start using the term regularly in this blog
- think about creating a quality landing page to cover the term
- consider adding a new web page and / or blog page, bearing in mind I don’t want to repeat myself or bore visitors
- mull over the relative merits of adding a pertinent ebook, white paper, .pdf or whatever
- see if I can buy a strong url including the term, on which to build a new site from scratch and drive fresh traffic my way
- decide if and how social media can play a part in the process
- get my content and creative ducks in order
- make a solid plan
- diary to work on it for an hour a day over the next few months
I’ll keep you posted.
Despite the vast amount of data sites like Facebook hold about members, the social network’s data-driven adverts are still way off target.
OK, I may be in my forties and engaged to be married. But that doesn’t mean I’m into anti-ageing products. I don’t want a massive great meringue of a wedding dress either. Nor am I interested in celebrities or concerned about my weight: I couldn’t give a stuff how much blubber Cheryl Cole has lost!
These days marketers have access to ‘big data’, which by rights should make targeting offers tightly to people’s needs, preferences and lifestyles much easier. But in real life, it doesn’t. The conclusions they come to are still far too simplistic.
In reality Facebook’s efforts are no better than thirty years ago, when data driven targeting was the direct marketer’s holy grail and we only had postcode, sex and buying history to play with.
If Facebook filled my account with adverts for stuff I’m really interested in, things like 1950s German art pottery, ’60s and ’70s oil paintings, antique rugs, craft materials, tickets for Radio 4 comedies, garden stuff, wood carving gear and good books, I’d be a happy bunny and would probably click through. That’s what I’d call targeting!
I fell into direct marketing more or less by accident back in 1989. Part creative, part logical, the bit I found most exciting was the concept of targeting: tailoring your offer to a carefully-chosen bunch of people who should be more likely than average to respond.
Targeting makes sense in principle. For example, I love gardening. So when I get an offer from a gardening-related business, by rights I should be more likely than average to buy. On the other hand I don’t drive – I never have – so firms who send me information about breakdown services are on a big, fat hiding to nowhere.
While it’s a logical concept, we’ve been banging on about targeting for more than two decades. So why is effective targeting still the holy grail for so many marketers, more than twenty years down the line?
I suspect targeting might have a relatively small part to play in marketing success. It’s handy. It helps maximise your chances of conversion, along with numerous other common sense factors. But it ain’t – and will never be – a magic marketing silver bullet.
Why? Economists are currently busy revising the world’s financial models, taking human frailty and lack of logic into account for the first time. Perhaps it’s time to admit that it’s equally difficult – if not impossible – to accurately second guess individuals’ buying behaviour. While I love grubbing about in the garden my buying triggers are multi-factoral, involving much more than just a general need for garden-related stuff. However fantastic your offer, if I don’t need or want it at the point you contact me, for whatever reason, I won’t buy.
Having said that, there’s a shiny new kid on the block. ‘Big Data’ is here, born of the mind-boggling amounts of information collected online. Fingers crossed the sheer volume and depth of data will help marketers unravel fresh information about what motivates us to buy. But I won’t be too surprised to see another twenty years of self-congratulation and pseudo-science, prettily wrapped around a tiny core of common sense just big enough to perpetuate the myth.
Yes, targeting helps. Without doubt it’s better than nothing. But will Big Data deliver astounding practical insights that marketers can use to increase sales? We’ll see.
We tend to click on stuff we like, approve of, enjoy and are familiar with. Search engines assume we want more of the same, which has its advantages. But the downsides are clear. We’re either being deprived of a whole load of potentially interesting stuff because it’s demoted and we never see it. Or being deluged with so much of the same kind of stuff we get stale, bored and frustrated.
Facebook does it too, applying invisible algorithmic editing to what we see. As does Ebay, which recently provided a case in point. I donate every time I buy something and Ebay throws up an animal charity button during the process. Fine so far. But when they started tailoring the ads that appear next to my Ebay searches to animal charities I disabled the targeting straight away.
How come? Having to scroll past a steady stream of charity banner ads showing horrific images of animal cruelty was sickening. So despite Ebay’s best efforts it was an example of targeting gone wrong. I still give to animal charities. But there’s no way I’ll authorise Ebay to tailor ads for me again.
Is the net’s personalisation a bad thing? Are we missing out on the richness and variety of information online because search engines and so on filter it so heavily… and so heavy handedly? It’s a hot debate right now and the jury’s still out. But if I was God of Google I’d give people the choice of switching filtering on or off. And I’d be very wary of making sweeping generalisations and assumptions. We humans don’t work logically like algorithms. We’re natural explorers – we enjoy roaming free-range!
Here’s some food for thought on the subject: http://www.webpronews.com/facebook-google-filter-bubble-2011-06#comments