How to Use Statistics in Marketing

Statistics are often used in marketing to add authority and act as proof of a claim. We love the specificity of figures too: “A 250% increase in turnover” sounds so much better than “a big increase in turnover” doesn’t it?

But the numbers don’t always add up.

Anyone remember the Whiskas adverts in the 80s? We were told with unwavering authority that “8 out of 10 cats prefer Whiskas”.

Did you ever stop to think about that?

How does a cat food manufacturer know that cats prefer their brand over all others? Have they surveyed them? Asked a representative sample?

Isn’t it just a bit too convenient that they preferred Whiskas. Would Mars have used the research if they discovered that only 2 out of 10 cats preferred their brand?

It seems unlikely.

The truth was less cat-chy

And – no great surprise – things weren’t quite as clear as they appeared.

What Whiskas really meant (and subsequently changed the line to) was: “8 out of 10 owners who expressed a preference said their cats preferred it”.

Not as strong is it.

And even that is a bit misleading as the researchers asked cat owners if their cats like Whiskas and, of those who answered, 80% said yes.

How many people did they ask and how many answered? And of those who responded, who really knew their cats’ preferences?!

I’m pretty sure my dog thinks dead sheep, dog food and tuna steak are all equally delicious.

Whiskas carrying out research
Copyright 2018 Mars

The truth didn’t matter. Everyone who saw the original Whiskas advert believed it and the campaign was hugely successful.

Later changes made no difference to what anyone thought about Whiskas and to this day everyone over a certain age has “8 out of 10 cats prefer Whiskas” embedded in their brains.

Whether deliberately deceitful or accidentally unclear, the ads did what they were intended to do.

Facts are stubborn things, but statistics are pliable
― Mark Twain

It’s easy to manipulate people with statistics and there are all sorts of cognitive biases at work when marketers use numbers.

Harnessing cognitive biases

Why do men buy diamond engagement rings for their other halves? And why does just about everyone on the planet associate getting engaged with diamond rings?

Because the people who sell diamond rings said we should!

De Beers pushed the association and also came up with the idea that people should spend roughly a month’s salary on an engagement ring.

In possibly the most successful ad campaign of all time, copywriter Frances Garety (from agency N.W.Ayer) came up with the famous line “diamonds are forever” and connected life-long love and marriage with diamond rings (before then ruby and sapphire engagement rings were popular, only about 10% of people bought diamond ones).

The campaign also did some amazing ‘anchoring’ by saying that men should spend a month’s salary – a twelfth of their annual income – on an engagement ring. And people did it.

Years later, in the 1980s, De Beers decided that wasn’t enough and upped it to two month’s salary. By moving the financial ‘anchor’ upwards they got people to spend more simply by telling them they should (and who could deny their betrothed?!).

Up to you whether you think it was a stroke of genius or highly manipulative but it worked.

How to use statistics in marketing

Here are a few tips for using statistics in your copywriting and some pitfalls to avoid.

When to use big numbers

Most of us are more like sheep than cats and we need to belong, to feel part of the group, which is why we – marketers – often use social proof to encourage readers to do something or buy something. There are lots of ways of tapping into this human need: through endorsements from clients, friends, celebrities, influencers and experts in the form of case studies, testimonials, TV adverts, sponsorship deals, articles, social media, ratings and reviews – and using big numbers.

Here’s a great example.

Basecamp

Basecamp is an online platform that helps you manage projects, stay on top of what needs doing and communicate with your team.

This is their homepage:

use large numbers as social proof

 

They use the number of businesses who signed up last week – 4,069 – to encourage new businesses to do the same, and it’s persuasive. As soon as you see how many other people have just signed up you assume that Basecamp is a reputable company, provides something useful, and you want a piece of it too.

If, instead, Basecamp’s homepage asked you to join two other businesses you’d draw very different conclusions.

TIP ONE: Use big numbers as social proof.

When to use small numbers

If you want to motivate someone to join something, do it with a big number, if you want someone to care, do it with a small one. Stories about one person or a small group tug at our heartstrings in a way big numbers don’t.

BookTrust gets children and families reading and reach “3.4 million children across the UK with books, resources and support” every year. That’s awesome you might think. I do, which is why I’m a regular supporter and happily donated to their Christmas appeal last year.

In December, they raised money to send books and other goodies to looked-after children all over the UK. They included a postcard in each parcel that the children could write and send back if they wanted to.

Here’s one of them:

use one person's story to engage emotionally
BookTrust shared Tiffany’s postcard and a few others on their website and Twitter. I’d challenge anyone not to feel moved when they read them (they made me cry).

The cards show how much the kids loved getting the parcels and what they meant to them, it makes you feel that your donation has truly made a difference to someone’s Christmas.

And that’s the idea. Hopefully, lots of people read them, have the same reaction and donate.

Hearing one person’s story is much more persuasive and emotionally engaging than seeing a simple number. As Trevor Bragdon explains in this article about persuading with numbers, you need to connect readers with emotion first, then logic. If BookTrust had said something like, “1020 children really enjoyed receiving our parcels this Christmas” instead, it wouldn’t have struck a chord.

TIP TWO: Use small numbers to engage emotionally.

Translate massive numbers into something we understand

We find it very hard to visualise big numbers, which is why people often talk about the size of a football pitch or the size of Wales (why are so many things the same size as Wales?!).

One of my favourite examples is used to explain how long humans have been on earth.

Try and picture Nelson’s Column in Trafalgar Square (it’s 52 metres high) with a huge pile of 1p coins next to it. If the height of the statue represents the time earth has existed, the very last penny at the top of the pile is how long there’s been life on earth.

Then imagine someone (in a crane?) very gently placing a sheet of paper on top of the coins. The thickness of that piece of paper is how long humans have been around.

It’s still hard to relate to that sense of scale but it’s a lot easier than trying to think about earth existing for 4.543 billion years and humans being here the last 0.004% of that. Try doing that calculation in your head!

TIP THREE: Express huge numbers in a way we can relate to.

Is it ‘1 in 10’ or 10%?

According to Trevor Bragdon (cited above), “we think people, not numbers”.  We tend to relate more to statistics presented as ‘1 in 10 people in the UK’ rather than ‘10% of people in the UK’, despite them being the same amount.

Depending on what you want to achieve, you can choose which way to write out your numbers.

HMRC uses nudge theory [this blog – How to Use Nudge Theory in Copywriting – explains what it is] to get us to pay our taxes on time with ‘nudges’ like “9 out of 10 people pay their taxes on time”.

They sent out a letter that included that exact phrase to 200,000 people who hadn’t paid and it raised an additional £4.9 million in tax revenue.

In general, ‘1 in 4’ is easier for most people to understand than 25% as percentages are abstract. We hear that 40% of marriages end in divorce and it seems like nothing to do with us. Someone tells you 4 in every 10 married people will get divorced and you start looking around at your friends and family.

Yet there are times it makes sense to express a figure as a percentage rather than as a number. Say your widget sales increase from 100 to 200, it’s a 100% increase. You’re still only selling 200 so it might be better not to mention that but a 100% increase sounds impressive.

The opposite can also be true. Imagine you have to put your prices up £50 from £250 to £300. If you talk about raising your prices 20% that seems like a lot. Pay rises aren’t 20%, inflation isn’t 20%. But you’re only charging another £50? That’s not very much money.

TIP FOUR: Use numbers or percentages wisely.

Be careful! Correlation does not mean causation

There are a couple of pitfalls to avoid when you use numbers in your copy too, the first one is that correlation is not the same as causation.

I wrote a blog about this a little while ago and my made-up example was that you don’t make it rain by carrying an umbrella… but it does seem to rain more when people are carrying them?!?

Another example I’ve come across a few times is that the rate of physical assaults rises when ice cream consumption goes up. Statistics show that when people eat more ice cream, there’s a higher incidence of assaults.

more ice creams eaten and more assaults in hot weather
Don’t be fooled. They’re not as innocent as they look.

But that’s not the whole picture.

People eat more ice cream when there’s nice weather. And when it’s hot outside, people get irritable and aggressive and are more likely to get into arguments and fights.

Eating ice cream and assaults are both correlated with the weather, not each other.

TIP FIVE: Don’t fall into the causation/correlation trap. Eating ice cream does not cause fights.

Be careful with scales on graphs

Another way we’re sometimes misled is with the scales used on graphs, so make sure yours make sense.

Here’s an image published by New South Wales government a few years ago about nurse recruitment in the health system:

example of misleading scale on graph

You can see that the numbers are increasing year on year, especially after 2011/11, and you’d be forgiven for thinking they’re going up much faster than they really are.

Between 2010/2011 and 2011/12 there’s an increase of 3168 nurses – or 7.3% of the total number (43,405). But it looks like a lot more.

And from 2011/12 to March 2013, nurse numbers go up to 47,500+. If they get to 47,500, that’s an increase of just 2% on 11/12 but it looks way bigger than that.

In the graph, 4 pink nurse icons represent 43,000 nurses (up to 2010/11). That 4 jumps to 32 pink nurses to represent 46,573 (2011/12), then to 40 pink nurses for 47,500 real nurses (to March 2013). You could argue it’s a bit misleading.

You see a graph and assume the scale is the same for each year and one nurse or row of nurses represents the same number. But on this one there’s roughly one icon for every 10,750 nurses to start with, then one for every 1362 nurses, then one for every 1188.

Of course, if you look at the scale and do the rough calculations in your head you can work out how big the increases really are, but if you just glance at it quickly you’d think numbers have sky-rocketed. Which is exactly what you’re meant to.

TIP SIX: Use a sensible scale on your graphs.

Statistics can add weight to copywriting as a way of backing up a claim or showing you know what you’re talking about. Even though 74.63% of statistics are made up on the spot we believe them and if you write a statistic in bold no will dare question you [honestly, that’s a thing, have a read about it here: How to Make Your Copy More Convincing (With the Appliance of Science)].

Hopefully these tips will help you use statistics effectively in your copy:

  • Use big numbers as social proof
  • Use small numbers or one person to engage emotionally
  • Use objects/sizes/numbers we can relate to to explain scale/size/numbers we can’t
  • Use numbers or percentages depending on what you want to achieve
  • Try to avoid confusing correlation and causation
  • Use a sensible scale on graphs

Thanks for reading!

Sally

What are your thoughts? Feel free to leave a comment.

This site uses Akismet to reduce spam. Learn how your comment data is processed.