Martin Lewis: I confess, I did a 'decoy effect' experiment on my Twitter and Facebook followers… and it may be why they're buying more Jaffa Cakes

The 'decoy effect' polls

I ran two polls, a week apart (to put some distance between them) – telling people it was an experiment, but not what the experiment was. Many tried to work out what we were testing, but no one put it together (or at least no one told me they did). In fact, the key wasn't either individual poll, but the combined result of both.

Of course, my polls aren't a statistically random sample of UK society, but the number of voters was pretty substantial, and so a nice indicator of how things works. Results were similar on Twitter and Facebook, but as Twitter had more votes and is easier to compile, I'll base the results on that.

Poll 1 – 23 June 2020

Imagine you and the family WANT the new hot TV subscription service just out, 'NetBox+ Prime'. The TV offering is constant, but you've searched and found THREE PROMO CODES all give access to a different number of HOT/TOP FILMS. Which is your best deal? (See the original post on Twitter and Facebook.)

Poll 2 – 3 July 2020

Imagine you and the family WANT the new hot TV subscription service just out, 'NetBox+ Prime'. The TV offering is constant, but you've searched and found TWO PROMO CODES all giving access to a different number of HOT/TOP FILMS. Vote split by male (M) and female (F). Which is your best deal? (See the original post on Twitter and on Facebook.)

NB: The gender split here was a deliberate red herring. Men were very slightly more likely to go for the higher price option, but that's irrelevant for what's being tested, so I've combined the result below.

The polls showed the decoy effect worked

The only difference between the two polls is that in the first there is a third higher-cost deal. It's clearly poor value at £20/month for 700 films, as you also have the option of paying £15/mth for 750.

Even so, some people went for it, probably due to a combination of retail snobbery (ie, assuming that the higher price must mean you're getting the very best, even though there's no tangible detail that shows that), clicking errors, or not paying attention. All of which happen in the real world too.

Yet what really counts here is the impact that throwing in that dud higher price offer had on the number of people plumping for the £15/mth deal. Let me isolate just that...

While many individuals still made the same choice, across the mass of people the inclusion of the decoy deal had a material impact on the total results. Nearly twice the number of people went for the £15/mth deal. My assumption is they just thought: "That's decent value compared to the £20 one".

So where do Jaffa Cakes come into this?

Not long after running these polls, I walked into a supermarket and spotted this.

As you can see, a twin pack of Jaffa Cakes is reduced from £1.80 to £1. And right there next to it is a single pack of Jaffa Cakes, also priced at £1. As Adam Oliver notes, many firms use the decoy effect, even though they may not know the theory behind it.

And while this may just be accidental, it's hard not to suspect the supermarkets know exactly what they are doing. If a twin pack is the same price, why would they leave out the single pack? Let's answer that via the lens of the decoy effect.

By leaving that box out, it sets a reference price that shines a bright light on the relatively cheap price of the twice-pack deal, so people say: "I'll grab those, that's a bargain." In fact, even though I was aware of this when I saw it, I still bought the Jaffa Cakes, because... well, I like Jaffa Cakes.

The next week I then saw a similar deal in reverse. Here, the single-pack price was reduced, while the twin-pack price was left as a reference.

A word from Adam Oliver

That's my bit done, but I've invited Adam Oliver from the London School of Economics to give his more scientific view...

"Standard economic theory makes a number of assumptions about how people behave, act, choose and decide. One of those assumptions is that each person attaches a value – sometimes called a utility – to any particular good or service that they face, and that the introduction of other goods or services cannot affect the value that the person attaches to the original good or service.

"What this means is that – according to standard economics – if a person prefers, say, good A over good B, then the introduction of another good – say, good C – cannot affect that person's preferences such that he or she now prefers B over A rather than A over B.

"However, as Martin has demonstrated with his Netbox example, the introduction of a third option into a choice set can sometimes affect people's preferences between two goods, or rather, in this case, services.

"This can happen when the service that is introduced makes the original least-preferred option suddenly seem like a very good deal. In these circumstances, the third option – say, option C – is called the decoy, and if it makes people switch their preference for A over B to B over A, we call this the 'decoy effect'.

"The decoy effect is not trivial. Companies can use it to increase substantively their revenues with virtually no increase in costs. Martin highlights this with his Netbox example, where if the company in question were targeting, say, one million subscribers, and Martin's figures held up in real-world choices, the decoy would generate almost an additional £1 million in revenue each month (from almost £11 million to almost £12 million).

"Some might say that companies should not be allowed to do this, but they do use decoys, and we as citizens, with some basic knowledge of behavioural economics, can perhaps protect ourselves against such practices."