“Oh! But he was a tight-fisted hand at the grindstone, Scrooge! a squeezing, wrenching, grasping, scraping, clutching, covetous, old sinner! Hard and sharp as flint, from which no steel had ever struck out generous fire; secret, and self-contained, and solitary as an oyster. The cold within him froze his old features, nipped his pointed nose, shrivelled his cheek, stiffened his gait; made his eyes red, his thin lips blue; and spoke out shrewdly in his grating voice. A frosty rime was on his head, and on his eyebrows, and his wiry chin. He carried his own low temperature always about with him; he iced his office in the dog-days; and didn’t thaw it one degree at Christmas.”

MetaMetrics blog | A Christmas Carol for our time

Ebenezer Scrooge: I have always been an Alastair Sim manmyself 

 Very festive. And you know what happened to Scrooge of course, his icy heart melted, and he became the kindest most generous of men all year round, not just at Christmas. 

Isn’t that nice? 

Nobody likes to think of money, especially at Christmas. After all, it’s the thought that counts. Except, we would argue, if you are buying TV ratings. You see, the thing is that normally when things double in price you just get half the amount 

Like, say, a bag of humbugs. If humbugs start at £10 per kilo and, due to poor hedging of sugar futures, the price rises to £20 per kilo, your £5 now only buys 250g, whereas previously it bought 500g. Inflation sucks. 

MetaMetrics blog

small bag of humbugs, costing proportionately less than a large bag 

The odd thing with TV is that it doesn’t work like that. If you double the cost, you can actually get LESS THAN HALF of the effect. That would gladden old Ebenezer’s heart. 

But how could that be the case? Surely if we spend £1m at £2,000 per rating we get 500 ratings. If the cost goes up to £4,000 per rating we get 250? Well, yes this is indeed true. 

The problem comes with our old friend – no, not Tiny Tim, but diminishing returns. Imagine, if you will, a fairy-tale world of peace on earth, zero carbon emissions and linear returns. If you were to halve the amount of media you deployed in your campaign from, say, 300 ratings to 150 ratings, that would indeed halve your media impact. Such a world is depicted in the chart below, where halving deployment reduces the media uplift (i.e. impact) by exactly half, from 60% to 30%. The problem is that such a world doesn’t exist (the clue’s in the title). 

MetaMetrics Blog - A Christmas Carol for our TimesMetaMetrics Blog - A Christmas Carol for our Times

Instead, to speak momentarily of the facts, TV response often looks rather more like the chart below, in practice. The curve is S-shaped. This means that, as you’d expect, there is a saturation effect. If you see an advert for a well-known sofa retailer advertising their Xmas sale for the 175th time you are likely to throw your tin of Quality Street at the TV. Doubling spend does not double the impact. Twice as much is never twice as good (unless we’re talking about the big purple ones in Quality Street). See part 5 of our ebook “Guide to econometrics for the modern marketer” which you can download here. However, there is also something interesting that happens at the bottom of the curve. If you do too little activity, you just never drive your adstock up high enough to cut through in consumers’ minds.

MetaMetrics Blog - A Christmas Carol for our TimesMetaMetrics Blog - A Christmas Carol for our Times

Between these regions, there is a part of the curve where media deployment is just right and which we, for want of imagination, call the “efficient range”. 

MetaMetrics Blog - A Christmas Carol for our Times

Now, let’s say that, at a certain amount of spend we can buy enough ratings to get our adstock in this “just right” zone. 

MetaMetrics Blog - A Christmas Carol for our Times

There. That’s nice. 

Now let’s imagine our TV ratings double in cost. So, we can only buy half the number of ratings and therefore deliver half of the adstock. A bit like this. 

You might think that this deployment would deliver half the overall effect of the first campaign, but in fact the effect is only 37% of the first campaign. This is because the adstock languishes down in the bottom of the S curve and just never makes it up into the efficient zoneYou can see in the chart below how the second campaign – bought at double the cost, and therefore comprising half the number of ratings, of A – never makes it into the green efficient range. 

MetaMetrics Blog - A Christmas Carol for our Times

So, the Campaign suffers a double whammy in effect – where the efficiency effect on top of the additional cost significantly reduces the overall impact of the campaign. And, as the chart below shows with made-up, but realistic, figures, that efficiency effect can be quite chunky. Consider it a Christmas present from the TV industry. 

MetaMetrics Blog - A Christmas Carol for our Times

What this means in practice is that, not only must we be acutely aware of media costs, but we must work out their implications for the optimum deployment. Don’t just halve everything when cost doubles, but re-optimise your flighting to ensure you are getting the best returns on your investment at the new cost levels.

It is a counter intuitive finding, admittedly, but hopefully useful. After all, even as he went through his ordeals that night, Scrooge never forgot the importance of econometric modelling for optimising his media deployment:

“Spirit,” said Scrooge submissively, “conduct me where you will. I went forth last night on compulsion, and I learnt a lesson which is working now. To-night, if you have aught to teach me, let me profit by it.”

Happy Christmas!