Misconception of price comparisons
Price comparison campaigns are becoming increasingly popular and are often exploited as a powerful communication tool by many retailers, especially in maturing (or mature) markets. The advantage of comparison campaigns over quality advertising is their direct relation to measurable quantities. At least in principle. In actual fact, a realistic price comparison would be so complex as to be useless in advertising and the average reader would find it indigestible (in the case of quasi-specialist publications). As a result, retailers have a lot of room for manoeuvre when it comes to presenting price comparisons casting themselves in a favourable light and the media usually stick to a simple model which is of no value whatsoever to anyone at all.
In order to understand why the available price comparisons are of practically zero value, let's take a brief look at how they are done:
- First, a basket of products is compiled from a wide range. The selection criteria give the first degree of freedom in influencing the comparison results. The criteria may be the overall market sales value, sales items, buying frequency within a specific time frame. We might pick the “top n” products, or “randomly” pick “n out of top m” using one (or a mixture) of these criteria.
- For the selected basket, the product prices are recorded in different shops. At any moment in time there is a mixture of regular shelf pricing and promotional offers. If really done randomly (and this is one big IF), and on a fairly big sample of products – the final result would also reflect promotional intensity at a given store which is in fact of some benefit to the shopper. However, sensitivity to promotional share also makes the comparison results vulnerable to any intentional fine-tuning.
- Once the price collection is complete, the basket price is calculated. The most common method is to simply add up the prices of all the products. The total sum gives the score, the score creates the rank, end of story. Right? Well, to me it’s not really fair to say that customers buy one piece (or unit) of every product in a unit of time. In a hot summer I might buy a 250g jar of coffee, and 10 bottles of water per week. Even if the water is much cheaper than the coffee, I would spend a similar amount of money on both (so any % difference in price matters to me!). In a price comparison basket there would be one unit of each, making the water almost irrelevant.
I honestly do have a problem with this approach. Mainly with what such comparisons actually communicate (SPOILER ALERT! Nothing really). The composition of the basket is based on averages, and there is nothing like an average customer. The selection criteria are not those used by any customer. I choose certain products simply out of personal preference not because they feature in the top sellers ranks! I do not buy one piece of every product every time. Nor do I buy the same products every time. The contents of my basket would differ depending on my income, my diet (vegan or burger buff?) or on the occasion (everyday meals vs. BBQ or wedding). The composition is not just the range of products, but also how much of each I buy. The straight sum of prices does not even pretend to be correct. Why? Well, suppose you are offered: (a) a salary of 1000EUR for the first month and then 2000EUR per month for another year, or (b) alternatively 5000EUR for the first month and then 1000EUR per month for another year. Price comparison common “logic” would recommend picking the latter, just because the numbers are relevant, while the quantity is not. Adding a volume weighting based on average retail sales would again apply to a non-existent customer, but it would be much better.
Is there then any hope for a comparative tool that would support consumers in choosing the best shop for them? Actually there is. It’s simple enough to tailor the message more accurately and to become a little more innovative. A step towards the right solution would be the “lifestyle” basket comparison, as it determines shopping habits and therefore the “shopping basket”. A bodybuilder consumes litres of water and lots of protein whilst avoiding sugar and alcohol. Their basket would be completely different to that of a party animal who drinks above average volumes of beer and buys frozen ready-made meals. A sample of this approach is presented in the “Which store is the cheapest?” example, but it’s not perfect as it’s based on an arbitrary selection of products, while lifestyle has an impact on not just how much is purchased, but to an even greater extent, what goes into that basket. In a perfect world, the base for comparison would cover a huge sample of stores and products, and anyone could prepare a personal profile of to-be-consumed quantities of each of the products (i.e. a weekly/monthly shopping list). This personal weighting would provide a “the cheapest store for you is…” answer. Or even better “the cheapest store for you within 10 minutes drive from your home is…”. Or how about “… but it would take 30 minutes more, and cost 10EUR less to buy these products in Store A, and those in Store B”.
It’s not just a vision of a perfect world. As in many other business areas, digital easily overcomes physical material barriers. Smartphones with bar code scanners and price recording software already provide a powerful tool with which to build the huge store – product - price database. Once “price collecting social media services” become popular, personalized basket price information will be accessible, stable, reliable, transparent and furthermore, offer value for shoppers. And it’s inevitable. This poses new challenges for retailers, as they will be forced to concentrate much more on service quality and assortment than on prices and price communication if they are to meet their customers’ individual expectations. The alternative would be of course to become the cheapest on every product. This doesn't seem to be the most sustainable choice, does it?
Which store is the cheapest?
To fully appreciate the lack of value in classic price comparisons, let’s consider the example illustrated in the table below. The brands and quality of unlabelled products are the same across stores, so we are comparing apples with apples. A quick look at the result…so we should buy at Store B, and avoid store A.
|Store A||Store B||Store C||Store D||Store E|
|Flour 1 kg||2.22||2.35||2.29||2.49||2.85|
|Sugar 1 kg||1.65||1.55||1.99||1.95||2.10|
|Box of 10 eggs||3.69||4.45||3.89||4.65||4.85|
|Milk 3.2% 1L||2.25||2.35||1.75||2.05||2.15|
|Ground coffee 500g||20.99||16.69||14.99||19.99||21.59|
|Bottled beer 0.5L||1.69||2.53||2.39||2.35||2.25|
|Bottled water 1.5L||1.45||1.60||1.89||1.55||1.92|
Correct? Because the overall score is the sum of all the recorded prices, the price is simultaneously the weighting of the basket. It turns out that the most significant products are vodka and ground coffee, while bottled water or milk are 10 times less relevant!
|Store A||Store B||Store C||Store D||Store E||Weighting in the basket|
|Flour 1 kg||2.22||2.35||2.29||2.49||2.85||2.6%|
|Sugar 1 kg||1.65||1.55||1.99||1.95||2.10||2.0%|
|Box of 10 eggs||3.69||4.45||3.89||4.65||4.85||4.7%|
|Milk 3.2% 1L||2.25||2.35||1.75||2.05||2.15||2.3%|
|Ground coffee 500g||20.99||16.69||14.99||19.99||21.59||20.4%|
|Bottled beer 0.5L||1.69||2.53||2.39||2.35||2.25||2.4%|
|Bottled water 1.5L||1.45||1.60||1.89||1.55||1.92||1.8%|
This does not make sense unless you always buy the same number of cartons of milk as bottles of vodka. I guess such a shopper might exist but I bet they do not buy 1kg of flour as frequently… To make the comparison more realistic let us assume we have the monthly consumption levels of the average shopper (I know there isn’t one), and three types of shoppers with varying lifestyles:
- Shopper A – a beer lover, likes high fat food, but no sweets
- Shopper B – has a sweet tooth, doesn't drink alcohol
- Shopper C – loves dairy products, visits the gym often
Monthly consumption might look something like this (in units presented in the price comparison table, e.g. 10 x 10 eggs = 100):
|Average||Shopper "A"||Shopper "B"||Shopper "C"|
|Flour 1 kg||2||0.2||4||0|
|Sugar 1 kg||1||0.2||3||0.2|
|Box of 10 eggs||3||10||10||15|
|Milk 3.2% 1L||10||2||5||10|
|Ground coffee 500g||5||3||5||3|
|Bottled beer 0,5L||15||60||0||10|
|Bottled water 1.5L||10||0||2||30|
If we multiply each price by the monthly consumption we get monthly expenditure on these goods. For the (non-existent) average shopper the result would be:
|Average||Store A||Store B||Store C||Store D||Store E||Weighting in the basket|
|Flour 1 kg||4.44||4.70||4.58||4.98||5.70||1.4%|
|Sugar 1 kg||1.65||1.55||1.99||1.95||2.10||0.5%|
|Box of 10 eggs||11.07||13.35||11.67||13.95||14.55||3.6%|
|Milk 3.2% 1L||22.50||23.50||17.50||20.50||21.50||5.9%|
|Ground coffee 500g||104.95||83.45||74.95||99.95||107.95||26.4%|
|Bottled beer 0.5L||25.35||37.95||35.85||35.25||33.75||9.4%|
|Bottled water 1.5L||14.50||16.00||18.90||15.50||19.20||4.7%|
So, considering average monthly consumption we should mark Store C as the cheapest, Store A as the most expensive. At least which is the most expensive is clear. However, if we repeat the same operation for our shoppers we get the following:
|Store A||Store B||Store C||Store D||Store E|
Here we can see that Store A – the most expensive – turns out to be the cheapest option for the beer lover. Store B is not the cheapest for the average shopper nor for any of the selected shoppers. Store E emerges as the most expensive for shoppers A, B and C but not for the average shopper. So, which store is the cheapest? Well, that rather depends on who's asking.
Image: Apples & Oranges - They Don't Compare by TheBusyBrain