The Miss­ing Mil­lions: How Super­mar­kets are Acci­den­tal­ly Giv­ing Away Millions

Gareth Har­vey — DECIDE


Gareth shares some of the research he does into prod­uct pric­ing in super­mar­kets and restau­rants, and explains why big­ger dis­counts aren’t always the best way to get more purchases.


By load­ing the video, you agree to YouTube’s pri­va­cy pol­i­cy.
Learn more

Load video

Full Tran­script:

Jo Ever­shed 0:00  

But first of all, we’re going to talk to Gareth about how super­mar­kets are acci­den­tal­ly giv­ing away mil­lions to our ben­e­fit, of course. Over to you Gareth!


Gareth Har­vey 0:12  

Hel­lo. So, as I said, I’m Gareth Har­vey. And I used to be a pro­fes­sor of con­sumer psy­chol­o­gy at ASHRAE in Gene­va. But now I am a direc­tor of behav­iour­al sci- try that, again, I’m the direc­tor of behav­iour­al sci­ence at Decide which is a mar­ket­ing and cre­ative agency. And unlike most of you guys, my world is not like­ly doing through we say the aca­d­e­m­ic research, but it’s tak­ing all the great find­ings that we have out there in the world. So things that you guys are work­ing on now. And to put it real­ly crude­ly find­ing a way to make mon­ey from it. So there’s so much great research out there, but it just stays in the aca­d­e­m­ic silos. So my remit is to try and read about that, and work out how it can be used for brands.


So my par­tic­u­lar inter­est is in super­mar­ket design. So I do a lot of work design­ing super­mar­ket lay­out. So if you can’t find the baked beans or the bread, it’s prob­a­bly my fault, or some­one like me. Or on the oth­er hand, I do a lot of work, which is eye track­ing and look­ing at prod­uct design. So mak­ing sure that prod­ucts stand out on the shelf.


Now today, I was going to talk a lit­tle bit about actu­al­ly how super­mar­kets end up giv­ing away a sil­ly amount of mon­ey. Now, if you go around any super­mar­ket, you’ll see all sorts of spe­cial offers. Lots of them that look some­thing like this, some will be two for five pounds, some­times you’ll be see things buy one, get one free, some­times 50% extra free. And if you look at these, some of them will be, shall we say? Not so great offers. So for exam­ple, some of them is it said at one and the bot­tom, one pound each, or two for two pounds. Peo­ple often think that this is a case of super­mar­kets being, shall we say delib­er­ate­ly try­ing to manip­u­late you. And the sim­ple truth is most of the time, it’s not that. It’s a sim­ple case of human incompetence.


Human incom­pe­tence explains a lot. But when we look at it, actu­al­ly though, there’s such a vari­ety of dif­fer­ent offers, and in effect, they’re giv­ing away more mon­ey than they actu­al­ly need to. If we think about it, it’s a sim­ple case. So what’s more effec­tive, buy one, get one free, 50%, extra free, half price, what grabs atten­tion. And actu­al­ly, a lot of the time when we actu­al­ly analyse data in the store, we realise it’s, the offer itself is kind of irrel­e­vant to a cer­tain extent. Con­sumers are so busy, are so busy try­ing to look and work out what stands out what’s on spe­cial offer.


And just to show how big an effect this is, I’ll talk about an exam­ple I did instal a num­ber of years ago, we’re work­ing in the alco­hol sec­tion, look­ing at beers, we col­lect­ed our base­line data. So we record­ed some­thing in the region of about near­ly 10,000 shop­pers, going down the alco­hol sec­tion. And then we made an inter­ven­tion, we made some changes. And we want­ed to see, could we increase the sales of a par­tic­u­lar type of beer. And sud­den­ly, after our inter­ven­tion, a beer we did noth­ing to we did no work to, it’s sales increased by some­thing like 50%. And we thought, what the heck has hap­pened here. And actu­al­ly it was a sim­ple case of peo­ple noticed the one prod­uct that was by a sign on the shop shelf. That sign it sim­ply said shoplifters will be pros­e­cut­ed. Noth­ing to do with actu­al any­thing to sales. But it grabbed peo­ple’s atten­tion. Peo­ple looked at it, and they end­ed up see­ing it. And then they looked at the beer that was next to it. They saw the beer that was next to it, or just above that sign. Oh, I might as well get that one. So they end­ed up buy­ing it. And actu­al­ly this kind of goes to show such a key point when we work in the super­mar­ket. Visu­al salien­cy is key. Visu­al salien­cy is prob­a­bly one of the biggest fac­tors that pre­dicts whether some­one’s going to buy some­thing or not. So we need to design prod­ucts that stand out and shelf. And it’s the same for spe­cial offers. So although super­mar­kets are great, all of these dif­fer­ent offers, actu­al­ly mak­ing sure the prod­uct stands out onshelf, and that spe­cial offer stands out on shelf, is prob­a­bly the most key factor.


Now, as I said, we’ve got all these dif­fer­ent offers. But we can still find ways to opti­mise it. So basi­cal­ly find ways that they can actu­al­ly get you to buy a prod­uct with­out giv­ing away more mar­gin than they need to. Because if you think about this, and we take it from a very, shall we say cog­ni­tive approach, we’re com­ing back to just notice­able dif­fer­ence. What the super­mar­ket is try­ing to do is try­ing to actu­al­ly work out what’s the biggest dif­fer­ence or biggest change they can make to the price that you’ll notice. They just want to go over just over that thresh­old and then stop.



Now we do all sorts of con­joint analy­sis for brands try­ing to work out how much we can do that, how big a change do we need to make. But actu­al­ly what we found out a lot of the time, we can actu­al­ly make dif­fer­ent changes to the way that we’re pre­sent­ing these things, we can change the way these offers are pre­sent­ed. And actu­al­ly, we can get some­times the same effect with­out actu­al­ly hav­ing to give any mar­gin away. So just the obvi­ous one, we start off with charm pric­ing, the idea of reduc­ing the dig­it by one, every­one knows that cog­ni­tive­ly, it’s the same thing. But mak­ing that a slight change makes a dif­fer­ence. If you look at some of the stats here, if a prod­uct is placed, or priced at 99 pence, con­ver­sion rate online is about 3%. For some of the brands we’ve worked with, if it’s a pound it’s about 1.88. So you’re get­ting vir­tu­al­ly a 1% increase in sales. Now does­n’t sound very much, but it’s a huge effect when we start to look at these, the sam­ples we col­lect in the com­mer­cial world, we’re look­ing at sort of hun‑, or 10s of 1000s. So you start to see these big dif­fer­ences. So these lit­tle changes we’re going to talk about today have some pro­found effects.


So here’s some lit­tle tips that we use, when we actu­al­ly try and design some­thing, some­thing we saw in the aca­d­e­m­ic lit­er­a­ture, and we’ve test­ed it in stores or super­mar­kets, the brain gets con­fused between size, numer­i­cal size and visu­al size. So these two spe­cial offers exact­ly the same thing, by chang­ing how we present them makes one actu­al­ly sell a cou­ple of per­cent bet­ter than the oth­er. So the one on the top is actu­al­ly the one that sells bet­ter. By putting the price in a small­er font than the ref­er­ence point, it actu­al­ly the brain feels like it’s a small­er price. On the one on the bot­tom, actu­al­ly, what you’ve got going on there is you’re mak­ing the font, the price feel like it’s more expen­sive, you’re mak­ing it feel like it’s more expen­sive, because it’s big­ger than the ref­er­ence price, oh the ref­er­ence font, the new low price. So just chang­ing how you present that makes a dif­fer­ence about 2 or 3% in terms of how many peo­ple are going to buy a prod­uct, all with just chang­ing the graph­ic design, the sort of lit­tle things that we can do.


Some­times where you can use exact­ly the same tech­nique, and what actu­al­ly what we’re try­ing to do here, we’re try­ing to make, not nec­es­sar­i­ly the price, but we’re try­ing to make a num­ber field big­ger. So on the first one on the left, is it say­ing we’re mak­ing the prices, the big thing that grabs your atten­tion $19 a month, $49 a month, or 99, it’s big­ger than all the ref­er­ence fonts. So we’ve changed it in terms of we’ve just reframed in terms of the bronze pack­age, the sil­ver pack­age and the gold one. So that’s the big thing that catch­es your atten­tion, the prices slight­ly smaller. 


We’ve changed it here, this is try­ing to make the, there we go, if you try to make it the size of how much data you get with an online inter­net pack­age, we want the num­bers to feel as big as pos­si­ble. So this times we’ve made it big­ger than the ref­er­ence of the thing. And it does­n’t nec­es­sar­i­ly need to be a font, you can do exact­ly the same thing in terms of a visu­al or graph­i­cal ele­ment, although the effect is slight­ly small­er. So by mak­ing the num­ber big­ger than ref ref­er­ence point, it feels like it’s either big­ger or more expen­sive, depend­ing on what you’re try­ing to do.


Like­wise, when we’re design­ing price tick­ets, real­ly sim­ple tricks you can end up doing to make the spe­cial offer or the dis­count feel like it’s bet­ter. So in the top one, if a prod­uct was 100 pounds, and you now reduc­ing it down to 50, you can make that field big­ger by sim­ply increas­ing the space between those two num­bers on the tick­et and actu­al­ly chang­ing it graph­i­cal­ly. So if we put the new price in a dif­fer­ent colour font, or a dif­fer­ent colour font, and a dif­fer­ent actu­al font, it actu­al­ly changes that and peo­ple feel like the price is going to be an even greater one. Some­where in the region, again, about one, one and a half per­cent when we’ve done this in brands. But again, that one and a half per­cent change actu­al­ly starts to add up con­sid­er­ably, espe­cial­ly when we’re look­ing at this at a super­mar­ket level.


Some­times you can even tell peo­ple, and if you tell what that dif­fer­ence is, say some­thing like say 50%, it works real­ly well. The chal­lenge for this is if you change it and it’s saved, save 23%, or should we say a non round fig­ure. At that point, peo­ple start to feel like the dif­fer­ence is small­er. So if you’re actu­al­ly going to tell them on the tick­et, what the dif­fer­ence they’re sav­ing is in terms of a per­cent­age. We only want to do that when we actu­al­ly have a nice round fig­ure. Typ­i­cal­ly, there’s not done though, because it just makes up that mess­es up the look and feel of a spe­cial offer sign. And it makes it look more clut­tered. And actu­al­ly the big­ger thing we’re using nor­mal­ly to dri­ve sales is visu­al atten­tion. visu­al atten­tion is the key thing that we’re look­ing for.



Oth­er things that you can see here, we’ve got one offer on the top, but it’s sim­ply changed and we change the order of where those that infor­ma­tion is pre­sent­ed. So in this one, all I’ve done is change it instead of it being in hor­i­zon­tal, we’ll put the change in prices ver­ti­cal­ly. A real­ly, real­ly small thing, but it makes it so much such a big dif­fer­ence of three to 4% in this case. Why? Well, it’s actu­al­ly eas­i­er for us to do the maths, it makes it more obvi­ous what the dif­fer­ence we’ve made it.


Anoth­er thing we’ve done here is we’ve actu­al­ly made sure that the change is dif­fer­ent for all the dig­its. So both the right hand col­umn and the left hand col­umn, each one is small­er. If they’re not both small­er, some­times peo­ple get a lit­tle bit maybe not con­fused, but they’re slow­er to process. And eye track­ing research in store typ­i­cal­ly shows that peo­ple are going to be look­ing for about half a sec­ond max­i­mum at a prod­uct. So it’s all about try­ing to max­imise the time peo­ple are spend­ing look­ing at the prod­uct, and mak­ing it eas­i­er and quick­er for them to make it to make a decision.


The oth­er thing that their bot­tom tick­et does, it gives them a rea­son of what that why the prod­ucts on dis­count rather than just say­ing it’s on dis­count. So this one just use the clas­sic exam­ple of a Jan­u­ary sale. And the rea­son that’s impor­tant is it stops or at least reduces the like­li­hood that anchor­ing is going to occur. So anchor­ing is the idea that the dig­it you see is it anchors or sets your expec­ta­tion of how much you’re going to pay. And this is always the risk, when you put a prod­uct on spe­cial offer that peo­ple start to think that, oh, it’s not real­ly val­ued at 79 pounds, it’s real­ly only cost 63. So by so by giv­ing them a rea­son, it just reduces the like­li­hood that anchor­ing occurs, and that they still val­ue the prod­uct at what its full price should be. So it’s just one of those lit­tle tricks we can play around with to try and try and stop that hap­pen­ing. So it’s a real­ly sim­ple thing that we start to do.


Anoth­er thing that you’ll often see some­times done, is actu­al­ly start­ing to start­ing to change the pre­sen­ta­tion of the price itself. This is some­thing that typ­i­cal­ly is more like­ly to occur, I sup­pose on a restau­rant menu. And again, I do a lot of work design­ing restau­rant menus. But it’s a sim­ple case of we drop the cur­ren­cy sym­bol. Now on restau­rant menus, when you drop the cur­ren­cy sym­bol, you can start to see about a some­where between 12 to 15% increase in the amount of mon­ey peo­ple spend on on an item. And like­wise, we’ve also dropped the dec­i­mal point. So rather than just say 89 pounds, point 00. This time, we’re actu­al­ly say­ing 89. And the rea­son this works is it helps to make the price less salient. Basi­cal­ly, we don’t want peo­ple to be con­cen­trat­ing on the price, we want to be con­cen­trat­ing, espe­cial­ly in a restau­rant set­ting where it’s a real­ly hedo­nis­tic expe­ri­ence, we want them to be con­cen­trat­ing on what they’re going to be enjoy­ing. We want them to con­cen­trate on the food they’re about to be eat­ing. So it’s all about try­ing to focus on that, we real­ly want to down­play the idea of price.


In an ide­al world, we’d love for peo­ple to pay for the food first and then eat it, it’s been shown that’s a much bet­ter way of doing it. The prob­lem is, we’ve kind of been trained, you pay for food first in cheap restau­rants, in posh restau­rants, you do it after­wards. But prob­a­bly the worst exam­ple of actu­al­ly where we get pric­ing salien­cy wrong is if you think of a Lon­don black cab, you get in a Lon­don black cab. And as you dri­ve, there’s not a lot to do. So you sit and watch that metre slow­ly tick­ing up, you watch it tick­ing up, it makes it real­ly real­ly salient. You final­ly get to your des­ti­na­tion, you get out. But the last thing you have to do is you pay. So you’ve made the pain of pay­ment, even greater. Con­trast that with Uber, but you take your phone, you pay vir­tu­al­ly so it feels less. So we know pay­ing by card is less painful than pay­ing by cash. So you can pay on your phone, you pay upfront. So you’ve actu­al­ly paid before the Uber arrives, and there’s no metre that’s tick­ing over so you can slow­ly watch it.


Now, the tech­niques we’ve got on the slide here, like remov­ing the pound sym­bol, or the cur­ren­cy sym­bol, yeah, I would­n’t rec­om­mend that for my clients in the super­mar­ket for the sim­ple rea­son. Well, typ­i­cal­ly, I just find it reduces clar­i­ty. We want every­thing to be as easy as pos­si­ble to process in the store, we’re real­ly try­ing to increase the con­cept of pro­cess­ing or per­cep­tu­al flu­en­cy. We want to make the pur­chase deci­sion as easy as pos­si­ble for the con­sumer, and pos­si­bly not hav­ing a cur­ren­cy sym­bol, it makes it a lit­tle ambigu­ous in store. But on restau­rant menus, and espe­cial­ly in the more classy ones. This is some­thing that typ­i­cal­ly gets done quite a bit.



But we test all of these, these are not just things so that we’ve said and read in the lit­er­a­ture and we think works. But all of these have come from peer reviewed exper­i­ments. And we’ve test­ed them. And actu­al­ly, the rea­son we test them is I’ve test­ed lots of ideas that have been shown to work in lab based stud­ies, but I’ve nev­er been able to recre­ate them. Some­times things don’t quite work out. Now it’s not a case of pow­er, because gen­er­al­ly when we test some­thing in a super­mar­ket, we’re not talk­ing about a sam­ple size of a cou­ple of 100 we’re talk­ing 1000s to 10s of 1000s. So our datasets are far more pow­er­ful than any­thing I ever had in my PhD. But this is a clas­sic exam­ple it was taught at Nudge­stock. And they were talk­ing about which web­site design is going to be more effec­tive. Option A or Option B. Chose the snow­boards because try­ing to drop a hint to some­one. But does it make more sense to put the pic­ture of a snow­board on the left of the text or the right of the text?


Now there’s aca­d­e­m­ic papers out there that say we eval­u­ate things dif­fer­ent­ly depend­ing where you’re pre­sent­ing them. So if the stim­u­lus we put it on the left hand side, then we’re going to process it slight­ly dif­fer­ent­ly. How­ev­er, some would argue, or some neu­ro­sci­en­tists would say, well, the right hemi­sphere is slight­ly bet­ter for pro­cess­ing the pic­to­r­i­al infor­ma­tion, it makes more sense of pre­sent­ed on the one side. Now I’ve test­ed this in 10 dif­fer­ent con­cepts on dif­fer­ent online stores. And yeah, I’ve nev­er been able to show it. We’ve test­ed it with prod­uct pack­ag­ing design. And again, we’ve nev­er been able to show it. Now prob­a­bly it does­n’t work in prod­uct pack­ag­ing, because sure, you might have the image on the right, and the text on the left, got the wrong way around. But actu­al­ly, there are lots of oth­er stim­uli that are com­pet­ing with and grab­bing your atten­tion. So that pos­si­bly explains why the super­mar­ket side, but pre­sent­ing it when I can con­trol what image appears on the screen for an online shop. As I said across 10 dif­fer­ent online stores, the effect has nev­er worked for me, maybe I’ve got my exper­i­men­tal design wrong. But just because it’s in aca­d­e­m­ic papers, it’s been talked about in Nudge­stock, this year, a num­ber of years, we’ve still nev­er been able to get it right.


So all of these ideas, all of these con­cepts have talked about, they’ve come from the recent research and papers, prob­a­bly lab based stud­ies. But we always made sure they work first, nor­mal­ly start­ing off by pilot­ing it in small steps, small scale exper­i­ment, some­times it just one super­mar­ket, we test it in that one indi­vid­ual store, before we roll it out, after we know it works. So hope­ful­ly that’s giv­en you a few ideas about how super­mar­kets are acci­den­tal­ly giv­ing away mil­lions, and how they can actu­al­ly some­times get the same effect by mak­ing things visu­al­ly salient, mak­ing sure it catch­es your atten­tion and actu­al­ly putting it on on the shelf. And by care­ful design. So there are oth­er ways of doing it. If you’ve got any ques­tions, let me know my email address is on the screen there if I prob­a­bly actu­al­ly hide myself so you can actu­al­ly see it. Yeah, I will make sure my dress this I’ll dis­ap­pear in a sec­ond. But yeah, thank you, Jo. And any ques­tions, I will hap­pi­ly answer them.


Jo Ever­shed 17:35  

Gareth, that was fas­ci­nat­ing. Thank you so much. I feel a bit like with the last ses­sion, which was all about gam­i­fi­ca­tion. Nev­er see a game the same way now that Nick has shown how all these effects are cre­at­ed, I think it’s going to be the same now with walk­ing around super­mar­kets. I’m going to see all of these price labels, and be just notic­ing all the tricks that they’re putting on my eyes and also notic­ing all the things they’re not putting on my eyes going ha you could have done bet­ter there.


Gareth Har­vey 18:02  

Yes, some­times they get it right. Some­times it’s a sim­ple case of ah, we need to make a leaflet or some­thing very quick­ly and it goes out the window.


Jo Ever­shed 18:10  

Yes, I think often­times when we’re work­ing, we all know that feel­ing of hav­ing to just get some­thing out of the door. Now, does any­body in the audi­ence have any ques­tions for Gareth, please put them in the q&a. That was fas­ci­nat­ing. I’m try­ing to think of a good ques­tion for you. Gareth. Is there a ques­tion you par­tic­u­lar­ly like me to ask you,


Gareth Har­vey 18:34  

if you’re going rather than look­ing at super­mar­kets, the one area I think it’s the most fun is going to look at restau­rant menus is one of my favourite com­mer­cial jobs is design­ing restau­rant menus. So it’s a case of it’s not just try­ing to work out to get peo­ple to spend the most. But it’s often think­ing about well, which dish­es have the high­est prof­it mar­gin. So it’s try­ing to work how we encour­age them. So you’ll always thought hope­ful­ly, you’ll see anchor­ing is a com­mon thing. But you’ll real­ly see we dropped the cur­ren­cy sym­bol, we dropped the dec­i­mal point. There’s things where you play around with the fonts. Yeah, there may be the more com­pli­cat­ed of fonts more cal­lig­ra­phy based, the hard­er it is to process and from a per­cep­tu­al flu­en­cy, the more expen­sive it feels. Now, that may be a good thing if it’s a lux­u­ry restau­rant, if it’s a cheap­er one. Actu­al­ly, it’s not real­ly so great. 

Jo Ever­shed 19:19  

That is so tricksy, I am actu­al­ly a mas­sive fan of the new thing they’re doing in the UK where they show you the calo­ries as well as the price and it has com­plete­ly changed my behav­iour. I know some peo­ple aren’t a fan. They know it’s a com­pli­cat­ed dis­cus­sion point. But I feel like I am spend­ing both mon­ey and calo­ries on a day by day basis. And it’s amaz­ing because some of the dish­es that I thought were healthy turns out they’re not.

Gareth Har­vey 19:40  

Yeah, the oth­er one is some­thing they’ve strug­gled with with the menu design because in the past, they often put healthy sec­tions have a healthy sec­tion or a low calo­rie sec­tion. And they found that when they did that it gen­er­al­ly decreased the like­li­hood that some­one would order it part­ly because if it’s a low calorif­ic, or low fat, we assume it’s less tasty. So they actu­al­ly they had that was one neg­a­tive. So when they actu­al­ly put it in as part of the main menu and put the calo­rie infor­ma­tion in had a slight­ly bet­ter influence.

Jo Ever­shed 20:06  

Ah, very inter­est­ing. There was one ques­tion from the audi­ence, which actu­al­ly is a ques­tion I had as well. What’s this amaz­ing tech you’re using where you get to sit in front of your slides, it’s so much more engaging.

Gareth Har­vey 20:15  

I’m using OBS. OBS. So it’s com­plete­ly free. And it was a sim­ple case of it’s nor­mal­ly much bet­ter than this, but I’m work­ing at a friend’s house today and I did­n’t have time to set it up prop­er­ly, but um, OBS Open Broad­cast System.

Jo Ever­shed 20:31  

I will put the link into the chat for every­body else who might be interesting.

Get on the Registration List

BeOnline is the conference to learn all about online behavioral research. It's the ideal place to discover the challenges and benefits of online research and to learn from pioneers. If that sounds interesting to you, then click the button below to register for the 2023 conference on Thursday July 6th. You will be the first to know when we release new content and timings for BeOnline 2023.

With thanks to our sponsors!