You, my friend, are hard­wired to find mean­ing; you can not help but con­nect the dots.

Case in point, the tri­an­gle to the right doesn’t exist. The only shapes in that pic­ture are three black pac-man shapes.

Yeah, the neg­a­tive space left by those pac­man shapes include wedges of white — but the larger tri­an­gle that you see con­nect­ing those wedges of white into a mean­ing­ful pat­tern only exists in your mind.

And yet, if the pac­mans are there, you can’t help but see that tri­an­gle, can you?

In fact, the only way to not see the tri­an­gle is to remove two of the pac­man fig­ures, ’cause as long as the dots are there, you WILL con­nect them.

Design­ers refer to this as “clo­sure,” and it’s more than just a par­lor trick or visual illusion.

Clo­sure and Image-Text Interaction

Clo­sure, as it turns out, not only comes into play between ele­ments within a pic­ture, but also between image and text. And this inter­play was espe­cially on dis­play in a recent post by the always-interesting Derek Halpern (h/t Melissa Breau)

Halpern ref­er­ences recent psy­cho­log­i­cal stud­ies show­ing that state­ments accom­pa­nied by related images are con­sid­ered more believ­able than the same state­ment with­out an image. So, a state­ment like “The liq­uid inside a ther­mome­ter is mag­ne­sium” was more fre­quently rated as true when it was accom­pa­nied by a pic­ture of a thermometer!

Sim­i­larly, state­ments about whether some obscure “celebrity” was alive or dead were also more fre­quently rated as true when the state­ment was accom­pa­nied by a pic­ture of the celebrity. And this effect was the same regard­less of whether the pic­tured celebrity was pro­nounced dead or still living.

Clearly, pic­tures have per­sua­sive power beyond what any­one has ever suspected.

And just know­ing this is incred­i­bly use­ful, but in my opin­ion, the real meat of these stud­ies comes from ask­ing WHY. For­tu­nately, one of the posts that Derek links to nicely sum­ma­rizes the hypoth­e­sis formed by the sci­en­tists who con­ducted these tests [empha­sis mine]:

The rea­son for the dif­fer­ence lies in the sus­pected mech­a­nism at work. The “truthi­ness” researchers (New­man et al., 2012) spec­u­late that a not nec­es­sar­ily pro­ba­tive but rel­e­vant image, like the tire slide above, increases the “cog­ni­tive avail­abil­ity” of the con­cept. That means the mind finds it eas­ier to think about and elab­o­rate on the con­cept. In the process, that makes the claim seem more famil­iar which in turn makes it feel more true: “Truthi­ness” achieved.

There are also other mech­a­nisms that facil­i­tate elab­o­ra­tion. For exam­ple, the researchers refer to the notion of a “seman­ti­cally pre­dic­tive sen­tence,” which means phras­ing that leads a lis­tener to antic­i­pate what the upcom­ing words will be. For exam­ple, “the stormy seas tossed the boat” is more seman­ti­cally pre­dic­tive than “he saved up his money and bought a boat.” That expec­ta­tion causes a lis­tener to feel more famil­iar­ity and trans­late that into greater verac­ity (Whit­tle­sea, 1993). When peo­ple are engaged — by antic­i­pat­ing the final word in this case — they engage in more flu­ent pro­cess­ing and that leads to a feel­ing of truth.

That process extends past the role of imagery. In New­man and asso­ciates’ sec­ond exper­i­ment, they showed that includ­ing non-probative words instead of a photo pro­duced the same effect (e.g. accom­pa­ny­ing a polit­i­cal leader’s name with infor­ma­tion about eth­nic­ity, sex, hair color, etc. — fac­tors that cre­ate a pic­ture in the mind, but with­out telling the reader whether the fig­ure is alive or dead). The addi­tional infor­ma­tion led peo­ple to believe that the asso­ci­ated claim was more likely to be true.”

Ok, so first thing, what the heck does non-probative mean? Basi­cally, it means the photo does not log­i­cally prove the state­ment to be true or false. Non-probative images are merely decorative.

For instance, if you have a state­ment like “The US has the high­est incar­cer­a­tion rate of any coun­try” and you then accom­pany that state­ment with a bar graph like the one on the right, then that image would be con­sid­ered “pro­ba­tive” because it would log­i­cally “prove” the state­ment to be true, assum­ing that you took the image at “face value.”

This is opposed to a more dec­o­ra­tive image of a con­vict behind bars. That photo would be related to the state­ment about incar­cer­a­tion rates, but it would not log­i­cally “prove” anything.

Why “Non-Probative” Doesn’t Mean Non-Persuasive

Although a non-probative pho­to­graph may not “prove” any­thing, it can still sug­gest and imply.

So who says sug­ges­tion is any less per­sua­sive than out­right statement?

For instance, if that photo of the con­vict behind bars was black, it might remind the test sub­ject that the US jails a dis­pro­por­tion­ate num­ber of African Amer­i­cans — a visual sug­ges­tion that would surely color one’s judge­ment of the accom­pa­ny­ing state­ment, right?

Because peo­ple can’t help but con­nect the dots between image and state­ment.

It works the same way with the celebrity state­ments as well. because we believe in inter­nal con­sis­tency. If some­one hands us a state­ment with spelling and gram­mat­i­cal errors, we become less likely to lend cred­i­bil­ity to the state­ment or the per­son who wrote it. Any­one recall Dan Quayle’s Potato gaff?

So when someome men­tions a lit­tle known celebrity and pro­vides a pic­ture of said celebrity, we not only auto­mat­i­cally con­nect the dots between pic­ture and celebrity, but we con­nect the dots between know­ing who the heck one is talk­ing about with know­ing what the heck one is talk­ing about. The thought process goes some­thing like, you obvi­ously know who this guy is and I don’t, so you prob­a­bly also know whether or not he’s still alive… 

Why do I think this is a greater fac­tor than the psy­chol­o­gists’ “increased cog­ni­tive avail­abil­ity” hypothesis?

Because sci­en­tists who con­ducted the same test, but who accom­pa­nied the celebrity state­ments with facts and stats about the celebrity instead of a pic­ture recorded the same effect: the stats boosted the per­ceived cred­i­bil­ity exactly as the pho­tos did in the pre­vi­ous test. And my guess is that the stats “prove” to the test sub­jects that the peo­ple mak­ing the state­ment really know who they’re talk­ing about, in pretty much the same way that a pic­ture would. Makes sense right?

But would stats really help peo­ple hold an idea in their heads? Would stats make the celebrity more “cog­ni­tively avail­able” to the test sub­jects? I rather doubt it.

So it’s really less about help­ing peo­ple hold the idea in their heads, and more about sub­tly con­vinc­ing them you know what you’re talk­ing about.

And images don’t have to do that explic­itly, as impli­ca­tion and clo­sure work just fine, if not even better.

A pic­ture of an old-fashioned ther­mome­ter dis­plays a sil­very strip in the mid­dle of it, imply­ing the idea of liq­uid metal. Con­nect­ing the dots between image and state­ment, and sud­denly the idea of liq­uid mag­ne­sium seems a whole lot more plausible…

It makes me won­der if a pic­ture of a modern-day ther­mome­ter would have had the same results…

Using Clo­sure To Improve Per­sua­sion & Impact

So… we know this clo­sure between image and text cre­ates greater believ­abil­ity. But how would one use it for images alone?

Well, for images, the short answer is to give the viewer 2 + 2 rather than just hand­ing them 4. Cre­ate an image that makes them con­nect the dots between ele­ments of the image.  Here are some great exam­ples of that:

OK, so these are cheat­ing a bit because they’re both text-based images, but nei­ther of them make much sense until you con­nect the dots — allow­ing both ads to make their state­ments all the more strongly.

Here’s another exam­ple, this time with an hon­est, no-kidding image:

Again, the image is mean­ing­less until you men­tally “fill the gap” about what those sets of feet really indi­cate. Clo­sure at work. There’s also a nice gap/connection between the stockinged feet and the text.

And on a more purely visual note, much of the emo­tional impact of this image can be attrib­uted to the “gaps” that it forces your mind to fill in:

Great exam­ple of clo­sure used to increase men­tal engage­ment and impact. But what about using clo­sure to select more pow­er­ful imagery to accom­pany your per­sua­sive copy and messaging?

How to Use This In Web Copy

Here’s what I suggest:

1. Use the “I saw it with my own eyes, so it must be real” approach

If you’ve got a tes­ti­mo­nial, you could, as Derek sug­gests, place a pic­ture of the cus­tomer who gave it to you next to the tes­ti­mo­nial. That’ll work. Or, if you don’t have that, you could take a photo of the hand-written tes­ti­mo­nial and place it next to the testimonial.

It sounds silly, but just imag­ine the dif­fer­ence between some­one say­ing “this per­son wrote in to say X” and some­one hand­ing you the actual hand-written note and say­ing “look what cus­tomer X had to say.” Which would be more per­sua­sive? The lat­ter, right? Because then you could say that you saw the tes­ti­mo­nial “with your own eyes.”

Of course, the “so it must be true” part would likely go unsaid, but it would be all the more pow­er­ful for it. And that’s why an image of the hand-written tes­ti­mo­nial would be more per­sua­sive than the state­ment alone.

So within your sales copy, deter­mine which ele­ments peo­ple would most want to see with their own eyes, then find images that would give them a sim­i­lar sense of verification.

Another exam­ple, I once worked with a metal roof­ing com­pany that claimed a no-kidding 50-year life span on their roofs. Now the claim and guar­an­tee is great. But what I advised them to do was  find the old­est roof they had ever installed (which turned out to be 30+ years old) and to get both an estab­lish­ing pic of the building/roof and a close-up pic­ture of the metal “tiles.”  It’s one thing to claim a 50-year life span, and another entirely to show a 30-year roof that looks brand new.

Just don’t do the cheese-ball thing of using blacked out bank state­ments to “prove” how much money you make!

2. Use images to sug­gest and emo­tion­ally prime belief

No one does this bet­ter than apple. Take a look at this screen shot from Apple’s page on the new iPad 4:

It’s not an acci­dent that the iPad sports an image of two Porsche’s about to race, or that the image is from a graph­ics inten­sive game. The mes­sag­ing is about speed after all. Speed achieved through high-performance engi­neer­ing. Don’t you think the image of “Porsche Race Cars” brings all that to mind rather powerfully?

Here’s another example:

So… what the heck is that black ring in the mid­dle of the picture?

It’s not a mag­ni­fy­ing glass. Nor is it a cam­era lens, is it? Maybe it’s some kind of weird bas­tard love child between the two…

But it doesn’t mat­ter, does it. We instinc­tively know that this is show­ing us that even when you mag­nify the pic­ture 2.5X, it’s still high-res enough to look crisp and un-pixelated. Of course, the copy never makes that claim. But the pic­ture cer­tainly sug­gests it, doesn’t it?

If Images Com­bined with State­ments Are Pow­er­ful, What About Video?

But were this really starts to come into it’s own is in explana­tory videos. But that’s a sub­ject for another post…



  1. Melissa on 10.26.2012

    Great fol­low up post here Jeff! Glad you found the link I shared helpful…

    - Melissa
    Melissa´s last blog post ..Cre­at­ing Char­ac­ters from Scratch

Leave a Reply

CommentLuv badge

WP-SpamFree by Pole Position Marketing