Put one of two pictures on a €200m product line and, by my rough math, it brings in about €5m more.

Same ad. Same brand. Same headline. Only the picture differs.

And in most companies, the choice between those two pictures is made in a meeting.

The €5 Million Show of Hands

The meeting usually goes like this. Someone finds one picture warmer. Someone else says the other one did better in the pretest. The most senior person in the room settles it, and the media money goes out.

That is no criticism of anyone in that room. It's just a strange way to make a decision of this size. A bit like setting a price by asking who around the table would pay it.

We asked the room in a recent webinar how teams choose their visuals. About half said: discussion and gut feel. Most of the rest named liking tests, click tests or A/B tests. Nobody named a method that measures the effect on sales.

Asking customers directly doesn't fix it either. People can like an ad and still not buy it. Ask them and they give you the reasonable answer, which is rarely the one that moves them. That's the say-do gap, now wearing a creative-review hat.

So we stopped asking and measured the gut.

What We Actually Tested

SUPRA and SightEffect joined forces for a study we call "Do Pictures Sell?" We used a pension ad from AXA as the test case, a product where customers commit for decades, so the stakes of the picture are not trivial.

Per Kasch built the test pictures. He has spent more than 15 years on brand imagery for SWISS, Lufthansa, Mercedes-Benz and SBB. We showed versions of the ad to 1,000 people in Germany, about ten seconds each, and measured what each picture did to the wish to buy, using our DeepCause methodology, which measures gut-level reactions and then works out which of them drive the purchase.

Each version changed exactly one thing. Same people, same place. Per checked every version for consistency, otherwise we would have biased our own study. With a real photo shoot you could never afford that kind of control for research. With AI you can.

The headline result, as we report it: picture A, the one with a worked-out look and a worked-out feel, raised purchase propensity by 2.4% over picture B, the one with neither. After a single impression.

2.4% sounds small until you put it into money. On a €200m product line it comes to roughly €5m, by my rough math, from one look at one picture. (Treat the euro figure as an illustration of scale, not an audited forecast.)

Look and Feel Are Two Different Things

The most useful thing I learned from Per: "Look and feel are very distinct things." He splits every picture into two parts.

The look makes the brand recognisable: light, sharpness, perspective, colour, composition. The feel tells the story: the people, the moment, what they are doing.

For AXA, the version without the right look used the flashy wide-angle style many brands use right now. The version without the right feel showed a couple just drinking coffee. The full version showed them sorting out their finances at the kitchen table.

In our data, a better look added about 1.2% to sales propensity, and a better feel added a similar amount. Both together added the 2.4%. Two levers, each worth something, stacking.

Look vs. feel: the split to use in your next creative review

  • Look = recognisability. Light, sharpness, perspective, colour, composition. Does it look like you, or like everybody's stock library?
  • Feel = story. Who is in the picture, what moment is it, what are they doing? Does the scene say what your product does for the customer?
  • Judge them separately. A picture can have a great look and the wrong feel, or the reverse. Lumping them into "it's nice" hides which lever you pulled.

The Space Suit Problem

Yes, a child in a space suit makes a retirement pension ad interesting. But it does not make customers buy.

The space-suit picture is AXA's original campaign motif from Switzerland. Every creative pro loves such a motif. At the gut level it scored well on "interesting" (14.2 on our scale). But only 1.2 on honest and 3.6 on secure.

The kitchen-table version scored lower on interesting, just 6.3. And much higher on trustworthy (13.3), honest (12.2) and secure (10.4).

That trade-off usually reaches a CMO as "the bold one" versus "the safe one." For a product customers commit to for 30 years, the safe one is the better buy. Interesting earns a second of attention. Trust turns that second into a purchase. The campaign may well serve other goals. For the purchase decision, the kitchen table wins.

Both routes, look and feel, worked through the same perception: trust.

Attention Is Not the Goal

In my book THE TOP 5% I called attention a vanity metric. This data backs it up. In both categories we looked at, "attention-grabbing" did nothing for purchase on its own.

A picture needs to hold attention for the right reason, a reason that is aligned with what the product does for the customer. Otherwise it's free entertainment and wasted marketing budget.

And here's the part that surprised me most. After the messaging is done, the rightly crafted picture can add another 35% in impact. In our last study on still ads, that first gut reaction explained 35% of the propensity to buy. My caveat on all of these numbers: they come from one study with 1,000 people in one country, so read them as reported results from our own research, not as laws of nature.

How to Choose a Visual Without a Show of Hands

You don't need a €5m ad budget to apply this. You need a different question. Not "which picture do we like?" but "which picture changes the wish to buy, and through which perception?"

That is what deep implicit research is built for: measure the gut reaction in milliseconds, before people get to rationalize, then use Causal AI to work out which of those reactions actually drive the purchase. If you want the broader logic of why asking people why they buy fails, start with why customers can't tell you why they buy.

If your creative process is more gut-feel than evidence, you are in the same situation as the whole webinar room. The AI Diagnostic is the quickest way to see where your own decisions rest on stated opinions instead of measured drivers.

A €5m choice deserves better than a show of hands. A or B: which did your gut pick?

Do Pictures Sell: frequently asked questions

Do pictures in still ads really change whether people buy?

In SUPRA's 'Do Pictures Sell?' study with SightEffect, about 1,000 people in Germany saw versions of the same AXA pension ad for roughly ten seconds each. Only the picture changed. The version with a worked-out look and feel raised sales propensity by a reported 2.4% over the version with neither, after a single impression. On a €200m product line, Dr. Frank Buckler estimates that at roughly €5m.

What is the difference between look and feel in advertising imagery?

Per Kasch, who has spent more than 15 years on brand imagery for SWISS, Lufthansa, Mercedes-Benz and SBB, splits every picture into two parts. The look makes the brand recognisable: light, sharpness, perspective, colour and composition. The feel tells the story: the people, the moment and what they are doing. In SUPRA's study, a better look and a better feel each added roughly 1.2% to sales propensity, and 2.4% together.

Why does an attention-grabbing ad not necessarily sell?

In the AXA test, the space-suit motif scored high on gut-level 'interesting' (14.2) but low on honest (1.2) and secure (3.6). The kitchen-table version scored lower on interesting (6.3) but much higher on trustworthy, honest and secure. For a pension product customers commit to for decades, trust turned out to matter more than attention. Attention-grabbing alone did nothing for purchase in either category studied.

How can teams choose between two ad visuals without relying on gut feel?

Measure the effect of each version on purchase propensity instead of asking who likes it. Change one element per version, show each to a sizeable sample for a few seconds, and capture gut-level reactions. Then use Causal AI to work out which perceptions, such as trust or honesty, actually drive the purchase. This is the logic of SUPRA's deep implicit research, which avoids the say-do gap of direct questions.

Dr. Frank Buckler advises leadership teams on growth decisions that depend on why customers choose. He founded SUPRA, holds a PhD in marketing and was among the first to apply causal AI to marketing questions. His latest book is THE TOP 5%.