The Approval Cliff · Problem

The Compromise Logo

How “meeting in the middle” produces work nobody chose. A merge settles the argument between the people in the room. The audience never stood in the middle.

Quick answer

Should you combine two logo concepts when your team can’t choose between them? Usually not. A merge settles the argument between the people in the meeting. It does little for the people the logo is for, because audiences respond to a strong direction as a whole, and a merge breaks up exactly the qualities that made each direction strong.

The scene is familiar. Two logo directions are on the wall, each one has a camp behind it, and somebody suggests taking the type from the first and the color from the second. Everyone relaxes and the meeting ends on time.

The merged logo is the one option nobody in that room picked, and the audience won’t pick it either. Averaging two strong preferences puts the work in the empty space between them, because audience response clusters. A perception map almost never has a midpoint. It has clusters, and the space between two clusters is where nobody stood.

Why does merging two logo directions feel like the right call?

The compromise solves a problem, just a different one from the problem the logo has. Two senior people disagree, the project has stalled, and a merge lets both of them keep something. As a way to settle an argument between colleagues, it’s fair and it’s fast.

The trouble is who it’s fair to. The merged direction was shaped for the two people in the meeting. The logo has to work for the people who’ll see it on a truck, a badge, a pitch deck and a homepage, and none of them were in the room.

There’s also a hidden assumption in the merge: that each camp’s favorite is good because of its parts. If that were true, a mix of good parts would be at least decent. Visual work rarely behaves that way. People read an image as a whole before they notice any element of it, and the elements that make a direction work tend to depend on each other.

So the compromise answers “how do we stop arguing?” The brief asked a different question: what will this audience read as right? Those two questions have different answers more often than anyone in the room wants to admit, and the merge only answers the first.

If you’ve been the person who proposed the merge to get a stuck project moving, that was a decent instinct. You were protecting the relationship and the timeline. The fix is to give the disagreement somewhere better to go.

What does the audience do when two camps disagree?

A perception map is a board of images that an audience sorts one image at a time, placing a green dot for a match or a red dot for a miss against a single question, so you can see where agreement forms. That is audience perception mapping: an audience responds to images instead of describing them, and the pattern shows where the group agrees, where it splits, and why. When a group of people responds individually to a range of images like this, the responses don’t spread evenly across the board. They pile up. Separate people bring a shared read to the same image and land in the same places, and a perception map shows those piles as clusters. How to read a perception map walks through the three groupings: the positive cluster, the negative cluster and the middle.

If a room has two camps, an audience often has two as well, and the map shows two clusters. The space between them is mostly low engagement: images people looked at and didn’t feel strongly about. That middle doesn’t hold a hidden consensus waiting for someone to find it.

The more dangerous version shows up when the two groups’ responses get pooled. In When the Segments Disagree, two bars on the chart are “the whole reason to run the test segmented,” because “averaged together they would have read as two mediocre elements rather than one strong preference each way.” The same page puts it more bluntly: without segmentation, the responses “average together into a middle that represents nobody, and nobody in the room knows the average is hiding a fork.”

Figure 1 · Two camps, two clusters

Two camps on a segmented perception map pile their green dots on two different directions, the merge sits in the empty middle, and the same responses pooled read as two split bars Two panels. The top panel, labelled segmented, two camps, two clusters, is a board with direction A on the left, camp one's favorite, under a pile of green dots, and direction B on the right, camp two's favorite, under another pile of green dots, each with a few red dots from the other camp. In the middle a dashed tile reads A plus B, the merge, where nobody stood, with four grey dots around it. The bottom panel, labelled pooled, the same responses averaged together, shows direction A and direction B each as a bar split half green, half red, with an arrow to the words reads as two mediocre elements, and a middle that describes no actual person. A closing line reads a perception map almost never has a midpoint; it has clusters, and the space between them is where nobody stood. SEGMENTED two camps, two clusters A direction A camp one’s favorite B direction B camp two’s favorite A+B the merge where nobody stood POOLED the same responses averaged together direction A 50% green 50% red direction B 50% green 50% red reads as two mediocre elements and a middle that describes no actual person A perception map almost never has a midpoint. It has clusters, and the space between them is where nobody stood.
Segmented, the map shows what each camp’s audience actually did: two strong piles and a thin middle. Pooled, the same responses flatten into two split bars, and the merge is built toward the one number on the page that describes nobody.

That is the compromise logo, drawn as data. Two strong preferences, averaged, look like mild interest in the middle. A design team that builds toward that middle is building toward a number that describes no actual person.

Why does a merged logo underperform both originals?

A strong visual direction usually works because a few qualities move together. Split them up and each half stops doing its job.

The clearest published evidence is the Jazz It Up study. Twenty-five people sorted 33 designs against one phrase and placed 396 dots. An image counted as a match only if at least 65% of the dots it received were green. Nine images cleared that bar, and every one of them did two things at once: loud, clashing color and a surface that looked like you could touch it, such as goo, chrome, fur or glossy plastic. The study’s own summary is that “the dials only count when they move together.”

The losing corners are the instructive part. Loud pink cutouts on a flat, pale photo took 67% red, color without body. Beautifully embossed letters in a single soft color took 85% red, 17 of 20 dots, body without color. Each loser kept one of the two qualities that made the winners work.

Now picture the merge. One camp loves the energy of the bold direction and the other loves the refinement of the restrained one. The compromise takes the bold color from the first and the restrained surface from the second, and lands straight in a losing corner. The merge keeps one ingredient from each camp and drops the pairing that made either camp’s favorite land.

Figure 2 · The dials only count when they move together

The Jazz It Up grid of loud color by a touchable surface, where only the corner with both cleared the bar, and the merge taking one ingredient from each camp into a losing corner A two-by-two grid from the Jazz It Up study, 25 people, 33 designs, 396 dots. Columns are the surface, flat and polished or you could touch it; rows are the color, loud and clashing or soft and single. Loud color on a flat surface is red, 67% red, color without body, loud cutouts on a flat pale photo. Loud color on a surface you could touch is green, nine matches at 65% or more green, both dials moving together, goo, chrome, fur, glossy plastic. Soft color on a flat surface is grey, nothing to sort on, neither dial. Soft color on a touchable surface is red, 85% red, 17 of 20, body without color, embossed letters in one soft tone. On the right, the merge: camp one loves the bold direction and takes its loud color, leaving its surface behind; camp two loves the restrained direction and takes its flat, polished surface, leaving the rest. The two combine into loud color on a flat surface, and a dashed red line runs straight into the losing corner. A closing line reads the merge keeps one ingredient from each camp and drops the pairing that made either camp's favorite land. THE JAZZ IT UP GRID 25 people, 33 designs, 396 dots SURFACE flat, polished you could touch it soft, single COLOR loud, clashing COLOR nothing to sort on neither dial 85% red, 17 of 20 body without color embossed letters in one soft tone 67% red color without body loud cutouts on a flat, pale photo 9 matches, 65%+ green both dials, moving together goo, chrome, fur, glossy plastic THE MERGE camp one loves the bold direction takes its loud color, leaves its surface behind camp two loves the restrained direction takes its flat, polished surface, leaves the rest loud color on a flat surface straight into the losing corner The merge keeps one ingredient from each camp and drops the pairing that made either camp’s favorite land.
Nine of 33 designs cleared the bar, and all nine sat in one corner of the grid. The compromise takes the color from one camp and the surface from the other, and the study already measured where that lands: 67% red.

The Curious study shows the same thing from the middle of the map. Thirty-three people sorted 24 images against the word “curious,” and three images finished at 52% to 53%, a dot or so under the study’s 54% bar, so neither a match nor a miss. The report calls each of them “a half-match for the rule”: a profile with the right gaze inside a polished ad, a figure with a body and no face to read. Its verdict: “Half a match earned half the room.”

A compromise is a half-match on purpose. And half the room, on a perception map, is a split. Half the people said yes and half said no, and the average of those two answers looks like a shrug.

Isn’t a merged logo a form of consensus?

In the review, the merge looks like agreement. Both camps nod and nobody blocks it. Design by committee is the longer version of how a review room produces that kind of yes.

Constellations’ pricing page draws the line this way: “Approval means nobody blocked the work. Alignment means everyone believes the direction is right and can defend it.” A compromise logo is approval without alignment, and that gap shows up later.

Each camp looks at the merged mark and sees what was taken away from them. So the first time a salesperson calls it generic or a board member asks why it looks like a competitor, there’s nobody in the building who chose it and can argue for it. The direction gets reopened, usually after production money has gone into it. That’s the moment the approval cliff describes: the point where a change costs the most and the evidence for the original call is thinnest.

How do you choose between two logo directions instead?

Change what goes in front of people, and who it goes in front of.

Don’t bring the two finished directions to the audience. Take the research and instinct behind each camp and break it into elements: type, color, form, texture, imagery. Build a range of images from those elements and put the range in front of the audience the logo has to serve, segmented by the groups each camp believes it’s speaking for. The direction emerges from where the responses cluster, and results come back in under 48 hours once the test is live. How to test a logo before you launch it walks through the setup.

About 30 responses is enough to see a pattern. For a decision a brand will live with for years, plan on 300 to 500.

Then read the map in two layers.

Start with where every segment agrees. That part becomes the core of the identity, the part that has to work everywhere and never needs arguing about again. Very often it contains pieces both camps cared about. That overlap is the only honest version of meeting in the middle, because the audience put it there and no one had to give anything up to get it.

Then look at where the segments fork. A fork is a choice, and now it’s one you get to make on purpose. Lean the core toward the segment that carries the buying decision, or hold the core steady and let individual campaigns lean toward each audience. Either is defensible, because both are decisions. A merge is a way of avoiding one.

Figure 3 · Read the map in two layers

The elements behind each camp's favorite become a range of images, the segmented audience sorts them, and the map is read in two layers: the overlap is the core, the forks are a choice From left to right. Five chips, type, color, form, texture, imagery, the elements behind each camp's favorite, lead to a grid of twelve tiles, a range of images built from those elements, which leads to the map, read in two layers. The map shows two overlapping circles, segment one in blue on the left and segment two in cyan on the right, each with its own dots, and their overlap filled green with its own cluster, labelled the core: where every segment agrees. Beside it, layer one, the overlap: the core of the identity, settled by the audience, nothing given up; and layer two, the forks: a choice, and now one you get to make on purpose. Three pills underneath: about 30 responses show a pattern; 300 to 500 for a decision a brand lives with; results in under 48 hours. A closing line reads the overlap is the only honest version of meeting in the middle; the audience put it there. WHAT GOES IN FRONT OF THE AUDIENCE not the two finished logos type color form texture imagery the elements behind each camp’s favorite a range of images built from those elements THE MAP, READ IN TWO LAYERS segment one segment two the core: where every segment agrees layer one, the overlap the core of the identity, settled by the audience, nothing given up layer two, the forks a choice, and now one you get to make on purpose about 30 responses show a pattern 300 to 500 for a decision a brand lives with results in under 48 hours The overlap is the only honest version of meeting in the middle. The audience put it there.
The merge tried to find the middle by hand. The map finds it by asking: the overlap is what both camps’ audiences agreed on without anyone giving anything up, and everything outside it is a decision you get to make with the evidence in front of you.

How do you stop a clear result drifting back to the middle?

Even with a clear cluster on the map, teams drift. Someone asks to soften the color a little. Someone else wants the blue from the other option back, just in the secondary palette. Each request is small, and together they walk the work back toward the empty middle one revision at a time.

A few habits hold the line.

Name the quality the audience sorted on, in one sentence, and keep it in front of every review. In Jazz It Up that sentence is “loud color on a surface you could touch.” Every requested change can be checked against it: does this edit keep the quality, or trade it away to please someone?

Put the burden of proof on the change. Anyone who wants to move away from what the audience converged on needs evidence, and a preference from the person with the biggest title doesn’t count as evidence. That rule protects the work, and it also protects the people who’d otherwise have to fight that battle alone.

Keep the map in the room. The question at each review stops being “which of us is right?” and becomes “what do we do with what the audience saw?” That question has an answer, and the merge never did.

If two camps are about to split the difference on your next identity project, book a scoping call and we’ll frame the question the audience should answer before anyone starts merging.

Frequently asked questions

Should you combine two logo concepts when the team can’t choose?

Usually not. A merge settles the argument between the people in the meeting and does little for the people the logo is for. Audiences read a direction as a whole, and the qualities that make a strong direction work tend to depend on each other, so a mix of parts from two directions keeps the parts and loses the pairing.

Why does a merged logo underperform both originals?

Because a strong direction works when a few qualities move together. In the Jazz It Up study, every design that cleared the bar paired loud color with a surface you could touch; designs with only one of the two took 67% and 85% red. A merge takes one ingredient from each camp and lands in exactly that losing corner.

What does a perception map show when two camps disagree?

Two clusters, with a thin middle between them. Separate people bring a shared read to the same images and pile up in the same places, so an audience with two camps shows two piles. The space between is low engagement, not hidden consensus, and pooling the two groups turns two strong preferences into one middle that describes nobody.

How do you choose between two logo directions?

Break the research and instinct behind each camp into elements, build a range of images from them, and put the range in front of the audience the logo serves, segmented by the groups each camp speaks for. About 30 responses show a pattern; plan on 300 to 500 for a brand decision. Where every segment agrees becomes the core; where they fork is a choice you make on purpose.