Analyzing your designs

Find correlations between performance and different design elements

Table of Contents

Design dimensionsObjects in contentObject in content - individual comparisonFace in contentLogo in contentPrimary colorPrimary saturated colorBrightnessSaturationTemperature

When it comes to designing creatives for paid social advertising, there is often a lot of confusion and misconceptions within the question: “How to create the best possible design?”. 

The results your ads can achieve depend on a lot of factors, and they’re not universal to all advertisers. That means the best way to truly understand how to design ads for maximum returns has to come from analyzing what works, and doesn’t work, in your own designs.

That is exactly what the Design insights in Confect Analytics aim to do. Through the help of artificial intelligence, you can see what design choices correlate with the highest performance to apply this in your future designs.

Design dimensions

Inside Confect, the elements you can analyze and make comparisons with are referred to as dimensions.

This encompasses a wide range of design choices, from which objects and elements correlate with the highest performance, to the color options that achieve the highest results for you.

Objects in content

In this insight, the various objects that you have previously used in your designs will be identified by artificial intelligence, and displayed in a two-dimensional word cloud.

The horizontal axis refers to the amount of impressions that ads with these objects have received. The objects on the left have less than the average number of impressions, while the ones on the right are more commonly used in your ads.

The vertical axis displays the metric you’ve chosen to optimize for (i.e. ROAS, Cost Per Purchase, etc.). The objects in the bottom half score lower than average, while the top half includes objects that correlate with above average levels of this metric.

Please note that when optimizing for “Return” and “Rate”, higher means better performance, but when using “Cost Per X”, the lower the better. Always take the metric into consideration when analyzing insights.

In this example, we want the highest return possible, so the best objects to use will be in the upper half. If we used cost per purchase, we want it as low as possible, so ideal words will be located in the bottom half.

Object in content - individual comparison

In addition to seeing all the objects in a word cloud, you can also compare a single object to the average, or another object, and see the correlated performance.

To do this, simply search for the object to see a comparison to the overall average, or write multiple words separated by the | sign to compare different objects. For example: “Sneakers| Denim”.

Face in content

Including human faces in paid social content can be a valuable tactic to improve performance in many cases.

In this insight, you can see whether your ads with faces in them achieve better results than ads without faces, as well as the percentage difference between them.

Logo in content

Your logo is an indispensable part of marketing, however it doesn’t necessarily mean it should be included everywhere by default.

Certain advertisers and industries see a lift in performance once the logo is included in the design, but there are also cases when having the logo in paid social content may be overly salesy, pushy and lead to worse results.

With this insight, you can verify whether your specific audience reacts favorably to seeing your logo in the ad, or whether you should leave it out for better performance.

Primary color

The colors inside the creative is another extremely important consideration you can analyze in Confect. Different colors communicate different feelings and grab attention differently, so it’s a good idea to consider which colors resonate with your specific audience - based on data.

The primary color refers to the color that takes up the most space in the creative itself. This will be the dominant color that your ad uses, filling the majority of the design.

Primary saturated color

In comparison to the “primary color”, the “primary saturated color” doesn’t necessarily take up much of the design, but has the highest saturation and stands out from the rest of the ad.

An example of this could be a vibrant yellow element on an otherwise dark Black Friday ad.

It can help answer the question: “Which color grabs attention the best?”.


This insight is used to analyze and compare the performance of designs based on their brightness.

There are 3 possible values: light (bright content), neutral, and dark (low brightness).

For example, an ad with a soft white or yellow color will be categorized as light, an unedited photo will generally be neutral, and an ad using black or gray tones will be seen as dark.


Saturation refers to the vibrance and intensity of the colors inside the ad. 

There are 3 possible values: muted (low intensity and soft), natural, and vibrant (intense and sharp).

A vibrant ad will be very intense and bright, while a muted one will be reminiscent of softer, pastel-like colors.


The temperature of the content has 3 possible values: cold, temperate, and warm. Again, different colors have different emotional associations and perform differently for advertisers with different industries, or audiences.

Cold colors are: blue, green or purple. Warm colors are: red, yellow or orange.

A temperate ad uses a mix of warm and cold colors.