FAQ

In short

Short answers to the questions customers ask most often, grouped by theme. Each answer links to the page that explains it in full.

Short answers to common questions. Each one links to the page with the full explanation.

Results & KPIs

What is the E-Score: a synthesis or the main KPI?

Both. The E-Score is a synthesis because it measures nothing beyond its two components, Emotional Engagement and Attention. It is also the headline KPI that Foresight uses for lists, sorting, Compare and reports. Use it to rank and triage, and read its two pillars to understand a result.

How is the E-Score weighted?

The E-Score combines Emotional Engagement and Attention, so a high E-Score needs strength on both. Foresight doesn't publish the exact weighting; both parts matter, and a very strong pillar cannot fully hide a weak one. Valence (emotional tone) is not part of it, and the Mental Demand number is not used directly.

How is Emotional Engagement calculated?

Foresight combines all eight predicted emotions into a single intensity value, with higher-energy emotions counting more than calmer ones. Whether an emotion is positive or negative plays no part. The intensity is then positioned against reference creatives on a 0–100 scale, where 50 is roughly typical.

What does Clarity measure? Is it about comprehension or legibility?

Clarity measures how concentrated and how grouped predicted attention is on your creative: visual organisation and hierarchy. It is built from Focus and Spread. It does not read your text, check legibility or judge whether your message is understood, so a confusing message can still score high if the eye goes to one place.

What is the difference between Focus and Spread?

Focus measures how small an area of your creative draws predicted attention. Spread measures how close together the main points of attention are: it is highest when attention gathers on one main area, and more competing areas lower it. Together they make up Clarity.

Why do narrative or story-driven creatives get a very low Mental Demand?

Mental Demand is driven entirely by visual complexity: the number and density of elements, the detail and the clutter on screen. Storyline, dialogue, audio and the meaning of the copy have no weight. A demanding story told with clean, simple shots therefore reads low, which is expected rather than an error. The recommended zone is 25–75.

Is a score of 65 a percentage of viewers?

No. Most Foresight scores are rankings: they say how much of the reference library your creative is ahead of, with 50 in the middle. A 65 means your creative ranks above most reference creatives on that measure, not that 65% of people reacted in a certain way. Valence is the exception: it describes tone on a −100 to +100 scale.

Why is a creative I like only amber?

The colour bands are fixed: green is 70 and above, amber 40–69 and red below 40. Because scores are rankings against real creatives, a typical creative lands in the amber band. Amber is not a failing grade or "below average".

Emotions

Does Valence affect Emotional Engagement?

No. Emotional Engagement measures how strongly people are likely to react; Valence measures whether the reaction is positive or negative. Valence is not an input to Emotional Engagement, and Engagement is not an input to Valence: they read the same eight emotions in two different ways. Valence is not part of the E-Score either.

Is a strong negative emotion considered less engaging than a strong positive one?

No. A strong negative emotion counts like a strong positive emotion of similar energy, and a negative emotion is never marked down for being negative. Differences come from energy: tense, fear-driven ads usually score high, while quietly sad ads often score lower because sadness is a calm emotion, not because it is negative.

Which emotions does Foresight predict?

Eight evoked emotions: amusement, anger, awe, contentment, disgust, excitement, fear and sadness. Four are broadly positive and four broadly negative. They describe the feeling your audience is likely to have, not the emotions shown by people in the image.

Why don't the eight emotion values add up to 100?

Each emotion is a predicted strength from 0 to 100, predicted independently of the others. Several can be high at once, and all can be low for a neutral creative. They are not shares of a whole, and "40" does not mean 40% of people feel that emotion.

Does music or voice-over affect the scores of a video?

Not directly. Sound is not analysed directly; music and sound effects do not affect the scores. Spoken words can be reflected indirectly when the content description mentions them. Judge audio with other methods.

Attention

Is the heatmap real eye-tracking?

No. The attention model learned from real eye-tracking studies, but no one's eyes are recorded when you analyse a creative. The heatmap, fogmap, gaze plot and element shares are all predictions of where a typical viewer is likely to look.

Does Foresight detect faces, logos and text automatically?

No. In the AOI tab you draw a rectangle around an element, then give it a type and a name. Areas of type Face, Logo or Text get a reference value for that element type; a Custom area is measured the same way but without a reference.

Which benchmark is used for faces, logos and text?

Each area of type Face, Logo or Text is compared with the typical attention share of that element type at a similar size relative to the image. This reference is the same for every sector and platform, and the app labels it as a general benchmark, not a sector one. It is shown for images only, on areas whose type is Face, Logo or Text; custom areas have none.

Can I compare face, logo and text results with each other?

Within one creative, yes: raw attention shares are all percentages of the same creative's attention, so "the face draws more attention than the logo" is a valid reading. Gaps to the reference are a different matter. Each element type has its own yardstick, so do not rank face, logo and text by how far they sit above or below their references.

Does the gaze plot have a benchmark, and does it show where people look first?

The gaze plot has no benchmark: it describes your creative alone. Its numbered stops mark the main hot spots of predicted attention, joined in a suggested sequence that you should treat as indicative only. All stops are the same size, so they do not show viewing time.

Benchmarks

What are my scores compared against?

Every score is positioned against a large reference library of real-world images and videos, analysed with the same models as your creatives. If you pick a comparison group, the group's average is shown next to your score as a "+N / −N" difference. The group only sets that comparison value; it never re-scores your creative.

Can I compare against my sector or category?

For images, yes: pick a creative type (Advertising or Packaging) and, optionally, an industry sector. Only groups with enough reference creatives are offered, so some combinations can appear unavailable. Images analysed inside a platform have no type or sector groups, and videos show no comparison value.

What do the score colours mean? Is there a good / average / poor scale?

The colours are fixed bands: green for 70 and above, amber for 40–69 and red below 40; on video scenes they read Good, Medium and Low. They do not depend on your comparison group. A creative can be above its group average and still amber. Mental Demand uses its own 25–75 optimal zone, and Valence has no bands.

What happens if I don't select a comparison group?

Your scores are still computed and positioned in exactly the same way, against the reference library of the same media type. Foresight simply shows no comparison value and no "+N / −N" difference. The same applies to every video and to images analysed inside a platform. Without a comparison group, AI insights are less reliable about how scores compare with typical creatives.

Why did the benchmark value on an older analysis change?

It shouldn't. Your scores are calculated once and never change, and the comparison values next to them come from the fixed benchmark generation the analysis was run against. If two values differ, check whether you are looking at a re-analysis: a new analysis of the same creative uses the current generation, which can have different group averages.

Platforms (frameworks)

Why does Clarity change a lot with TikTok while the other KPIs stay stable?

With TikTok, attention is simulated on your video inside the full app screen, and the interface competes for attention, which Clarity is built to detect. Emotion is measured on your original upload, with only a small, fixed contribution from the TikTok interface added, so it moves only slightly. Attention and the E-Score usually move less than Clarity, because Clarity is only one of the things they are built from.

Does picking a platform change the algorithm, the criteria or the benchmark?

The models, the formulas, the reference library and the colour bands stay the same. What changes is the screen the attention model sees, so the attention KPIs and the E-Score can move. For images, picking a platform also removes the type and sector comparison, so no comparison value is shown; videos never show one.

Does the campaign objective change my scores?

No. The objective (Brand building, Consideration, Conversion or Trust) only shapes the AI insights: the summary and recommendations are written around that goal. Scores, heatmaps, benchmarks and colour bands are the same whatever objective you choose.

Images vs video

Are image and video KPIs directly comparable?

Not directly. Images and videos share the same KPIs, the same 0–100 scale and the same colour bands, but every score is ranked only among creatives of the same media type: an image among images, a video among videos. A 60 for an image and a 60 for a video therefore mean "above the typical image" and "above the typical video", not the same thing.

How is a video's overall score calculated?

Foresight predicts emotion on every frame and attention on frames sampled throughout the video, then summarises the whole timeline with every moment counted the same. The overall score is therefore not the video's best moment: long videos soften short peaks. Use the timeline and the scene scores to see the strong and weak moments.

Why can't I put an image and a video side by side in Compare?

Compare only accepts creatives of the same media type, because image and video scores sit on different benchmarks and are not directly comparable. The first creative you add sets the type. The same rule applies to exports.

Is an animated GIF analysed as an image or a video?

As a video. A GIF gets a timeline and scenes, and it is compared with videos.

Using Foresight

I can't find campaigns I tested a few months ago. How do I recover them?

Foresight has no separate archive: every analysis is kept indefinitely in Projects(opens Foresight in a new tab) until someone deletes it. Search the project name and press Enter (this also searches inside folders), open your folders, scroll or sort by Oldest, and check "Shared with me" in case a colleague ran it. If it still doesn't appear, contact your Emotiva account manager or support with the campaign name and approximate test date.

Do my old results change when Foresight is updated?

No. Each analysis keeps the scores and heatmaps it was produced with. The comparison values next to the scores stay fixed too. The models may change over time, so comparing a new analysis with an old one is not recommended: re-analyse the old creative as a new analysis instead.

Why does an older analysis show "This study has no AI insights"?

Analyses made before AI insights were introduced open normally, with complete scores, heatmaps and benchmark comparisons, but have no written insights. To get insights for an old creative, upload it again as a new analysis. If insights could not be generated for a new analysis, the Insights tab says "We couldn't generate the AI insights" and offers Regenerate insights (run them again) and Report a problem (tell the Foresight team; you can send another report after a short wait). Scores, heatmaps and benchmark comparisons are complete either way.

Do AI insights change my scores?

No. Copilot uses your scores and the content description and explains them in plain language, with prioritised recommendations. It never changes a score. Treat its suggestions as expert advice to test, and analyse a revised creative to measure the real effect.

Can I change the platform or comparison group after an analysis has run?

No. Objective, platform and comparison group are fixed once the analysis starts. To try different settings, upload the creative again, which uses another analysis from your plan.

What uses an analysis?

Each creative you analyse, image or video, uses one analysis when it starts. Your plan is managed by your Emotiva account manager.

FAQ ​

Results & KPIs ​

What is the E-Score: a synthesis or the main KPI? ​

How is the E-Score weighted? ​

How is Emotional Engagement calculated? ​

What does Clarity measure? Is it about comprehension or legibility? ​

What is the difference between Focus and Spread? ​

Why do narrative or story-driven creatives get a very low Mental Demand? ​

Is a score of 65 a percentage of viewers? ​

Why is a creative I like only amber? ​

Emotions ​

Does Valence affect Emotional Engagement? ​

Is a strong negative emotion considered less engaging than a strong positive one? ​

Which emotions does Foresight predict? ​

Why don't the eight emotion values add up to 100? ​

Does music or voice-over affect the scores of a video? ​

Attention ​

Is the heatmap real eye-tracking? ​

Does Foresight detect faces, logos and text automatically? ​

Which benchmark is used for faces, logos and text? ​

Can I compare face, logo and text results with each other? ​

Does the gaze plot have a benchmark, and does it show where people look first? ​

Benchmarks ​

What are my scores compared against? ​

Can I compare against my sector or category? ​

What do the score colours mean? Is there a good / average / poor scale? ​

What happens if I don't select a comparison group? ​

Why did the benchmark value on an older analysis change? ​

Platforms (frameworks) ​

Why does Clarity change a lot with TikTok while the other KPIs stay stable? ​

Does picking a platform change the algorithm, the criteria or the benchmark? ​

Does the campaign objective change my scores? ​

Images vs video ​

Are image and video KPIs directly comparable? ​

How is a video's overall score calculated? ​

Why can't I put an image and a video side by side in Compare? ​

Is an animated GIF analysed as an image or a video? ​

Using Foresight ​

I can't find campaigns I tested a few months ago. How do I recover them? ​

Do my old results change when Foresight is updated? ​

Why does an older analysis show "This study has no AI insights"? ​

Do AI insights change my scores? ​

Can I change the platform or comparison group after an analysis has run? ​

What uses an analysis? ​

FAQ