How It Works

Understanding the Hippoc solution


The 3 Basic Principles


Instant Attention

Instant Attention

See which parts of your ads attract your customers eyes in the first 3 seconds



Lasting Recall

Lasting Recall

See what parts of your ads get recalled by your customers



Impact Score

Impact Score

Combining Instant Attention & Recall, see which parts of your ads stand out in your customers minds


Results rooted in science

We tested 31,171 consumers

16M+ images and videos analyzed

Our predictive model is 94% accurate


We withdrew about 20% of our data set before training and validating our AI models to astonishing results: Hippoc scores are the highest achieved in our field

Prediction accuracy
for Lasting Memory Recall

Prediction accuracy
for Instant Attention


  • Our dataset has been collected according to the standards used in cognitive neuroscience research.
  • 31,171 consumers were exposed to thousands of images and measurements of attention and memory were taken using eye tracking or standard Memory test.


  • 31K consumers across North America and Europe were tested. 
  • Users were aged between 10 and 60 years old, with an average of 33.5 years.
  • 46% were women and 54% were men.
  • The 16M+ visual content pieces were shown to 80 consumers each.


Sociodemographic consistency

  • Using visuals we tested: We analyzed the cognitive responses of participants who saw the visual and compared it to the reactions of other participants, studying 70% for recall and 80% for attention.
  • This ensured consistency in our findings in both categories, despite the sociodemographic differences in our participants. 
  • What we found was that, over the course of a lifetime, the human brain is shaped by evolution rather than individual experience. 
  • In other words, some visual cues are more appealing and easier to remember than others—regardless of the consumer and target demographic.



We trained our algorithm using state-of-the-art machine learning and cognitive neuroscience methods.

This automated tool predicts what garners attention and can test what will really stand out in an ad. After all, how an ad performs isn’t just about relying on our attention, and its effectiveness can’t always be studied. 

The biggest takeaway is what a consumer remembers. And that’s where our proprietary Neuro-AI software comes in, an intelligent tool that accurately predicts attention span and memory. 

We’ve trained our machines to provide the outmost reliable predictions, without even having to collect additional data from your audience. What this means is that you can perform user testing even if no other human has laid eyes on your visuals. 

Our predictive analytics tool can be applied to all of your visual content, from digital and web ads, social media, print ads or out-of-home.


Our predictive model has an accuracy rate of 94% for instant attention.

  • Before training our Neuro-AI models, we randomly sampled 20% of our total data set and evaluated its accuracy (not used for learning).
  • Conclusions come from using the area under the curve (AUC Judd) metric, currently the main metric in MIT saliency benchmark.


We achieved 96% accuracy for lasting memory recall.

  • We validated the solution’s accuracy by comparing our prediction to the results, basing it on attention and memory recall. This method resulted in the best score attainable in our scientific domain: 94% for instant attention and 96% for lasting memory recall.



All data is anonymized and randomized to ensure the protection of respondents’ private data and to respect GDRP laws and regulations.