1 of 17

CulturePulse

Predict Human Behavior

Lightweight

Cost-Efficient

Psychographic Mapping

September, 2021

All commercial information and intellectual property contained in this document are strictly confidential for the client, its employees, and its collaborators.

2 of 17

Overview

If you know someone’s beliefs, you can predict their behavior.

We predict human behaviors with quick and affordable algorithms that map beliefs and provide more actionable market insights.

3 of 17

Problem

Companies are making decisions by counting likes and views but can’t access the beliefs that drive those customer behaviors.

Popular AI models (e.g., BERT and GPT) are costly to deploy and require constant retraining. Historical analysis fails to predict emerging trends and unprecedented events because the training data does not yet exist.

The market lacks quick, flexible, and affordable systems �for predicting how humans will receive new information.

4 of 17

Opportunity

More data doesn’t guarantee better insights.�Better algorithms guarantee better insights.*

CulturePulse has developed algorithms that map �beliefs and values and enable the prediction of human behavior.

These psychographic maps don’t require constant retraining �and work with any text in over 40 languages.

*Greco, Polonioli, & Tagliabue (2019). “Less (Data) Is More: Why Small Data Holds the Key to the Future of Artificial Intelligence“. https://arxiv.org/abs/1907.10424

5 of 17

Solution

We compare how a message fits with a market by profiling the market, modeling belief systems, and simulating thousands of responses.

We translate how the message fits the market into a ‘Resonance Score’: a single actionable metric for measuring the success of a message.

Based in evolutionary psychology, not machine learning, the resonance score is a metric for finding audiences where a message resonates the most and finding the audience that is more likely to share the message.

* The CulturePulse system is 130mb. BERT is > 1.3GB and can be > 4GB when trained. GPT2’s deployment is >5Gb.

6 of 17

Technology

  • >100x more efficient than traditional models.*
  • Instantly functional across 40+ cultures without any historical data.
  • Goes beyond sentiment and includes anger, anxiety, personality, morality, family, friends, finances, inclusivity, racism, and 50+ other categories**.
  • Understanding of concepts can be varied by concept without retraining the entire model, unlike models such as BERT and GPT***

Developed from 10 years of lab research and provides accuracy levels that are acceptable in clinical research settings (>95% reliable correlation).

Proven to provide deeper insights in multiple industries including social, finance, entertainment, security and defense, government, etc. over 4 years of consulting (revenue generated and reinvested in development).

* The CulturePulse system is 130mb. BERT is > 1.3GB and can be > 4GB when trained. GPT2’s deployment is >5Gb.

** See full taxonomy for all categories in CulturePulse’s ontology on Slide 25

*** This is also true for all word embedding systems, which do not have definable representations for specific elements in the array

7 of 17

Business Model - SaaS

Consumer Products

  • Self-Service�
  • $25-150 p/mo.�
  • Platform access�
  • A la carte modeling services

Business Products

  • Managed Service�
  • $5-50k p/mo.�
  • Modeling services�
  • API Access

API Licencing

  • Single domain or multi-domain*�
  • AWS hosted or private hosted�
  • $2k/mo/dimension or $10k/mo for all dimensions�
  • Modeling services

*Domains include personality, morality, sentiment, social concerns, etc. See slide 32 in Appendix for full list of dimensions

8 of 17

Use of Funds

We’re raising a $500K SAFE with three goals:�

  • Develop product and sales pipeline — Create SaaS platform for existing technologies and APIs for platform partners.�
  • Ramp up sales — develop marketing materials, scale inside sales team and develop outbound sales practice.�
  • Emphasize thought leadership — Drive lead generation via content marketing and research partnerships.

9 of 17

Financials

  • We expect profitability in 12 months.
  • Main expenses through 2025 will be sales and marketing
  • Costs kept down by utilizing in-house teams for R&D and development in Central and Eastern Europe (team leads in Bratislava and Zagreb, already in place)

10 of 17

Case Studies

11 of 17

#1 Electric Vehicle Consumer Concerns

Problem

EV adoption is lagging in the US compared to the EU and consumer concerns are hard to track. Nationally representative surveys are expensive and time-consuming.

Strategy

Track conversations related to EVs on social media, and query a digital twin with key topics to see what will be most concerning to the market.

Results

CulturePulse was able to replicate market concerns found in nationally representative surveys. Results were provided in hours (not days).

Key Market Concern

Our Predicted Tweet Rank

% Ranking Concern as Most Important (Survey)

Cost/Price premium

1

26%

Driving range

2

24%

Lack of charging infrastructure

3

22%

Time required to charge

4

10%

12 of 17

#2 COVID-19: Norwegian Government

Problem

Emotional contagion and misinformation were spreading quickly in Scandinavia related to COVID-19.

Strategy

Track 24 COVID-19 related hashtags using CulturePulse’s social listening system to understand what makes COVID-related tweets successful.

Results

Increased strategic knowledge for Norwegian ministry of health related to the emotional signatures of successful information and what kinds of topics the population was concerned with. Helped shape communications in later waves of the pandemic. Able to create a digital twin of the Scandanavian Twittersphere to replicate real-world results.

13 of 17

#3 Market Fit For Scripts

Problem

Entertainment marketing agency wanted to predict a script’s performance in specific markets.

Strategy

Utilize CulturePulse for testing messages to see if the script is compatible with Netflix audiences.

Results

Found that in key areas related to cultural beliefs and values, the script resonated with key subsets of the Netflix user-based (from Reddit data). Script sold to Netflix for USD20m.

14 of 17

#4 Market Fit For Scripts

Problem

Entertainment marketing agency wanted to test script compatibility with leading Hollywood actor.

Strategy

Utilize CulturePulse for testing messages to see if the script is compatible with public perception of the actor.

Results

Found that the top choice for the actor was not the best choice for script fit. Movie moving to pre-production and casting.

15 of 17

The Team

Justin E. Lane

D.Phil, University of Oxford

CEO, Co-Founder

F. LeRon Shults

2x Ph.D, Princeton

CRO, Co-Founder

Brian Swichkow

7 years as Growth Hacker

CMO

Maja Roknic

7 years as PM

Lead Product Manager

16 of 17

Advisors

John Zdanowski

Former CFO at Second Life, General Assembly, Assembled Brands, etc.

Jim Goodman

Creator of Ticketmaster Online and Tickets.com, formerly with Disney and Intel

Will Cady

Global Director of KarmaLab at Reddit

17 of 17

“No one ever made a decision because of a number. They need a story.”

― Daniel Kahneman

(2002 Nobel Prize Winner: Economics)