Intro
The still somewhat-burgeoning cannabis industry has issues adequately matching products to customer tastes and desires due to a complex interplay of cannabinoids and individual psychological responses. More traditional methods of capturing user preferences (based on simple questionnaires) often fail to take into account nuanced effects that different cannabis strains can have on users, while those who have tried come up short. This incongruity can lead to poor first experiences which may deter people from trying the product again. This product aims to refine the onboarding and product selection for cannabis consumers by leveraging a more intuitive and user-friendly interface which utilizes multimedia inputs to better understand user preferences.
The Journey
The customer journey was broken into seven distinct steps which are explained in greater detail below.
Stakeholder Requirements
Embark (Matching Precision): Kicks off the user’s tailored experience, setting the foundation for precise matching by capturing initial interest and preferences.
Introduce (User Interaction Quality): Invites users to share detailed preferences in an intuitive setup, ensuring the quality of interaction is high and information is rich for accurate matching.
Engage (Accessibility): Delivers a multi-sensory engagement platform designed to be navigable and enjoyable for all users, regardless of ability.
Analyze (Growth Capability): Features robust back-end processes that handle increasing amounts of user data, ensuring scalability and the capacity for complex data analysis.
Discover (Potential for Recommendation Errors): Utilizes advanced algorithms to minimize the margin of error in strain recommendations, ensuring users discover only the most compatible options.
Set Sail (Privacy & Security Standards / Regulatory Compliance Flexibility): Wraps up the journey with a secure option to save and share feedback, ensuring data privacy, and facilitating easy updates to maintain compliance with cannabis regulations.
Implementation Expense: Not explicitly represented in the states, but implicit in the design's efficiency and the streamline of processes which aim to keep operational costs in check.
Documentation
State Descriptions and Transitions
Embark
Purpose: To welcome users and set the tone for a personalized experience.
User Input: Users begin by clicking 'Start' after a brief intro video or text explanation.
System Output: Transition to the 'Introduce' state.
Transition Trigger: User clicks the 'Start' button.
Introduce
Purpose: To collect initial user preferences and past experiences.
User Input: Users answer preliminary questions through a simple form or selection interface.
System Output: Compilation of user data and transition to the 'Engage' state.
Transition Trigger: Submission of the introductory form.
Engage
Purpose: To capture detailed user preferences through interactive multimedia content.
User Input: Users interact with images, music, and video, providing feedback on preferences.
System Output: Detailed user profile update and transition to the 'Analyze' state.
Transition Trigger: Completion of all interactive elements.
Analyze
Purpose: To process user interactions and formulate strain recommendations.
User Input: Implicitly provided through prior interactions.
System Output: A curated list of cannabis strains recommendations.
Transition Trigger: Successful analysis and processing of user data.
Discover
Purpose: To present personalized cannabis strain recommendations to the user.
User Input: Review of the recommendations.
System Output: Detailed information on each recommended strain and the option to provide feedback or save preferences.
Transition Trigger: User selects a recommendation for more details or opts to provide feedback.
Set Sail
Purpose: To gather user feedback and conclude the current session with an option for continuation.
User Input: Feedback submission, saving preferences, or choosing to start a new session.
System Output: Feedback acknowledgment, saved preferences, or reset of the user journey.
Transition Trigger: User selects 'Finish' or 'Restart'.
Interface Elements
Buttons: Present in each state for navigation ("Start", "Submit", "Next", "Finish", "Restart").
Forms: Used in the 'Introduce' state for initial data gathering.
Media Players: Utilized in the 'Engage' state for interactive content delivery.
Recommendation Display: In the 'Discover' state, showing the strains and detailed descriptions.
Feedback Module: Available in the 'Set Sail' state for users to provide their insights and preferences.
Algorithms
Matching Algorithm: Employed in the 'Analyze' state, it could use machine learning techniques to find patterns in user responses and match them with strain profiles, taking into account the chemical properties of strains, user reviews, and empirical effectiveness data.
Feedback Algorithm: Used in the 'Set Sail' state to integrate user feedback into the system, enhancing future recommendations and adapting the system to user evolution.
References / Sources
- Sharp, H., Preece, J., & Rogers, Y. (2019). Interaction Design: Beyond Human-Computer Interaction. United Kingdom: Wiley. Chapter 11.
Why It's Better
Behind the shiny, intuitive UI lives a proprietary recommendation algorithm powered by machine learning techniques that analyzes complex datasets of user responses, cannabinoid/terpene profiles, and user feedback. Many recommendation engines fall short in that they are overly constrained by the concept of phenotype which are roughly built around two dominant species of the genus Cannabis. There are no longer “pure” C. indica or C. sativa plants on Earth (except perhaps deep in the Hindu Kush region), so a better way of classification for therapeutic purposes is needed. The Cannabis Preference Interface (CPI) will use a more nuanced approach to categorizing cannabis products based on their chemical profiles, which will be used to make recommendations based on user preferences.
The system follows guidelines set forth by the WCAG (Web Content Accessibility Guidelines) for ensuring inclusivity on web applications, as well as considerations for neurodiversity. Some users may not be able to see or hear, so this tool can provide usefulness even if one of the two sensory inputs aren’t available.
Evaluation of Alternatives
Criteria for Evaluation
- Matching Precision: The effectiveness in aligning user preferences with cannabis strains.
- User Interaction Quality: Measures how engaging and intuitive the solution is for users.
- Accessibility: The extent to which the solution is usable by people with a wide range of abilities.
- Growth Capability: The ability of the solution to accommodate an increasing number of users and data complexity.
- Privacy & Security Standards: How well the solution safeguards user information.
- Implementation Expense: The upfront and ongoing financial requirements of the solution (higher is less expensive).
- Potential for Recommendation Errors: The likelihood of the system to suggest inappropriate strains (higher is better).
- Regulatory Compliance Flexibility: The ease with which the solution can adapt to legal changes in the cannabis industry.
Alternatives
- A: Multimedia-based preference interface for strain recommendation.
- B: Standard questionnaire-based matching system.
- C: In-person, budtender-led strain selection.
Evaluation
A (Preference Interface)
Precision in Matching (20%) | 9 |
User Interaction Quality (15%) | 9 |
Accessibility (10%) | 8 |
Growth Capability (15%) | 8 |
Privacy & Security Standards (10%) | 7 |
Implementation Expense (10%) | 6 |
Potential for Recommendation Errors (10%) | 8 |
Regulatory Compliance Flexibility (10%) | 8 |
Total | 7.875 |
B (Questionnaire)
Precision in Matching (20%) | 6 |
User Interaction Quality (15%) | 5 |
Accessibility (10%) | 6 |
Growth Capability (15%) | 7 |
Privacy & Security Standards (10%) | 7 |
Implementation Expense (10%) | 8 |
Potential for Recommendation Errors (10%) | 4 |
Regulatory Compliance Flexibility (10%) | 7 |
Total | 6.25 |
C (Budtender)
Precision in Matching (20%) | 8 |
User Interaction Quality (15%) | 8 |
Accessibility (10%) | 4 |
Growth Capability (15%) | 3 |
Privacy & Security Standards (10%) | 9 |
Implementation Expense (10%) | 4 |
Potential for Recommendation Errors (10%) | 7 |
Regulatory Compliance Flexibility (10%) | 6 |
Total | 6.125 |
Overview
- Matching Precision: A's interactive and dynamic feedback mechanism potentially offers the highest precision, leveraging nuanced user inputs. B and C are more static, with C somewhat mitigated by the personal touch of budtenders.
- User Interaction Quality: A's multimedia approach is likely most engaging, providing a rich, user-centric experience. C's personal interaction is also engaging but less scalable. B is functional but less dynamic.
- Accessibility: A and B present opportunities for tailored accessibility features, with A potentially offering more innovative, inclusive interfaces. C's accessibility depends heavily on the physical store's accommodations (assuming no "virtual budtender" options).
- Growth Capability: A and B offer better scalability due to their digital nature, with A requiring more sophisticated backend support to manage multimedia content and process the results.
- Privacy & Security Standards: While all solutions pose privacy risks, digital solutions (A and B) can implement robust cybersecurity measures. C benefits from less digital data collection and the fact that budtenders generally have bad memory.
- Implementation Expense: A's advanced technology could entail higher initial costs, whereas B is less technologically demanding. C requires paying budtenders, which can get expensive.
- Potential for Recommendation Errors: A's reliance on interpreting user interactions and inferring personality traits introduces a risk of misinterpretation, whereas B and C rely on more established, but still fallible methods of understanding user preferences.
- Regulatory Compliance Flexibility: Digital solutions (A and B) can adapt quickly to regulatory changes through software updates, whereas C might struggle with more logistical adjustments.
Naturally, the proposed interaction-based recommendation system (A) scores highest in the evaluation matrix, indicating its potential as a superior solution in terms of engagement, scalability, and adaptability, despite higher implementation costs and the need for rigorous data privacy measures. Its success relies on effective execution, especially in areas of inclusivity and accuracy of recommendations. Future iterations should prioritize minimizing the risk of inaccuracy and enhancing the inclusivity of the design.
FAQ
How is this different from other recommendation systems?
The Cannabis Preference Interface (CPI) provides an easy-to-use UI that uses visual and auditory stimuli to infer a user’s subconscious preferences and associating them with potentially preferred cannabis effects.
What's going on in the background?
A recommendation algorithm powered by machine learning techniques analyzes complex datasets of user responses, cannabinoid/terpene profiles, and user feedback. Many recommendation engines fall short in that they are overly constrained by the concept of phenotype which are roughly built around two dominant species of the genus Cannabis. There are no longer “pure” C. indica or C. sativa plants on Earth (except perhaps deep in the Hindu Kush region), so a better way of classification for therapeutic purposes is needed. The CPI will use a more nuanced approach to categorizing cannabis products based on their chemical profiles, which will be used to make recommendations based on user preferences.
How do you make this personalized to me?
A system architecture that supports dynamic learning from user engagement, which allows for continuous refinement of recommendations based on feedback and evolving preferences.
I have issues either seeing or hearing, is this for me still?
The system follows guidelines set forth by the WCAG (Web Content Accessibility Guidelines) for ensuring inclusivity on web applications, as well as considerations for neurodiversity. Some users may not be able to see or hear, so this tool can provide usefulness even if one of the two sensory inputs aren’t available.
I'm worried that the wrong people will know I used this; is my data secure?
Ensure that industry-standards for user privacy are implemented in the final product, such as encryption and promises of nondisclosure in the EULA. Regulations are also very tight around accessing products dealing with cannabis, so age-restriction needs to be account for in the UI.
Who paid for this fancy thing?
The development was sponsored by a profit-seeking organization in an attempt to increase business from current customers, lure cannabis-users from other brands to which they may be loyal, and tap into a potential latent market of those curious to try the product but perhaps too intimidated by the perceived (potentially real) inaccessibility.
References / Sources
- Web content accessibility guidelines (WCAG) 2.1. W3C. (n.d.). https://www.w3.org/TR/WCAG21/
About
Hi, I'm Michael Poplin and I'm a senior in the Industrial and Management Systems Engineering program at Montana State University located in Bozeman. This site was the culmination of the semester project for our 'EIND 410: Interaction Design' course and the result of a rigorous design approach used by professional interaction designers in industry.
For better and/or worse, cannabis has shaken a good deal of its stigma and become part of mainstream American culture. It was legalized in Montana for all adults in 2021 and since then, there seems to more cannabis dispensaries than coffee shops on the main streets of towns. As someone who’s worked in the industry since its infancy, I am deeply invested in promoting the responsible use of cannabis. Like many things, it has the potential for abuse but it undoubtably has potential for therapeutic use. We cannot legally research the medical benefits quite yet, so it is difficult to understand the exact interplay of the biochemical compounds that make the plant special (Russo E. B., 2011).
The have been efforts over the years to build so-called “recommendation engines” that attempt to categorize individual batches of cannabis into "strains" based on chemical profiles, but the solutions always tend to come up short. While the concept of a strain is on scientifically shaky-ground (Stockton, N., 2015), it is my belief that shortcomings in preference engines are due to the lack of a proper interface for potential consumers to input their preferences. Most solutions utilize a questionnaire that asks a user to explain what sort of experience they’re looking to pair with their cannabis, but its becoming evident that people of particular temperaments are more likely to prefer certain strains more than others, perhaps due to brain chemistry (Weedmaps, 2022).
Science confirms different people like different art, and everyone gets impacted in their own unique way... right? How unique is it really? Do certain personality types tend to like the same kinds of art? It has been observed that certain personality disorders tend to have fans of more "heavy" music (Rentfrow, P. J., & Gosling, S. D., 2003). Conversely, high levels of "openness" tend to be associated with those who prefer abstract and modern art (Furnham, A., & Avison, M., 1997). By utilizing a multimedia experience that engages a person’s senses, it is hoped that we will be able to understand how the person's brain ticks and offer algorithmically-generated recommendations vetted by feedback from like-minded users.
References / Sources
- Russo E. B. (2011). Taming THC: potential cannabis synergy and phytocannabinoid-terpenoid entourage effects. British journal of pharmacology, 163(7), 1344–1364. https://doi.org/10.1111/j.1476-5381.2011.01238.x
- Weedmaps. (2022, June 20). Cannabis Terpenes & the entourage effect. https://weedmaps.com/learn/cannabis-and-your-body/terpenes-entourage-effect
- Stockton, N. (2015, August 27). Sorry, but the names for weed strains are kinda meaningless. Wired. https://www.wired.com/2015/08/sorry-names-weed-strains-kinda-meaningless/
- Rentfrow, P. J., & Gosling, S. D. (2003). The do re mi's of everyday life: The structure and personality correlates of music preferences. Journal of Personality and Social Psychology, 84(6), 1236–1256. https://gosling.psy.utexas.edu/wp-content/uploads/2014/09/JPSP03musicdimensions.pdf
- Furnham, A., & Avison, M. (1997). Personality and preference for the arts. Personality and Individual Differences, 22(4), 659-663. https://www.researchgate.net/publication/282855726
Try It Out
Are you ready to find out which products are best suited to your personality?
Get Started
Note: This is a simple, image-based demo with sample content and limited analysis.