The trade in high-value natural resources, particularly exotic plants and animals (henceforth biocommerce), can lead to both positive and negative impacts on wild populations. The exotic plant and animal hobbies are highly organized with internal cultures that are often poorly understood by non-members; however, understanding them is of central importance for biocommerce with both direct and indirect consequences for the conservation of wild populations.

Growing international demands for exotic species leaves simple questions related to the economics of biocommerce unexplored.  The long-term success of biocommerce businesses requires constant income, meaning they obtain and retain significant market proportions.  Some biocommerce products can be reproduced by hobbyists, and undercut business efforts based on sustainably produced exports.

We analyze a test case in Neotropical poison frogs where sustainable biocommerce is burgeoning, yet (export-based) businesses face direct market competition from the offspring produced within the countries which receive their exports, and therefore are forced to rely on fervent customer support. 

We survey customers’ behaviors where we find different classes of consumers, which differ in where they place higher values within the hobby. By taking into account these different classes we gain insights and potential solutions to promote the stability of this unique type of biocommerce.

Survey design: We created an online survey on the SurveyMonkey.com© platform to assess consumer preferences and purchasing behaviors (human ethics approval granted from UDLA CEISH committee project code: JY-210423-001). Our questions aimed to identify specific customer preferences related to their hobby-related interests, purchasing decisions and price stability, and specifics related to the direct competition dynamics between biocommerce animals and their offspring in shared markets.  We assess whether broader ethics such as supporting conservation, legality, or sustainability could influence consumer decisions.  Three types of questions were used in the survey to be able to quantify consumer preferences (Hinsley & ’t Sas-Rolfes, 2020).  Multiple-choice questions contextualized the individual’s participation within the hobby with regards to the diversity of species kept, time in the hobby, and annual spending.  Additionally, questions with ranked importance (from 1-5) were used to assess the importance of provenance and legality of purchases, the importance of price in dictating purchasing decisions, the importance of status (gained from owning rare species), and their interest in sustainability. Finally, short answer and multiple-choice questions sought to clarify details related to offspring selling expectations, including understanding the competition between biocommerce offerings and in-country produced offspring.

Surveys were distributed broadly through social networks, with personalized invitations sent to active hobbyists.  For two months repeated social media posts were made in specialized groups, and we created an Instagram profile for the survey which advertised the questionnaire using daily posts that featured in situ images of rare species of poison frogs to draw in participants.  We acknowledge potential downsides of our sampling: first, our questionnaire excludes hobbyists (albeit likely a minority) who are not active on platforms; second, responses were voluntary and required participant motivation, and may underrepresent casual hobbyists; and third, although surveys were anonymous, participants’ replies may be biased when providing responses that would be looked down upon by the hobby (Barnett, 1998)—relevant to questions related to their opinions related to the legality of animals, or the importance of the “status” in the hobby. 

Data analysis: We built a Logit model of latent classes (LCM) using Stata statistical software® (version 15.1).  We aimed to define types of hobbyists and define their interests within the hobby. This type of model has been shown to successfully identify heterogeneity of preferences (Hinsley et al., 2015). Due to homogeneity in ranked importance responses, a Logit model was developed with the variables: price (importance of price purchasing decisions), status (importance of rare animals/social status), tendency to purchase in-country offspring, and importance of (re)sales (ability to generate revenue from offspring). This model identifies the variables that best explain the preferences of each latent group or class, but does not determine the number of these, so statistical information criteria must be used to define discrete groups (Birol, Villalba, & Smale, 2009) where we used the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC), for modeling latent class models (MacDonald, 2018).

A latent classes model was built encompassing questions related to animal prices, hobby-status (owning rare/expensive animals), likelihood of directly purchasing/supporting biocommerce businesses, and resale potential using binary variables. In the last category data were averaged scores of importance, from different questions that were asked in the survey, these values ​​were considered as 0 if they belonged to the range between 1-3 and 1 if they belonged to the range between> 3-5, apart from the variable resale which was considered false = 0 and true = 1.