HowGood’s Latis impact platform houses information on the environmental and social impact of over 33,000 ingredients, chemicals, and materials globally.

Through an ongoing process of exhaustive data collection, analysis of peer-reviewed science, and a progressive heuristic approach to mapping and assessing the data collected, HowGood has developed the world’s largest product and ingredient sustainability database.

The foundation of HowGood’s data is a diverse and continuously updated collection of data sources, including peer reviewed journal articles, academic conference proceedings and texts, aggregated commercial databases, targeted industry studies, NGO research, and government publications.

We use a mix of qualitative and quantitative data sources and for each source, we perform a data certainty assessment based on the age and comprehensiveness of the findings. This process is completed for every impact metric in the HowGood system, and for every ingredient on which there is accurate and verifiable data.

One of these data sources is the USDA National Organic Program, which provides:

And that’s just one of 600+ data sources that HowGood uses.

Once the data is collected and analyzed, we map every single ingredient to its source crop animal or material. Using global import/export data and HowGood industry partnerships, we then map each source crop to its corresponding geographic location to account for the specific on-the-ground practices, impacts, and risks in each locale.

At this point in the process, we have a map of nearly every ingredient, chemical and material in the CPG industry and where and how it is produced. This map is used to aggregate data across geographic regions or ingredient categories and develop industry-average impact profiles for each metric and every ingredient.

Based on the ingredient mapping process, the Impact Platform assigns a default location and corresponding industry-average profile for every ingredient in a product. If deeper levels of data granularity are available (from a specific supplier, industry partner, or publication), these specifics can be applied to override the industry average values.

With each new data partner, publication, or customer relationship, we receive new insights that are then integrated into our larger systems map. Over time we build up more and more and more granularity. If for some reason, we are unable to find a perfect match for a particular ingredient, we use an internal proxying protocol to identify the most appropriate comparable data.

Perhaps the most important step in the HowGood methodology, we harmonize all the data collected on any single impact metric by plotting each practice along a single line: the impact spectrum. This is done while keeping the entire CPG ecosystem in mind, enabling a clean transition from theoretical to practical application.

On the degenerative end of the spectrum, we find damaging, extractive, oppressive, and/or abusive practices. Sustainable, the mid-point, is a "net-zero" perspective, that doesn't cause harm but also doesn't improve. Regenerative not only avoids harm, but it also improves, develops, and heals.

Using Labor Risk as an example, on the degenerative end you would find forced child labor, forced labor, bonded migrant labor, critical low pay, and permanent debt. Toward the middle of the spectrum, you might find the minimum viable income, critical debt, and consistent risk of loan default. Regenerative could include living wages, strong and healthy unions/collective bargaining, and socioeconomic development.

With the full spectrum of impacts in place, we determine the thresholds to set for each score. Most spectrums are not divided into even quintiles, but rather are carefully portioned out to avoid a bell curve of impacts with the majority of products falling into the middle.

Even quintile distribution is shown above, HowGood's threshold distribution is shown below.

The thresholds are set with one primary goal in mind: to deliver practical, actionable insights for differentiating between two ingredients or whole products.

Each quintile represents a score bracket of 2, which adds up to a score out of 10 for each metric. Each metric is then weighted equally and rolled up into a HowGood Impact Score out of 100.