Castillo or Caturra? A Simple Question

sheridan 1By Michael Sheridan, CRS Coffeelands

CRS is working with friends and allies in the specialty coffee and research communities to compare the cup quality of Castillo and Caturra, the two leading coffee varieties in the fabled Colombian coffee origin of Nariño, as part of The Colombia Sensory Trial. I presented the preliminary results of our research to the 2015 SCAA Symposium in a talk titled, “A simple question: Castillo or Caturra?” The guiding question behind the research may be simple, but the implications of the results are decidedly more complex.


Thanks to its resistance to rust and an integrated package of agronomic and financial incentives, Castillo has become the single most important cultivar in Colombia in just 10 years’ time.  Today it represents nearly 40 percent of the country’s coffeelands. Questions have persisted about Castillo’s cup quality, however, and Cenicafé’s reports of Castillo’s quality potential have been met with skepticism by many specialty coffee roasters. The debate over the sensory merits of Castillo has echoed loudly in Colombia’s coffeelands, leaving growers to make the important decision of which variety to plant under conditions of considerable uncertainty.


sheridan 2We partnered with the International Center for Tropical Agriculture (CIAT) in Colombia, World Coffee Research (WCR), the Sensory Analysis Center at Kansas State University (KSU) and cuppers from some of the world’s leading coffee organizations to create the Colombia Sensory Trial and inform decision-making at multiple levels:

  • On the farm, where growers seek greater clarity about the returns they can expect on their investments in different coffee varieties;
  • In the industry, where buyers are making purchasing decisions on the basis of partial information in a coffee landscape that is changing rapidly; and
  • In the policy process, where governments, coffee institutes, and public and privately funded projects determine which coffee varieties to promote, support, and subsidize through the provision of certified seed, extension support, credit, and grants and other services.

The preliminary results of the study are being delivered to all three of these communities of interest.


“A Natural Experiment”

Mark Lundy is a Senior Researcher at CIAT who was instrumental in helping us design the Colombia Sensory Trial as a “natural experiment.”   “A natural experiment,” he says, “happens when you identify and analyze interesting things that happen in real life—variation that you couldn’t randomize or examine as a controlled trial, but that occurs naturally.”

The conditions for a natural experiment exist in Nariño, where many farmers are growing, harvesting and processing Castillo and Caturra under nearly identical conditions. By selecting samples of each variety from 25 different farms in Nariño, we were (mostly) able to control for variability in terroir, husbandry, harvesting practices, and post-harvest processing, and to isolate the impact of variety on cup quality.

The Cupping Protocol

As part of the Trial, we enlisted Cup of Excellence veteran Paul Songer to lead two cupping panels at Intelligentsia Roasting Works in Chicago, the first in October 2014 and the second in January 2015. Leading coffee quality analysts from eight influential coffee brands participated in the panel:

  • Counter Culture: Timothy Hill, quality control director
  • Federación Nacional De Cafeteros: Néstor Perilla, coffee quality specialist
  • George Howell Coffee: Jennifer Howell, director of coffee sourcing
  • Keurig: Roman Bondarenko, coffee project manager
  • Red Fox Coffee Merchants: Aleco Chigounis, co-founder and president
  • Starbucks: Doug Langworthy, coffee quality manager
  • Stumptown Coffee: Adam McClellan, green coffee buyer
  • Intelligentsia Coffee: Geoff Watts, vice president of coffee

The quantitative results of these panels were subjected to analysis of variance tests to assess their statistical significance.

The Sensory Protocol

In addition to the traditional cupping panels hosted by Intelligentsia, the Sensory Analysis Center at Kansas State University [KSU] applied the new WCR sensory lexicon for the first time to samples from the trial.

From the more than 100 sensory attributes that constitute the new specialty coffee lexicon, KSU selected 35 key terms considered most applicable on the basis of a preliminary analysis of trial samples, previous experience with Colombian coffee, and a review of secondary literature. A panel of six highly trained assessors selected sensory attributes from the list to describe each sample. The results of these qualitative assessments were subjected to analysis of variance tests to assess their statistical significance.


Before sharing what the Colombia Sensory Trial told us, it is important to share what it did NOT tell us.

N = 25

It did not tell us about the potential of the two varieties for quality outside of Colombia. It did not tell us about the potential the two varieties have for quality in Colombian coffee origins beyond Nariño. And it did not even tell us about what the two varieties are capable of in Nariño for the 2015 harvest currently underway. With just 25 farms participating in the Trial, the sample size simply wasn’t big enough to generate statistically robust outcomes. That means that the results of the Trial are far from conclusive. They represent a single sensory snapshot taken with a narrow-angle lens.

No quantitative separation              

The data did not generate evidence of any significant separation between the two varieties in terms of average cupping scores, either overall or by category. The first round of the Trial in Chicago went to Castillo by the narrowest of margins, and Caturra took the second round by a half-point. Over the two panels, the average score for Caturra was just 0.3 points higher than the average score for Castillo. None of these narrow margins, either for overall cupping scores or for any of the specific sensory categories from which the overall scores were constructed, was statistically significant.

sheridan 4

Equal but different

Just because the two varieties earned similar scores doesn’t mean they taste the same.  In fact, the sensory assessors at KSU found significant sensory differences between the two varieties.

They evaluated Trial samples using 36 of the 108 attributes in the new WCR lexicon and found that all 36 attributes were present in both varieties.  In other words, it wasn’t that one variety presented one set of attributes and the other variety presented another.  The difference was one of degree: certain attributes were more intense in the Castillo samples and other attributes were more intense in the Caturra samples.  The differences in intensity, however, were statistically significant for only 13 of the 36 attributes: Castillo registered higher intensity than Caturra in 10 attributes and Caturra registered higher intensity than Castillo in three.  In summary, KSU did find evidence of statistically significant differences between Castillo and Caturra, but these differences came against the backdrop of a significant degree of overlap in the “sensory footprints” of the two varieties.


So, what can growers, policymakers, and coffee buyers learn from the results of the Colombia Sensory Trial?

What matters for growers

Castillo v. Caturra ≠ Quantity v. Quality

There has been a general sense among the growers we work with in Nariño that they can increase the probability of access to quality premiums by separating their Caturra into single-variety lots or by resisting the shift from Caturra to Castillo altogether. The overall results of this Trial call that assumption into question and suggest that growers who opt for Castillo over Caturra may be able to avail themselves of rust-resistant varieties without reducing in any statistically significant way the probability of producing higher-scoring coffee.

Where + How > What                               

For farmers choosing between Castillo and Caturra, what they choose to plant may have less impact on cup quality than where and how they grow it. The data from both the cupping panels at Intelligentsia and the sensory analysis at KSU show that cup scores are more strongly correlated with environment and management than with variety. In other words, growers may do more to increase cup quality through more active soil and shade management, careful harvesting, and improved post-harvest practices than through the intentional selection of one of these varieties over another.

Rethinking Caturra?

sheridan 3In a recent Roast Magazine article titled “The Geshas and the Rest,” specialty coffee pioneer Kenneth Davids assigns Caturra to a second tier of sensory quality and questions the viability of single-variety Caturra lots. He argues that Caturra and other traditional Latin American varieties “tend to produce a solid but conventional-tasting cup that may impress but does not stand out,” making it difficult for growers to succeed with a quality-based differentiation strategy based on these varieties. But just because Caturra didn’t achieve separation from Castillo doesn’t mean other varieties can’t. Growers who wish to achieve quality premiums could increase their chances of success with the cultivars Davids assigns to the top tier of quality: Gesha, Pacamara, Bourbon, SL28, and Maragogype.

What matters in the marketplace

Castillo can be (very) good

The single most important contribution of the Colombia Sensory Trial to the ongoing conversation about coffee breeding and cup quality may be the finding that Castillo can produce a good cup of coffee, and occasionally an extraordinary one, as the Castillo samples that earned 90 points or more attest. Against a backdrop of stubborn resistance in some segments of the specialty market to Castillo and other hybrid varieties, this is no small thing.

Prices, Profiles, and Probabilities

Cupping scores aren’t everything, as the KSU analysis suggests. Even buyers who award similar cup scores to both varieties may prefer one to another if it more consistently produces the specific cup profile those buyers seek. If that variety is Caturra, buyers may need to rethink their sourcing and pricing strategies. A single-variety purchasing preference may increase the probability buyers get the profiles they want, but will impose additional risks and costs on growers that may require both risk and quality premiums as well as the multi-year purchase commitments.

Who + How > What

In the marketplace as in the field, the Trial suggests that genetics may not be the most important determinant of success. For farmers choosing between Castillo and Caturra, what they choose to plant may have less impact on the prices they earn than how they sell it and to whom. More than 95 percent of Nariño’s coffee is sold exclusively on the basis of physical quality and almost never cupped as part of the price discovery process. This model is very good at bringing large volumes of good coffee to market, but not very good at rewarding growers who produce coffee of extraordinary quality. Regardless of what varieties they are planting, it is unlikely that large numbers of growers in places like Nariño will make significant investments in quality improvement until quality-focused business models that reward cup quality with price premiums are more prominent.

What matters for policy

The most good for the greatest number of growers

The Trial suggests that Colombia’s Castillo-only policy may be the best approach for the largest number of growers if it can reduce their exposure to production risk without compromising cupping scores in any statistically significant way vis-à-vis Caturra.

More intelligent extension and site-specific agriculture

The farms in the Trial that produced exceptional Castillo samples also produced exceptional Caturra samples, meaning that high quality was more a function of where farmers are planting coffee and how they are managing it than which variety they are planting. Some environments are simply better suited than others to produce high cup quality. The problem is that few growers know whether they are in one of those agroecological niches or not. We commonly use elevation as a proxy for quality potential, but are many other environmental variables that affect the suitability of a particular growing environment for high-quality coffee. Colombia has already collected much of this information in connection with its Denomination of Origin work, and may be positioned to support more “intelligent” or site-specific agriculture approaches that include information about the viability of a quality-based differentiation strategy based on a grower’s environment.

sheridan 5Michael Sheridan has worked in coffee for Catholic Relief Services since 2004. Since 2007, he has worked with smallholder farmers in the coffeelands throughout the Americas. Michael is currently based in Quito, Ecuador, where he directs the Borderlands Coffee Project in Colombia and Ecuador and advises other CRS coffee projects in Latin America and the Caribbean. From 2004-2007 Michael led the CRS Fair Trade Coffee Project. He publishes perspectives from the intersection of coffee and international development for the CRS Coffeelands Blog at