Warning: Creating default object from empty value in /nfs/c04/h03/mnt/68328/domains/vstapps.com/html/blog/wp-content/plugins/wp-e-commerce/wpsc-includes/purchaselogs.class.php on line 705
Underdevelopment- VST


One of the perks of being in this line of work is sampling many coffees from some of the world’s finest roasters. While in the process, I’ve had many tremendous coffees. More recently, however, I’ve observed a significant, damaging trend emerging among some roasters operating at the highest level. I speak of underdevelopment at roast.

While not strictly limited to ultra-light roasts, underdevelopment is certainly a more common defect of lighter roasts as compared to roasts taken further. While at first I was disappointed, lately I’ve been getting more dismayed with the number of specialty roasters delivering underdeveloped coffees.

I applaud and respect anyone who goes through the immense efforts required to run a specialty roasting business. Being in the roasting business can at times be an unforgiving, full throttle and exhausting effort that requires constant change and learning continuously for many years in order to develop all of the skills necessary to survive, let alone succeed. However, specialty roasters have a responsibility to taste and test their products before releasing them for sale, and setting a minimum standard of quality control. Yet, in the past few weeks I’ve received three premium espresso blends (out of six) that were markedly underdeveloped.

No one wants their products to shine more than the roasters themselves. This is why it’s so critical to add rudimentary procedures for QC that ensure only quality coffees are released for consumption. Refined further, simple QC procedures to check extraction levels at brew can help address underdeveloped coffees and aid in checking adjustments to roast profiles to address the issue.

One of the most noticeable symptoms of underdevelopment, apart from often tasting vegetal, is the requirement for drastic adjustments to grind, temp and or brew time to try to achieve good flavor and normal extraction levels. Despite unusual adjustments beyond the norm, coffees underdeveloped at roast will still not extract fully at brew, via any method. All things being equal, with coffees of similar roast levels, there should only be a reasonably limited range of adjustments necessary to dial it in; a couple of increments +/- on the grinder collar or a couple of degrees +/- in water temperature and/or +/- brew contact time. Jumping through hoops to get a normal extraction yield should raise flags.

Shots 1-4 were from original espresso (shipment 1). In this case the original espresso serves as the experimental control. Shots 5-8 are same espresso (shipment 2), same crop, same roaster, different date and batch. Brewing Protocol and equipment were identical for both batches.

The three premium espresso blends in question did not behave within any normal adjustment range, extracting to only 15-16% yield. In essence, the coffee behaves as if there is less coffee than dosed in the basket. The innermost core of the bean, the densest part, has failed to fully develop during roast, and is not able to deliver soluble solids –extract normally– the equivalent to the removal of a portion of extractable coffee. I frequently suspect the full batch didn’t properly make it through first crack before flame is reduced, but there are a myriad of ways coffee can be underdeveloped at roast. The “trend” to roast ultra-light is ruining as much as half of the coffees and espressos I have received from specialty roasters in the past several months, both from within and outside of the U.S. Bean color, and or agtron numbers are meaningless, they can all be spot on, but the coffee tastes and measures underdeveloped.

Most roasters are unaware they can use the Coffee Refractometer and MoJo to test their coffees, to confirm whether the roast is fully developed. MoJo will not provide any indication of “degree of roast”, but it will certainly measure accurately a roast that fails to extract normally. The following is a simple process that works and one any roaster can use as part of their QC routine:

1) Brew, taste and then measure the espresso and chart the result. If the coffee tastes under-developed and measures under-extracted using normal adjustments (see example charts), then continue with the next two tests.

2) Brew, taste and measure the espresso brewed as drip coffee. Many espresso roast profiles are slightly darker than those for drip, as such, an espresso is usually very easy to brew as drip well w/in normal drip adjustment parameters. Frequently, the water temp can be lowered slightly and or the grind coarsened because the slightly darker roast profiles (as compared to drip) allow a faster extraction to normal levels. However, if the same coffee that under-extracted when brewed as espresso also under-extracts when brewed as drip, it is highly likely there’s a problem with under-development at roast, rather than a problem with brewing equipment, settings or technique.

3) Finally, use an experimental control as your final test. That is, use a coffee “standard” that extracts properly using your normal brewing protocols, whether they be espresso, drip or both. For this we use Illy ICN Medium Roast beans. They are 100% Arabica, always roasted consistently and fully developed, and always extract as expected w/in normal adjustment ranges in either espresso or drip modes. These are available in 8.8-oz tins. We use the 3kg tins, which ensures enough to make it through a long series of tests w/o changing batches (variables).  Illy has incredibly reliable production roast level controls in place, and every batch we’ve tried always behaves predictably. If the control coffee taste returns to normal, and MoJo measurements confirm typical Extraction Yield and TDS for your standard shot, then you have confirmed with a high degree of confidence you have a coffee that was underdeveloped at roast, and no matter of brewing as espresso nor coffee will provide an acceptable result.

Any remaining coffee from the batch in question should be saved, and the roaster notified, in case they request it be returned to be tested [and hopefully addressed one way or another]. Unless you provide this important feedback, the roaster will be unaware there might be a problem. Obviously, you must be careful during the process to confirm your protocols and measurement are consistent. That said, this quickly uses up a small 250-gram or 12-oz bag. Keep the feedback positive, and document your results.  This can only help future products are improved.

Same Espresso brewed as Drip Coffee. Note similar Extraction Yield

I regularly chart and save most of the espressos and coffees I brew using either ExtractMoJo or using MoJoToGo. ExtractMoJo allows multiple plots on a single chart, which is useful for direct comparisons. MoJoToGo allows the full set of measurements, and detailed notes which may be emailed to yourself or any other MoJo user, and automatically added to the database of measurements for future reference. You can also export all coffee and espresso measurements from MoJoToGo as an excel .CSV file for database use outside of the APP. I add roast dates and if available batch numbers to the Notes. I frequently find the same coffee from the same roaster and from the same crop is delivered with vastly different [effective] roast profiles (see chart). In the case of  underdeveloped roasts, my previous brews of the same coffee that extracted normally can serve as the experimental control coffee.

It is understood that very light roasts often take longer to extract during brew (or require slightly finer grind and sometimes hotter brew temps) to normal levels than a comparatively darker roast. Making severe brewing adjustments in order to “coax” a normal extraction by using very high water temps and or extremely fine grinds can be done, but should raise flags if you find such severe changes in degree of adjustment is necessary. In my experience, this results in even more pronounced taste defects, but under no circumstances have I ever found it to result in an acceptable final quality. This, I think, is where most [perhaps less experienced] baristas fall prey to a roaster’s defective product, and what leads to the roaster never being taken to account for the lack of consistent roast quality in their product(s). The term “candied lemon” is, to me, too frequently an excuse to describe a failed roast. Candied implies a balancing level of sweetness with citrus, and can and does frequently accurately describe some  fully developed, light roasted coffees, but these underdeveloped coffees are extraordinarily sour, grassy, vegetal– “candied” is more often not even in the equation.

Upon discussion with various roasters about this problem, and having designed installed and used sophisticated roast profiling hardware and software systems for small batch roasters, a few common threads prevail. One refers to a previous [excellent] roast being from a different (i.e., normal) batch size versus the underdeveloped roast, done at something smaller than the “normal” size. Batch size, moisture content, water activity, starting green batch temperature must all be strictly monitored if a “standard roast profile” is expected to deliver a “normal” result. Changes in local weather conditions and barometric pressure can easily be enough to change convection performance of a small roaster, and drastically change final results, even if the profile is managed with great precision. The other common thread I heard was that “we were looking for a brighter taste profile”. There is a very big difference between a light roast and an underdeveloped one. The former is acceptable, the latter is not.

This is a problem all roasters struggle with. In all fairness to the roaster(s), an occasional problem is to be expected. For most small batch roasters, lack of adequate instrumentation and sophisticated profiling apparatus means roasting is not an exact science, so the problem of consistency is an extremely challenging one. However, since all coffees can be brewed, tasted and measured prior to shipment, end consumers have the right to expect reasonable consistency and quality. This may require a change to the way small batch specialty coffees are processed. In  order to QC a coffee, a roaster needs wait 18-24 hrs post roast, grind and allow the coffee to out gas for thirty to sixty minutes, then brew, taste and measure to see if it extracts normally. Pre-grinding and post roast time is necessary in order to allow CO2 to out gas prior to brew (as opposed to waiting 4-5 days post roast), especially in espresso. CO2 outgassing is very high immediately post roast. As much as 70-psi of CO2 can be contained within some of the densest [whole bean] coffees. Hot water [and pressure ] are catalysts that accelerate outgassing and can cause excessive foaming when brewed too early post roast. The ground coffee should be placed in a sealed bag with one-way valve so that it may out gas, but not absorb moisture from the air. Yes, this would delay fresh-roasted coffee deliveries by one day, perhaps that’s part of the problem – rushing the coffee to market before QC tests can be properly conducted.  Personally, I would much prefer a coffee one (1) day later, but properly developed than a 50/50 chance a day earlier.

Finally, another symptom I’ve noticed is that coffees underdeveloped at roast frequently fail to out gas in the sealed bag, even after 7-10 days. Most bags these days use a one-way valve. Assuming the coffee is bagged and sealed promptly post roast, i.e., w/in a few hours at most, the valve operates at a low pressure, but one that is higher than normal atmospheric (sea level) pressure. Without the valve, the seams of the bag can actually split.

As the coffee outgases, the bag should inflate noticeably before the valve operates. Some coffees are so dense, they can continue to out gas for 104 days in the whole bean state, but most of the outgasing occurs early. This CO2 helps to preserve the coffee, as it is partially inert, and displaces any oxygen in the air of bags not flushed with Nitrogen. Ground coffee will give up most of its CO2 in just 8-16 hours.

Coffee underdeveloped at roast fails to outgas at all; even after several days have passed, the bags are completely flat. Assuming the valve is working correctly, and the coffee was packed promptly after roast, and the sealing machine  is sound, there should be CO2 partially inflating the bag, until the valve operates.

More formal research needs to done to quantify these general observations. Note, this observation does NOT apply to coffees intentionally or otherwise allowed to out gas prior to packaging. What I’ve read to date about outgassing comes from the Fraunhofer Institute in Germany, and is extensive, relating outgassing, age and oxygen levels in sealed bags to freshness as conducted by a jury of cuppers.

I welcome your thoughts and comments.

10 Responses to “Underdevelopment”

  1. Victoria says:

    Where did you find the information from the Fraunhofer Institiute? I’d like to read more about out gassing and oxygen levels in regards to freshness.

    Thank you,


    • Vince says:


      The Publication of interest is one of these:

      Mrs. Dr. Radtke, Frauenhofer-Institut,
      Munich Food Technology 6, Page 36-42, 1979


      Deutsche Lebensmittel-Rundschau 77 (1981), No.6, S.203-210
      ISSN: 0012-0413


  2. Thanks for the great article Vince. I read it with great interest, and it certainly shows how the Mojo can be used as a diagnostic tool. I can’t help but think that under-development should be evident long, long before you are actually pulling shots, but another way to confirm the problem is always welcome. I have to say that the one missing piece of information here is exactly what I keep coming back to: What type of roaster resulted in these under-developed roasts, and what sort of roast times, curves, and finish temperatures are we talking about. Granted, I doubt you can or would say, but it would be interesting to know. We are talking about some massive incompatibilities between the use of the equipment and the thermal transfer of the particular roaster type. There is not doubt you can have issues with under-developed coffee bean interiors on most any machine, but small batch loads in traditional drum roasters, or fast roasts in hybrid or air machines that are hugely convective could give these results. Can you give any indications? Thanks again, a great read…

    • Vince says:

      Hi Tom:

      Was great to see you again in Oslo.

      Yes, you’d think underdevelopment would be evident before, but it’s not, at least not easily to the naked eye. In fact, roast profiles (for those lucky enough to have a profiling system) can be followed with great precision, and bean color can look right to the naked eye and even measure correctly (both as ground and w/b) using what ever method is your standard, but until you start brewing the coffee all seems to “appear” to be okay.

      The only exception I’ve noted that can be observed requires that you wait long enough and see that after 4-5 days there is virtually zero outgassing. Darker roasts are known to out gas more and sooner, but all roasts should out gas CO2 when fully developed (in my experience) – assuming packaging takes place promptly post roast.

      Some have said they can smell underdeveloped coffee at grind, even before it’s brewed-I can not.

      I’ve personally observed this issue with ALL types of roasting machines, from sample roasters to 35 kilo machines, including both those that are primarily convective as well as combination of convective, conductive and radiant types.

      A simple example of how easily this can happen is to think about processing two 35-kilo batches that are identical with one exception, one was dropped w/green at 25 Deg C, the second dropped at 18 Deg C. This happens all the time in warehousing where fall-winter temperatures drop. I am over-simplifying this a bit, but if you consider ONLY the ~11% moisture content of the green, you will need an additional 100,000 joules, or about 100 BTU to bring the batch to 1st crack. This figure is not accurate, because the beans give up moisture during roast, which decreases the effect of colder drop temps, but the entire mass of the cooler green must be included which raises the energy required considerably, perhaps as much as requiring several hundred thousand joules of additional energy.

      Looking at this another way, roasters can compensate for cooler green drop temps in a number of ways. Pre-heating the green to a constant temp (risks drying the beans) or storing more energy in the drum and dropping at higher temps.

      Changes in starting drop temp (charge) can be computed to match the additional energy required, but must be carefully managed because small changes in charge temp can cause huge changes in roast profiles when trying to compensate for cooler green storage temps. In a 22-kilo roaster, for example, with a 40-kg drum, a change in charge temp from 210 to 230 Deg C results in a change in stored energy of ~360,000 joules. So, ~3.7 million joules is stored vs ~3.3 million joules before drop (if my math is correct). I am using an estimate of 450 joules/kg-Deg C as specific heat capacity of the drum.

      Unless this change is paired properly with change in starting green temp and mass, controlling the profile can be extremely difficult at best.

      Regarding underdevelopment at roast, I’ve heard via direct eMail from a number of significant others out there about their recent experiences with underdevelopment being similar to mine or worse. I hope we here form them here on the blog to share their experiences.



  3. Chris Tacy says:


    Thank you so much for writing this up – and for bringing to light a growing and concerning issue.



  4. Tim Varney says:

    45 minutes certainly wasn’t enough for you in Oslo… great read. thanks again.

  5. Kees says:

    Very interesting doc. Thanks for renewed inspiration

  6. Drew says:

    Thanks for this, Vince.

  7. geir oglend says:

    Hi Vince, great article insightful and helpful with topic that isn’t covered all that well.
    Thanks Tom for digging. Having been fortunate to operate several different roasters over the years, this is an area that has frustrated me, thanks again for the eye-opener.