Actually I'm not thinking that either position is an absolute. Anything we can do at reasonable cost to make roasters more consistent and results more repeatable seems worthwhile. I am thinking, though, that QC may slip for the very high standards of boutique roasting when the process becomes routine and the person becomes less attentive. To take Jim Schulman's point, are we underestimating the amount of practice needed to develop sensory expertise?

1001 say they want more consistency, but it's BSing, like me saying I'd rather jump off a tower and flap my hands than hassling with a commercial flight.

Here's the proof: a fully automated Sonofresco roaster is cheaper than drum roasters, so are any number of highly programmable and highly accurate roasting devices, from commercial bread ovens to pottery kilns, that could be easily adapted to roasting 10 to 50 pound batches of coffee. Yet these are a niche market for dreamers opening a cafe "to get out of the rat race," in those first few months before they realize what they have signed up for.

An experienced craft roaster going this route would be admitting that specialty coffee roasters do bad roasts on good beans, while mass market roasters do good roasts on bad beans. That would be a short cut to bankruptcy.

How can anyone come to take this seriously? If you spend your roasting time poring over artisan profiles (what a deceitful name for this product!), with your head filled with pseudo-scientific roasting gibberish like RORs, etc, it's easy to get into a head space where nothing except a mass production engineering ethic seems real.

So here's a cure for that delusion:

A McDonald's french fryer is far more consistent with precut potatoes than a bistro line chef doing frites in duck fat with everyday market potatoes. Which would you prefer? If you boot the chefs, who can competently improvise with hundreds of different foodstuffs, with a set of ultra consistent robots and machines using ultra-consistent inputs, would it still be a bistro, or a fast food place with a pretentious menu?

Here in Chicago, get off any red, brown or blue line El stop on the north side, and beside the liquor and mobile stores, you'll find a few ethnic eateries, a hole in the wall theater, and a hole in the wall cafe. Take an Uber to Michigan avenue, and you'll find glossy restaurants, Broadway plays on tour, and Starbucks reserve. Guess where the life and energy are?

I almost agree, Jim.

But there are things I have learned from roast profiles that have improved my coffee far more in a few months, than in all the seeing, sniffing and listening.

Some beans just take a crooked road from Maillard to drop. Whether it is moisture content, density or terroir, it doesn't matter. What matters is straightening out the crooked road. It didn't make sense to me, but when I tried it my coffee took a major leap forward in "dynamics" and sweetness.

Without looking at the RoR it would have taken years of trial and error to figure this out. And I've tried it in reverse and can replicate these roast defects on demand.

I noticed a post on the coffee thread yesterday: Misty Valley Natural where a typically good coffee, from a typically good roaster tasted harsh no matter what was done to brew it. I would bet a sawbuck it was the roast and not the coffee. The person blamed himself stating he bought several bags over a few weeks, thus eliminating the roast as the reason and gave in to just not knowing how to brew this coffee. I bet a fresh sawbuck that the roaster just didn't know how to roast it.

Waiting for a certain smell at the end of a roast is a nice idea, but what if something happens way before that point that eliminates the certain smell. How do you find out what it was. Clear and correct roast data would help a lot.

And again, I'm not endorsing push button profile roasting. Quite the opposite. I want as much data and control as possible to be able to use my skill and experience to bring out the most a coffee has to offer. A person might be a great driver, but if he is blind, his skills are limited. I want to "see" as much as possible.

A foolish consistency is the hobgoblin of little minds, adored by little statesmen and philosophers and divines. With consistency a great soul has simply nothing to do. He may as well concern himself with his shadow on the wall. -- Ralph Waldo Emerson, Self-Reliance

That said, I've long been a proponent of the philosophy that consistency is the holy grail of espresso. After all, if your grinder produces a different particle distribution every time you run it, you can't properly dial in a coffee and produce the same cup shot after shot.

I'm still learning to roast, but it seems logical that being able to consistently repeat roasts, at least in a general way, is a good thing. Maybe I'm in for some disappointment, but in my limited experience so far I haven't had issues with dramatic differences when repeating roasts on my Quest M3, provided that the roaster is adequately pre-heated, at least one batch has been run through (I use seasoning greens for that), and I carefully repeat the charge temperature, MET settings, fan settings, and development time. I can get pretty-much the same color level on my Tonino and the same solubility from one roast to the next. Doesn't mean the repeated batch will be good, just very similar to the last one.

I use Artisan and I think it's a good tool, especially for learning what's going on in a roast. But my sense is that while more data is usually good to have, and the more we understand roasting science the better, roasting is more art than science. It's rather like cooking, where a temperature probe can be useful but doesn't tell the whole story or make a mediocre chef into a great one.

I have a feeling that development of roaster technology isn't going to move the needle much until we figure out how to measure temperature inside the beans while they roast. Perhaps tiny nanoprobes with temperature sensors and Bluetooth transmitters injected into a few dozen beans will be the future...

As an old steel mill guy from an old steel mill family, I love this discussion. It's very reminiscent of the conversation that prevailed for years about iron and steel making. At the beginning of the 20th century, the blowers, the men who knew how to get the best out of a blast furnace, were highly paid artisinals. Everyone deferred to them. They made judgments based on time, sound, sight, and years of experience. I guess it was only after Dr. Deming taught the Japanese to pay attention to data, which data included the quality of the input to the process, that automation became an essential part of making quality steel. But, there is still a need for the artist, for the guy who has the experience. Whoever that is now just has a lot more control over the process. Those guys are still highly paid. I'm guessing they have a lot fewer screw ups than they used to have.
Anyhow, as a guy who roasts about 14 ounces at a time in his basement, I have come to really respect the idea that comes to us from Information Systems I think. Garbage In/Garbage Out. With no real pressure on production, decent equipment and experience that I've gained on my own and that Ive been able to garner from Home-Barista, I have lots of confidence that I can produce delicious coffee as long as i buy good quality beans.
I bought some of that Misty Valley and I'm betting I will like my roasts, even if they aren't all exactly the same.

Almico wrote:I thought about offering on-demand roasted coffee to my retail bean customers. But roasting 1# at a time is just not efficient. I have my name on a January build Cormorant roaster and am still thinking seriously about putting it on the coffee bar. Of course it will be more for show and tell than productively roasting coffee for customers.

The problem with the Bellwether model is efficiency. What happens when two people want two different coffees at the same time? How about 3? It still takes ~10 minutes to roast coffee. In this mad rush of a world, how many people are going to wait around while their coffee roasts?

And if customers are using it in a market setting, is it a one profile fits all scenario? Or do customers get to pick their own profile?! Whoa!

Bellweather has a green coffee program that most likely will be tied in to accompanying profiles. Load the coffee, select the green coffee or scan a bar code, enter the weight, and press start. Is it going to be that simple, or is there a catch?

Almico wrote:That sounds like the same old tired excuse for not being able to produce consistent coffee. Call it arts and crafts and it's OK. (Love you, Jim!)

But most coffee drinkers want consistency. How else can you explain why the majority of coffee consumers still line up at Starbucks and Dunkin Donuts by the millions?

And I bet if you ask 1000 of the top small batch roasters if they would prefer more exacting control over their roasters, and be able to produce exactly the roast they intended all of the time, 1001 would say yes.

Starbucks and DD are popular for the milk and sugar that they add coffee too.

But in all seriousness I find this debate of ideas very interesting. It reminds me of the battle between Steinway & Sons Vs Yamaha pianos. I learned about this in school with this case study https://www.slideshare.net/SeshasaiMaha ... y-and-sons

TLDR: Each Steinway had its own unique sound (Like those duckfat fried potatoes...yummmm) but if you played on a Yamaha they all sounded exactly the same (McD french fries...yummmm ) In the end Yamaha proved that there is more than one way to compete. Maybe the same can be said here.

Who wins? Am I missing something? Have I got the absolute best out of this coffee? Should I roast it 512 times with 9 variants in a fulluy crossed experimental design, each taste tested blind, to find the optimal roast?

If you ask yourself these questions, you are already lost. Sure, you'll get better steel, higher QPR pianos, and more consistently good coffee. But will you enjoy yourself? I've been cooking a "prep exercise pasta" for about 20 years (peel and seed the tomatoes, chop the garlic, crack the peppers, while the water boils and the first few minutes of the pasta cooking, so you can saute the sauce and have it all finish inside 15 minutes). I'm proud to say I've never done it exactly the same way twice. Do I cook the world's best tomato and basil penne? Not even close. Do I have fun cooking? Yep.

Playing at John Henry with enough power tools to get a tie; is that really what either you or the customers of a high end cafe want?

Consistency for consistency's sake is indeed foolish. Starbucks coffee is consistently bad. That is certainly not the goal. Intentional inconsistency can't be the goal either.

The ability to repeat a successful process certainly is a worthwhile quest. And just because you can repeat a process, doesn't mean you have to, but it does demonstrate control of your craft.

Not only does it establish control, but it gives a basis for transferring that control. Having a roaster with a magic nose is great, but what happens when he/she's on vacation? The meat of the bell curve customer wants consistency. If they want a different coffee experience, they can buy a different coffee. But when they want XYZ blend and buy XYZ blend, they expect to get XYZ blend. Most people just want good coffee, not coffee art.

another_jim wrote:A McDonald's french fryer is far more consistent with precut potatoes than a bistro line chef doing frites in duck fat with everyday market potatoes. Which would you prefer? If you boot the chefs, who can competently improvise with hundreds of different foodstuffs, with a set of ultra consistent robots and machines using ultra-consistent inputs, would it still be a bistro, or a fast food place with a pretentious menu?

If those competent chefs always produced a winning dish, maybe. Unfortunately, the flaw in your argument is that there is far to much really bad coffee coming out of supposedly good artisanal roasting houses. Just like there is far more bad espresso than good being served, there is far more "less than stellar" coffee being roasted than consistently great coffee.

hankua wrote:Bellweather has a green coffee program that most likely will be tied in to accompanying profiles. Load the coffee, select the green coffee or scan a bar code, enter the weight, and press start. Is it going to be that simple, or is there a catch?

Simply following a BT roast profile will get you in the ball park on darker roasts, but I fear not with lighter roasts for the discerning palette. I don't see automation as the answer for consistently great coffee.

I find this discussion interesting. I believe There are a lot of variables that come into play. One questions I always ponder: is taste subjective? I know philosophers have pondered this question for awhile or in a more broad sense aesthetics in general. I don't think anyone has this figured out as much as people try and figure morality out. Is it inherent? This of course is beside the point.

What about personality? I am not type A and I find it stressful to track a ton of variables. This leads into what Jim was saying, the "art of roasting". Sometimes the less I technology I have in front of me the more I am concentrated on the roast and the more I enjoy it. Of course, this doesn't mean I am producing better roasts.

What about business? From a business perspective, is your increase in consistency even worth it. That extra 5% is it bringing in x more amount of customers? That would be an interesting thing to actually track although I am sure it would be very difficult because what are we considering consistency? Like you said McDonalds and Starbucks have consistent customers but their coffee sucks.

I think that it is a worthy pursuit. My mom is an artist and consistency in her product is VITAL if it comes to selling to someone like Pandora where they only accept almost perfect beads (my mom does Lampwork which is essentially melting glass in a torch and making beads etc.)

Another thing to think about is: Should we aim for consistency? You know, it is interesting, I actually PREFER to experience different tastes every single time I drink the same coffee. If I buy Panama Elida one year and it exhibits blueberry notes one year and grapefruit the next it doesn't bother me. But you see it all the time on this forum about people who are trying "go back to the days of that blueberry Harrar" those are a certain TYPE of people.

I find too many variables to exist. This isn't to say that we shouldn't narrow them down and take a hard look at it.