With a little bit of work, I’m happy to unveil this neat new feature of RoastLog.

The main idea is that everyone services their roaster from time-to-time.  However, just like changing your oil, sometimes you forget when you did that, or you forget when you should be doing that.  With RoastLog, it’s now trivial to schedule an upcoming servicing and track the history of what you’ve done.

For any roaster which you’ve added to your account, simply open up the detail page for that roaster and you’ll see a “Log service” button.  Once you click on that, you’ll be able to add the details about a past or upcoming servicing which you or someone in your organization performed.

Adding a service record for a roaster

When you visit the detail page for a particular roaster, you can quickly see what’s coming up and what’s already been done.  For the servicings which are within the next two days, we highlight that row just to draw some extra attention to it.

Servicing record in the roaster detail page

Alerts are sent via email to everyone in your organization.  Alerts can be controlled to be sent out on the day the servicing should take place, or on some day before that.  If you prefer, you don’t even have to get the email and simply visit this page on your own schedule.

Thanks to our users who continue to give us great feedback on what they need.  We love hearing about what’s working and what’s needed so that we can continue to evolve our platform like this!

RoastLog fully supports pre-roast blending, but many, many people exclusively blend post-roast. With a clear understanding of Roastables (the ways coffees are used) and the blending feature in RoastLogger you can track your post roast blends.

Let’s be completely clear on terminology by defining “pre” and “post” roast blending.

A pre-roast blend is two or more green coffees roasted together in the same batch. We built RoastLog with this exact workflow in mind, so that inventory is automatically adjusted when you roast this type of blend.

Post-roast blending is two or more batches of roasted, single origin coffee mixed together once they’ve left the roaster. With RoastLog, this is not inherently supported; however, it’s still possible, and here’s how:

Using a specific example of a blend called Gaute-maul-an, which is a 50/50 post-roast blend of two Guatemalan coffees, a Huehuetenango and Rio Azul.

First, create three Roastables:

1. “Guat Huehue PR” so you can profile that SO coffee specifically for the Guate-maul-an blend
2. “Guat Rio Azul PR” so you can profile that SO coffee specifically for the Guate-maul-an blend
3. “Guate-maul-an PR Blend” for your Guate-maul-an post-roast blend, made up 50% “Guat Huehue PR” and 50% “Guat Rio Azul PR”

Now, for the roasting. Here’s how to do a 20lb batch:

The Roast Queue in RoastLogger for setting up a post-roast blend

1. add “Guate-maul-an PR Blend” to the queue, entering 20 for starting mass. You’ll use this roastable to tell the inventory system:
   • 10 lbs of the “Guat Huehue PR” were roasted and used in your Guate-maul-an PR Blend
   • 10 lbs of “Guat Rio Azul PR” were roasted and used used your in Guate-maul-an PR Blend
2. add “Guat Huehue PR” to the queue, entering 0 for starting mass
3. add “Guat Rio Azul PR” to the queue, entering 0 for starting mass
4. Record a couple seconds and save the Guate-maul-an PR Blend from step #1. The point here is just to trigger the inventory system.
5. Now actually roast and log the two other batches.

So, in the end, inventory is reduced correctly and usage data is recorded. You’ll also have complete profiles for the individual components stored in a way that lets you easily find them later.

Next time, using this same technique, I’ll explain how to handle a batch of coffee that will be split up post-roast for different uses.

Not only did I not want to light myself, family or house on fire, but I also didn’t want to burn or overcook my turkey.  Enter RoastLog.

My TurkeyLog setup. Four gallons of 350F peanut oil in the pot on the left, MacBook Pro and RoastLog Data Bridge on the high tech laptop stand on the right.

RoastLog started off its life as a temperature recording system with a little web software to store and interact with that data.  Since then, we’ve grown to include a much richer set of features for managing inventory, analyzing green coffee purchase prices and fees, etc.  Still, we retain the ability to measure, record and visualize temperature readings in real-time of pretty much anything you can reach with a thermocouple.

I grabbed a type-J thermocouple and drilled the correct size hole in lid of my fryer to measure the oil temp, securing the TC with a small metal binder clip.  Since the Data Bridge can measure up to four temperatures, I put a type-k thermocouple at the base of the pot in between the pot and burner.  I didn’t know what to expect measuring the burner temp, but thought it’d be fun.  With that, I lit the burner and was off.

Oil temperature during the warmup

After about a minute I got a very nice linear warm-up for the oil temperature.  As you can see in the screenshot above, the burner TC was reading about 850F.  One of the neat features of the updated UI is the ability to selectively turn on/off certain TCs on the plot.  It was nice to see the burner temperature, but I didn’t always want it on the plot for scaling reasons.

Oil (red) and burner temperature (green) during the warmup

The warmup phase went well and was thankfully uneventful.  Once I got the oil to 350 it was show time.  I had my old welding gloves from college and wore those while very slowly lowering the bird into the oil.  Putting that sucker in four gallons of 350 degree oil was a bit scary at first.  The welding gloves were a late addition since I didn’t know exactly where I had put them, but I was damn happy that I had found ‘em.

Gently placing the turkey in the oil. Note the thermocouple sticking through the lid, attached to the Data Bridge.

Once the lid was back on the pot and thermocouple in the oil I could see that oil temperature was dropping.  Just like coffee roasting, this was the result of putting something cold into something hot.  I cranked up the heat of the burner to get back to 350F as quickly as possible.  During that phase of my fry the burner temperature got up to about 1100F.  It took roughly 2 1/2 minutes to hit the turn-around time where the oil starting increasing in temperature…a bit longer than a typical coffee roast.

Now, here’s the part where RoastLog really helped.  Because I had a visual representation of the temperature, I could see when I was getting close to my target of 350F and the rate at which the oil temp was increasing.  As I approached 350F, I backed the heat down bit by bit and tried to just touch and then stabilize at 350.  Below is the zoomed-in plot from roastlog.com.  You can see it for yourself here as well…to zoom in simply click and drag on the chart: http://roastlog.com/roasts/zambrano-coffee/46/

Oil temperature near 350F

As you can see, I didn’t stray too far from my 350F target.  I was continuing just fine through minute 24 when I noticed a slight ramp up in temperature.  In terms of degrees, it wasn’t that much above 350, but due to RoastLog I was able to see the trajectory and that it would quickly become an issue if I let it continue.  Back down came the burners so low that I actually killed the flame….doh!  If you look at the subplot on the website or burner temp (green) below you can actually see a huge drop in burner temperature when the flame was out.  It wasn’t that big of an issue really….once the burner was re-lit  I got back on track.

Full recording of oil and burner temperatures. There's a big drop in burner temp when I accidentally killed the flame.

Being able to remove the oil temperature from the plot while recording was great, since the scales were so much different.

Oil temperature

The result?  Delicious.  I didn’t light anyone on fire, had a great time and produced a great tasting turkey for the family.  I’m looking forward to doing this again.  With a longer thermocouple I think I can actually log the internal temperature of the turkey which would be the ultimate in turkey perfection.

With our 4-input Data Bridge, there are some tough UI issues which we need to tackle.  What’s the “best” way to display 1-4 roasting curves in RoastLogger?  In the first iteration, I came up something which I thought made sense: create one main plot which can be though of as the “most important” along with three sub-plots.

Our current 4-input RoastLogger UI

The feedback Ryan received from folks was that they wanted to have all of these plots on one chart.  Originally, I thought this was totally insane.  With up to four curves for the current roast, and up to four for a previous roast drawn in the background, we now have up to eight curves drawn on one chart.  Insanity!  Indeed, making sense of eight curves on one plot is hard to impossible.  Ryan came up with an initial layout which had some controls over which charts to show or hide.  But now, just because a curve is toggled off for a given thermocouple doesn’t mean people aren’t interested in the temperature.  Hrm, what to do?

Since we’re reworking the UI with the charts, it’s a good time to look at other aspects of the UI beyond the chart. When adding the three subplots, we took away some vertical resolution which meant that it’s a bit hard to see subtle changes in temperature.   Realizing we needed to give back some vertical resolution, it was pretty obvious that the notes panel at the very bottom of the window was taking up more that it’s fair share as well.

So, our goals are:

• maximize vertical resolution
• allow fine controls over charts
• provide clear readings for time and all four temperatures

With all of this and some prototyping, we settled on the following changes:

• all curves are drawn on one plot
• provide off/on toggles for curve 2, 3 and 4
• provide off/on toggles for all curves in previous roast profiles
• move all Heads-up Display elements into the chart (time, temp and buttons)
• move notes into a vertical panel

Wireframe of 4-input RoastLogger UI version 2

This really opens up the vertical space proving more room for the chart to draw curves.  Inside the chart, the four TC temperatures will be displayed in real-time while roasting (that little table-looking thing).  Checkboxes next to those thermocouple readings toggle the lines for the current or previous roasting profiles.  For users who don’t need all four TCs, the inputs which do not find a TC plugged in to the Data Bridge will be automatically hidden.  The Notes panel which doesn’t need a ton of real estate get’s kicked over to the right and is oriented vertically along with the events panel.  However, the Notes panel can now be hidden so if it’s not needed, simply close it and keep it out of sight.  The Roast Queue and Starred Roasts profiles panel remains the same, except that it can now be much wider since the HUD has been moved.

For those who really want to maximize the chart, it’s fairly simple.  Each of these panels is moveable and close-able.  Below is a wireframe of what it would look like when orienting the toolbar vertically and completely closing the Notes and Events panels.

Maximizing the chart in RoastLogger

These changes just need a bit more coding and testing, after which we’ll be releasing RoastLogger 2.1.2.  I think this is a much better design and hope that it’s easier to work with.

4-input Data Bridge

Jeez…we’ve been talking about this for so long even I’m getting a bit sick of it.  However, I’m extremely happy that this little guy is now available!  We rushed to get the hardware ready in time for SCAA in Houston, but still had a ton of integration work to do with our software.  There were a few database changes, UI tweaks and odds and ends to clean up so that we could easily support one or four charts.  For the last few weeks our testers have been logging roasts daily with these devices…yee-haw!  Of course, we identified a few things to fix and have been doing so diligently.

When you sign-up with RoastLog you’ll have the option of logging with our four-input device or our original one-input.  Once you log into the RoastLogger client software, it will automatically detect which device you have and the UI will adjust itself out appropriately.

There is still some work do to in order to use a subset of those four charts, but it’s coming.  I’d like to give users the option of using one, two, three or all four plots and have the layout be more flexible.  Some folks like side-by-side to get more vertical resolution while others prefer the layout below.  Drag-and-drop to reorganize?  With this update I completely switched out the charting library in RoastLogger, so we have a lot more options which is very very nice.

Software updates

One of the things I’ve been working on most recently is making our software better overall….fixing bugs, designing better in-application messaging systems, working on new-user experience, etc.  It’s been a lot of fun actually.  As a software guy, it’s always fun to tackle new problems and build new systems.  However, there are always these little annoyances (i.e., bugs) which eat away at me, and are sources of frustration for users.

The latest version of RoastLogger has a few of those fixes as well as some enhancements which make it easier for users to know when there is a problem.  At the same time, the website has undergone a few enhancements and fixes.  If you’re a customer, you may have noticed a little message recently announcing RoastLogger 2.1.1.  This system will allow us to better notify folks about system updates.

Support portal

One thing we all obsess over is providing excellent customer support.  The best part of running your own business is talking to your customers and deciding that you’re going to treat them like you want to be treated.  Two weeks ago we launched our new customer support portal which will help us do a better job of helping you:

http://support.roastlog.com/

Our goal is to get back to everyone within 24 hours with an answer.  Most often, we’ll get back within a couple of hours.  Better yet is that this system (powered by the good folks as Assistly) provides a great search system through common questions and issues.  Our hope is that this system will take the hard work out of finding answers to common questions.

Planning for 2012

Since joining us shortly after the SCAA show this year, Linsey has been busting his butt getting our sales and marketing efforts organized.  We’re also all signed up for the SCAA show in 2012, so plan on saying hi to us in Portland (Portland…thank you SCAA).  We still have a big vision for RoastLog and are doing our best to build software which every coffee roaster will want to have.  Hopefully 2012 will see us get even closer to that vision!

Green coffee inventory presents unique challenges and we have build a system to address these issues. Take for example the fact that most roasters have limited space to store the green coffee that they have purchased. You might have bought a container of coffee that will satisfy your roasting needs for the next six months but, you can only store a couple of pallets of that coffee in your roasting facility at one time. No problem, you can keep your coffee at a green warehouse like the Annex and pull bags and pallets as needed. But how does this impact your inventory? After you add in the price per pound charge for storage and shipping, those bags you pulled last month were cheaper than what you are pulling now. The same coffee stored at different locations is ultimately going to have an impact on your cost of goods sold. Chances are the system you use now, whether it’s QuickBooks, Peachtree or your own tricked out Google Docs spreadsheet, is going to require a lot of manual data entry to get the numbers to justify and create accurate reports. The RoastLog inventory system takes the pain out of tracking this kind of information; we did this by listening to coffee roasters, understanding their needs and building a system that works. The system accounts for multiple storage locations, multiple roasting locations, movement of coffee and pricing variability.

Words on the screen are probably too abstract to fully explain what the RoatLog inventory system is all about. Because of this we are planning to offer a series of webinars that will explain how the system works. More news coming soon.

Linsey is Back

If you visited the RoastLog booth at the SCAA show, you may have had the pleasure of meeting Linsey Fan. As I mentioned in a previous blog, Linsey was involved with the prototype version of RoastLog a few years back but, now he has agreed to rejoin the team in a more formal capacity. Linsey’s background in sales and marketing is going to be a big help for our small company. Already Linsey is taking charge of our sales program, following up with roasters who expressed interest in the RoastLog system. Don’t be surprised if you get a call for Linsey in the coming weeks!

Last year in Anaheim we arrived at the conference with zero customers and basically said, “RoastLog is open for business!”  People loved our vision, but there was still a tone we felt which like…”Who are you guys and what happens to my data if you close up shop next year?”  That being the case, we walked away from the 2010 show with our first dozen customers.

This year in Houston, we showed up saying “Roasters on four different continents have been using our system day after day for over a year.  Here is our new piece of hardware and a new inventory system built to solve your issues.”  The tone which I picked up on this year was, “OMG…this will save me so much time.  Your user interface is sweet….I need this.”  We did not receive a single question about us theoretically closing shop.

Basically it just felt like people are much more willing to trust us.  It makes sense…we’ve been out there for a solid year and talking to folks for six months figuring out what the inventory module was going to look like.  When you do that it must seem pretty clear that we are here for the long-haul and dedicated to pushing our little start-up forward, building tools for the coffee industry.

Personally, this was the first year (out of the four shows I’ve been to) where I really felt like I got to catch up with old friends who I rarely get to see.  Even though Richard from 5 Senses is a RoastLog user, I only see him once a year at the show.  I’ve chatted with Mark from Synesso for the past three years so now we recognize each other…it’s a blast saying hi and catching up as one of us passes the other’s booth.  Same goes for Lincoln from Alterra Coffee.  Lincoln was next door to us in 2009 when a pre-alpha version of RoastLog arrived in Atlanta just to test the level of interest from the community.  I’ve been promised a place to crash in Sydney from the very first guy who wanted to install RoastLog before it was even ready (Paul from Mecca Espresso), so now I just need to get going on a business trip to Oz!

Pre-alpha version of RoastLog circa 2009

But, one of the best things about these shows is finally getting to meet our users in person.  We exchange emails, talk on the phone, Skype…countless things to help folks get up and running or listen to their feedback.  Still, seeing someone face-to-face and getting a change to look them in the eye and listen as their talking is priceless.  I think I speak for all of when I say that this was probably the best part of the whole show….I only wish we got a chance to meet all of our users.

RoastLog is a project that has been in the making for over three years. Brian came up with the concept at the SCAA show in Minneapolis in 2008; followed by Brian and Linsey’s exploratory exhibit of a RoastLog prototype in Atlanta the following year. Later in 2009 Ryan and I joined Brian to develop the prototype into a real product that officially launched at the Anaheim show last year. Since then we have continued our focus on making software tools that roasters and other coffee professional need.

Over the last year we have gone from a handful of beta testers, to a dozen early adopters, to over 60 coffee companies logging roasts in North America, Australia, Asia and Europe. At first customers were taking a chance with us, an unknown new company that had a cool product, but through trust and proving ourselves on the customer service front, we continue to bring on new users month after month. We take great pride in this degree of customer acceptance because it validates our hard work with customer support.

Different customers have different needs and it is our job to meet whatever those needs are. We have found that many roasters need help determining something about their thermocouple set up, whether it is drilling a hole and tapping threads or replacing an existing thermocouple with a new one. If needed, we provide our customers with hardware and instruction on how to install it. If our customers have questions about how the system works we promptly respond by email and often set up Skype meetings with new and existing customers in time zones throughout the world.

We listen to our customers and do something with the feedback we receive. In the past year we have made changes, some small and others significant, that improve the system for all. We do not make changes for the sake of making them, we collect feedback from all of our users; we know what to do when everyone is asking for the same things.

As any regular reader of this blog knows, one of the things we have been working on is a multi-input data bridge. After a lengthy engineering and beta testing process, we had the data bridges manufactured and assembled and are now finally able to introduce them at the show. We are also in the final phase of development of the inventory management system. Stop by our booth (#445) to say hello, see the progress we have made, ask some questions and sign up for RoastLog.

With all of the work that we’ve been doing on our inventory system, I thought it would be interesting to describe how we approach product development.

The 50,000 ft. overview of Customer Development is changing the angle of approach when starting a business or developing a product.  Traditionally, someone has an idea, a company builds the idea after which a sales team tries to sell it.  Customer Development takes the approach of finding your customers first, then building what they need.  We like that.

Here are a few key things we’ve done with RoastLog which I think explains much of the positive feedback we’re getting from folks:

1. Exploration – I had an idea of what would become RoastLog and took a very rough prototype to the SCAA show in Atlanta.  A friend and I shared some booth space with Jay of CoffeeTec and simply asked people what they thought of the idea.  We got out of the building and asked real coffee roasters if a tool like this would be useful.  The response was overwhelmingly positive…Paul from Mecca Espresso in Sydney wanted to buy it on the spot.
2. Building fast - A few months after the show the first prototype went out into the wild with a handful of testers.  By this time I had roped in Ted and we had made some initial contacts in the industry.  Ryan also jumps onboard to help out with front-end/UI/UX work.  RoastLog was ugly, kludgy and awkward…but it was functional.  The idea here was to figure out what we had gotten right, and what was wrong.
3. Iterating – We completely revamp the UI to make it easier to use.  At SCAA 2010 we went from 0 paying customers to a dozen.
4. Listening – Last year at SCAA, even though we had just launched or flagship product, we were already spotting issues and hearing patterns when talking with people.  People wanted to track where their green bean inventory as well as measure more than one temperature at a time.

Now, switch gears to the next evolution of RoastLog….how did we develop our inventory system?

1. Exploration – Try to get out of the building some more and wrap our heads around what coffee businesses need in an inventory system.
2. Building fast - After getting an idea of what a professional inventory system build for roaster would look like, we built it and pushed it out as beta software to a small group of testers just like we did last year.
3. Iterating – We launched inventory earlier than we all felt comfortable with, but a system needs to be tested in the real world.  The feedback we’ve gotten has allowed us to quickly correct mistakes and incorrect assumptions.  We’re on track now thanks to feedback from our beta testers and long nights and weekend coding away…we’ve been doing big updates about once a week, iterating quickly.

Listening isn’t listed in the second set yet….we’ll be at the show and we want to hear from you.  Come by our booth (#445) to get a demo of the system and tell us what you think.  We’ll listen.  Every piece of feedback we get will not make it into the product, but the ones we hear over and over again will.

There will always be bugs in software, we’ll never get it right the first (or tenth) time, but we’re always going to developing systems from your perspectives.  We want to build software and tools which solve your problems and make your lives easier as coffee roasters.

I had bit of a reality check the other day when a RoastLog user mentioned to me that he had been talking to a roaster at another company who thought RoastLog was an automated roast profiling system. This got me thinking that we need to do a better job communicating what RoastLog is and what it isn’t. This blog entry is an attempt to do just that.

When we first looked at what kind of technology solutions were available to coffee roasters a few years ago, we saw a massive gap between what large coffee companies could afford and what was available to the small and medium size roasters, the ones we feel make up the heart and soul of specialty coffee.  Even when smaller roasting companies could afford the software and hardware that was commercially available; those systems usually did not satisfy the needs of roasters who are focused on improving quality. If you want a computer to control your roaster for you, no problem, pay $10,000+ and take your pick from a variety of products available on the market. RoastLog offers something different.

RoastLog is focused on providing tools to roasters who want a deeper understanding of the roasting process in order to improve quality and consistency. Automated systems might help with the consistency part of the equation, but this comes at the expense of quality. RoastLog does not take averages in order to determine average roast profiles. Instead RoastLog makes it easy to track information and then make that information easily accessible so roasters can come up with accurate quality-minded conclusions.

Based on experience, roasters can feel their way through the roasting process but, testing intuitive theories takes time, time that is not always afforded to the busy production roaster. RoastLog makes it easier for roasters to turn theories into knowledge simply by using the system. Load coffee into the roaster, track the profile, mark events like changes in application of heat and then compare profiles. We’re not only talking about lines on a graph, but more importantly the impact of changes in profile on the aroma, taste and tactile sensations of the cup. By relating cupping notes to specific roasts you can progressively improve the roast profile based on empirical evidence and not merely a gut feeling. RoastLog does not want to take the art and craft out of the roasting process. We just want to make it easier for roasters to make coffee taste better.

In the meantime, when we’re not making improvements to the system (readers of this blog know about the imminent arrival of the RoastLog inventory system and the progress we’ve been making on the multi-input data bridge) we’ll do our best to reach out to the coffee community to talk about what RoastLog is, how it work and what it can do for you. We know that in reality most roasters have never heard of Roastlog and if they have, they may not know exactly what the system can do. Don’t be surprised if you get a call or find an email in your inbox from one of us sometime soon.