what i learned as an oompa loompa

Engineer (left) going Oompa Loompa (right)

Last fall, I heard the spousal call of duty, “Elaine, help me get these doors open.”

My husband, Todd, and his co-founder, Cam, had spent the last few years building chocolate machines and scouting the world for great cocoa beans in an effort to open a micro-batch chocolate factory in San Francisco’s Mission District. The construction crew finished converting the brick automotive garage to a food-safe bean-to-bar workshop and it was time to figure out the nitty gritty details like moving, merchandising, and designing the café & retail space. Todd needed all the help he could get and I was happy to pitch in during those critical months. Now that I’ve come up for air and returned to tech projects, I’ve had time to reflect on what I learned from brick-and-mortar operations:

Even the welds break

  1. For loops are a veritable miracle — At the chocolate factory, something breaks every single flippin’ day. Each morning I gave my evil eye to the roasters, melangers, temperers, wrapping machine, dishwasher, or anything with a screw, fuse, gear, glue, belt, or oil level and asked, “Okay, which one of you little buggers is going today?”

    In comparison, code brings tears to my eyes. If that for loop worked yesterday, then barring catastrophic hardware failures or someone checking in code they shouldn’t, it’ll likely work today. That type of, “if you don’t touch it, it’ll keep working” certainty seems divine. I’ve always loved the Web but I have renewed appreciation for redundancy, unit testing, and monitoring now.

    Lots of chocolate; no email.

  2. Want to release faster? Don’t cut features; cut email. Dandelion’s chocolate makers don’t check email throughout the day (the execs aren’t so lucky). If you can depend upon eight hours of focused time each day, then goals magically happen. With just a back of the napkin estimation, if the team says they’ll make 1000 bars, stop worrying — they’ve got it.

    There are still complications and distractions (e.g. tours, businesses with supply emergencies, daily check-in meetings, and see #1) but there are no late nights catch-up email marathons, no pleas for Agile-enabled predictability, no emergency requests to add two more weeks to the schedule, and no unsustainable sprints to a goal followed by “whoa, let’s never do that again” post-mortem analyses.

    Conservatively, I think that at least 30% of Meebo’s productivity was lost to folks staying current with voluminous Inboxes. Startups want to be hyper-communicative and transparent but those cc’s and long winding email threads add up. However, the first startup that balances getting stuff done despite an ever-expanding Inbox will have a formidable advantage.

    From left to right: future Pulitzer prize winner, Fulbright scholar, and Stanford grad

  3. Tech people, get over yourselves — Dandelion Chocolate has two Fulbright scholars, a Harvard law school grad, Ivy Leaguers, and even outside of formal academic pedigrees, is one of the most talented teams I’ve seen. Tech recruiters spend a lot of time bending over backwards looking for niche skillsets and as a result, we tend to think of ourselves as the center of the talent universe. However, if the folks at Dandelion Chocolate learned to code or design, they’d whoop most of our startup tooshes. Other industries are teeming with extraordinary, passionate people and it doesn’t take an army of recruiters and hundreds of LinkedIn emails to find them. The next time I build a team, I’ll look beyond industry borders.

    Brandon analyzing bean sorting efficiency

  4. Engineering principles provide arbitrage opportunities — When I tell tech people I’ve been pitching in at a chocolate factory, I spot the concerned eyebrows, “Elaine, what about intellectual stimulation?” or a VC will wander to another networking table presumably thinking, “Oy, small potatoes.” Fermentation processes, tempering crystallization structures, and modeling roasting & melanging profiles provide lots of mental fodder. Chocolate making isn’t a traditional startup but engineering principles like small achievable goals, metrics, and A/B testing are very applicable, opening new opportunities in existing markets. The tech world is a fantastic training ground for non-tech industries where there aren’t support networks of meetups, workspaces, and mentors.

    The #1 troublemaker at Dandelion Chocolate

  5. HCI is a digital thing — A little backstory… the women’s bathroom door at Dandelion Chocolate constantly breaks (see #1). From the mezzanine, I can hear the frustrated click-click-click when guests jigger the finicky door lock. At the beginning of the day, I paused when I heard the metal clicking to make sure the bathroom seeker didn’t need help. But by noon, the high-pitched metal-on-metal screeching made my hair stand on end, my stomach churn, and I was ready to do anything… anything… even standing vigilant outside the bathroom to personally escort patrons in and out of the loo… to prevent that irksome sound again.

    When I initially told my parents that I was interested in HCI (Human-Computer Interaction), they didn’t get it. And now I understand why. Technology is so virtual that it takes logging, data mining, insightful researchers, and persuasive experts to convince a team there’s a problem on the Internet worth fixing.

    However, when you’re working with tangibles, you are walking and breathing your own User Experience experiment. You see the half-drunk cups when you take out the trash. You can’t help but overhear under breath comments on the street. And if something is broken, you don’t have to do a cost-benefit analysis to figure out whether fixing the bathroom door is a priority. Someone’s going to fix the bathroom door or they’ll go crazy.

    Chocolate for sale

  6. Simple business models solve so many problems — With Meebo, our business model went something like this: we built a product that people like. We monetized a fraction of those eyeballs with brand advertising. We tracked the percentage of users who clicked an ad and logged their engagement times. However, since correlating brand advertising with bottom line revenue is nearly impossible online, we also monitored ad partnership renewals. If all of those metrics were healthy, we were happy. (~60 words)

    At Dandelion Chocolate, the business model is: we make and sell chocolate. (5 words)

    It’s so much easier to build a sustainable organization around a simple revenue model. There are no tensions between ad partners, distribution sites, engineering, and sales teams. There are fewer points of failure. Instead, everyone is aligned towards a simple goal: make something people want.

    The Dandelion Chocolate team goes to Hawaii after hitting a big goal

  7. Equity isn’t a great long-term motivator — In tech, salary is only one component of your compensation package. If your company does well, your stock options can be worth far more than all of your combined paychecks. Equity is “skin in the game.”

    However, the compensation at Dandelion Chocolate is traditional — wages and salaries. And when things get tough (e.g. the temperer breaks during the December rush — see #1), employees’ visions of becoming overnight millionaires aren’t shattered. Instead, it’s an even-keeled, “Okay, what do we need to do to get through this?” What motivates the team is working with interesting people, more opportunities to travel and grow, and building something pride-worthy.

    As a startup leader, you fear an employee exodus at the first sign of trouble. For engineers with lots of options, trouble means it’s time to diversify your equity portfolio and seek greener pastures elsewhere. In some cases, equity can have the opposite effect and even encourage short-term thinking.

    Valencia St crowded with café experts
    (Photo courtesy of Flickr:tofuart via Creative Commons)

  8. The real world has tougher critics — Startups can shield themselves behind a slow beta roll-out. Your mom and dad might follow your company blog but they probably didn’t critique last Friday’s release.

    In contrast, almost everyone in the world is a café expert. On his day off Todd gets calls and texts, “Todd, the bathroom door seems to slide funny…” or “I drove all the way from Sausalito and you just ran out of marshmallows!”

    The feedback is always, always, always appreciated. But your skin thickens a bit. The chocolate factory is a public venue of our closest friends, friends of friends, neighbors, and supporters — the people you want to make most proud. Every one of those visitor has seen hundreds of cafés and everyone has opinions about how they’d make it better. Unlike tech, you’re very exposed. And if we have an off day, we can’t follow-up with an email newsletter or pop-up notification, “We listened to your feedback! Here’s version 1.2 just for you!”

    Customers in the face blind retail mode

  9. Retail face blindness (i.e. prosopagnosia) — Hopefully there’s a psychology student searching for their next phD thesis among this blog’s readers.

    Either I’m very bland looking (totally plausible) or there’s an undocumented psychologic phenomenon for how people mentally categorize service professionals…

    As a door greeter, cashier, or farmer’s market volunteer, I love talking with visitors. A quick exchange can expand into an intimate 10-minute conversation about our families, our personal food preferences, frequently an exchange of names, and sometimes even promises to follow-up via email.

    Afterwards, I’ll say goodbye and walk next door for a sandwich where I coincidentally spot them again. They look at me when I join them in line, look back, and order a croissant. Nothing. Not even a spark of recognition. I’m a total stranger! This has happened frequently enough that I’m now fascinated by the brain’s ability to immediately delete people information — especially people we don’t think we’ll see again. If you end up researching this further or know anything about this, let me know!

    The chocolate chip cookie — a reliable favorite for seventy-five years and likely to be popular for decades to come

  10. Tech is still harder — It’s easier than ever to start a tech company. The food industry is envious of tech’s lack of permits and regulations, workspaces, meetups, access to capital, and the built-in network of angels & mentors. If you have a good idea, the entrepreneurial community will bend over backwards to help you.

    However, it’s much harder to keep that tech company going. The Internet reinvents itself every two years making longterm planning difficult. It’s safe to assume that people will probably eat chocolate in ten, twenty years. It’s hard to guarantee that any burgeoning startup will be relevant in six months. Building a longterm Internet business within a mercurial market is extraordinarily stressful. In addition to the normal trials of starting a business, you have to be hyper-aware of trends and prepared to pivot tomorrow. It’s likely that the team you hired 18 months ago signed up for a different vision that what you’re working on now and that requires further management. Years 0-2 are easier but years 2-8 are harder.

I’m winding down my Oompa Loompa days with more appreciation for brick-and-mortar businesses. If the toilet paper’s out at a restaurant, I’ll try to change it myself. I tip more and curse the evil people who steal tip jars and iPhones (grrrr)! When the Giants make it to the World Series, I tune in to the police radio chatter and listen for riots. I can add “skilled in the Tiffany bow tying method” to my resume.

Above all, I am grateful to have been a part of the Dandelion Chocolate story and to have learned so much from the team. Thank you again!

And of course, feel free to drop by Dandelion Chocolate at 740 Valencia St (at 18th) in San Francisco.

Dandelion Chocolate Door

, ,

12 responses to “what i learned as an oompa loompa”

  1. Re point 9, you should read up on inattention blindness. Given the distractions of the retail environment, it’s not surprising to me that people fail to recognise faces of service workers.

  2. #4 ‘When I tell tech people I’ve been pitching in at a chocolate factory,… a VC will wander to another networking table presumably thinking, “Oy, small potatoes.”’ That’s happened to me. I told someone that I was a stay-at-home dad in addition to being an entrepreneur. He promptly “saw” someone he had “to talk to before they leave” and I didn’t hear from him again. That’s O.K with me because my three kids will do better then his 100+ companies.

  3. Great read :). I am also in the tech industry and I too would like to spend some time as an oompa loompa!

  4. As an HCI pro myself, I should remind you that memory is contextual, meaning that memories are formed and retrieved around contexts, not in isolation. Everyone has had the unpleasant experience of running into someone who exclaims about how good it is to see them again, when the face triggers no recollection at all. And memories form primarily around emotional events, which means almost all of our daily encounters will go into the mental bit bucket.

    And as a former factory electrician and electronics tech, I can confirm that working in high-tech is fundamentally different from working on actual stuff. There’s a wholly different feeling when firing up a formerly ailing machine and hoping that nothing else was getting ready to go, than there is from a software launch. More basic, more primitive, as if the machine were a child with a disease and you can’t quite be sure it’s entirely over with yet.

  5. I worked in tech and after being made redundant I decided to be a self-employed tinkerer and small scale manufacturer. I will never go back to tech. All of a sudden I feel like I actally have a life and I don’t have to sit through endless meetings with bristling egos just to get a decision made.

  6. Point 1 sounds familiar. I have very little practical-work experience, but I do my IT at a biology lab – and I am ever so thankful when my work just keeps … working. The poor lab techs have ever new and exciting equipment failures / bad batches / unstable output and such to deal with, while my four year old java hackjobs keep ticking quietly.

  7. If you’d like, I can ask a friend here at UC Berkeley who has worked with prosopagnosics about your question; there’s a bunch of us here in the Psychology department with interests in face perception. My gut instinct is that it isn’t prosopagnosia per se, but something closer to a selective failure to encode a face/identity match in a particular setting. True prosopagnosia isn’t that common, although lots of people have some degree of difficulty with name/face encoding, especially on a single meeting.

  8. Ben, I bet you are right about selective encoding. That sounds closer to what I’m observing that prosopagnosia. The most startling example was a 20 minute conversation at a Marin farmers market with a woman around my age about blood sugar levels, health, and exercise. We had a great conversation and she even sent me an email a few days later following up on our conversation with some links to her favorite research articles. However, not even five minutes after our conversation, I bought some salmon from the tent directly across from us and saw her again. There was absolutely no recognition when I saw her again. It was amazing.

  9. […] What I Learned as an Oompa Loompa (Elaine Wherry) — working in a chocolate factory, learning the differences and overlaps between a web startup and an more traditional physical goods business. It’s so much easier to build a sustainable organization around a simple revenue model. There are no tensions between ad partners, distribution sites, engineering, and sales teams. There are fewer points of failure. Instead, everyone is aligned towards a simple goal: make something people want. […]

  10. Glad to learn more about dandelion!

    No offense to the author (I know this was not the intent), but this seems to me an example of “software engineer learns tech firms aren’t the best at everything.”

    I work in SF, and spend 70% of my time on computer stuff — IT + coding.

    And I got to tell y’all. Yes, software engineering is really hard, and the intertubes are complex. And tech startups have made some key innovations in business. But the industry needs to realize that it can learn from other industries, and it’s not the best at everything it does. The startup world is breathtakingly insular, flitting from one trend to another, with little cross pollination. It also seems to suffer from leadership that is often quite immature (see brogrammers) and inexperienced (find me a brilliant 25 year old leader and I’ll show you a list of mistakes that person will regret when she’s 45).

    So I’m totally not surprised to see a software engineer find key software engineering innovations in other fields. It’s all the same. It’s all people. And if you think, for a moment, your sector has it nailed, you’re wrong.

  11. Hi Elaine,

    I feel the things you are writing about since we are doing the same thing as a couple as you guys in Europe. The mixture of these two worlds are absolutely a life changing experience because we talk about it a lot. The physical and digital world met in our life. Im always fascinated about the dematerialized digital world developing products, but building a physical product is something totally different. And makes me relax from my digital self:) Anyway great post. 🙂

  12. Brad — absolutely no offense taken, you’re right. The more perspective we have, the more we realize how little we know and how humbling that is. Travel is great for that exact reason.

    Thomas — I bet you are fascinated by companies Apple, Square, and FitBit who work with both worlds too. When I’m coding, I want something to hold and wrap up in a bow to give to friends on special occasions. When I’m working with chocolate, I’m cursed by Murphy’s Law. The grass is always greener and both types of work can be relaxing for different reasons.

Leave a Reply