Posted on

The Refried Beans of Agile

As a young computer programmer, I moved to Dallas, Texas, without ever having been there before, and with no idea what the place was like. There were three attractors for me: (1) Reputedly good prospects for a career in information technology, (2) no State income tax, and (3) good Mexican food. Those of you know know me IRL will be unsurprised by which of those was the strongest attractor for me. Indeed, as I rolled into town the first time, I was delighted to see a Mexican restaurant on nearly every corner.

Some years later, I was working with a guy who had just completed an introductory Six Sigma class and had started the journey toward Green Belt status. In other words, an Expert. As he listened to the almost-daily discussion about which Mexican restaurant we should go to, he had a brilliant idea. He would assess Mexican restaurants by the quality of their refried beans.

The idea was inspired by a thick, heavy, expensive academic book by two PhD holders in management science. As an example of how to apply DMAIC, the authors suggested that we could judge the quality of a steak restaurant by measuring the quantity of green peas customers left on their plates after they had dined. Never mind the steak.

This is something academic types do a lot, and I’ve expressed my humble opinion about it before. They were trying to find a common denominator to compare across various observations. A worthy intention, even if it entirely missed the point. People don’t choose a steak restaurant because of the peas, or a Mexican restaurant because of the refried beans.

Now, you might say that’s so obvious I needn’t have typed it. I know you know that. It wasn’t for you. Occasionally an academic reads this blog, and they wouldn’t intuitively know that steak is more important than peas at a steak restaurant, and enchiladas are more important than beans at a Mexican restaurant. They’re good at counting things, but not so good at comprehending the meaning or value of what they’re counting. They count for the same reason as Count von Count counts. It’s what they do, that’s all.


So, the guy explained his rationale. The one thing all Mexican restaurants have in common is refried beans. Therefore, logically, the best way to compare Mexican restaurants is by comparing the quality of their refried beans.

The rest of us laughed. When we realized he wasn’t joking, we laughed harder.

We explained that when we’re in the mood for enchiladas, it’s Uncle Julio’s. When we’re in the mood for carnitas, it’s Cantina Laredo. When we’re in the mood for tortilla soup or stuffed jalape├▒os, it’s On the Border. When we’re in the mood for good home-style cooking, it’s Escondido.

For genuine street food, there’s this Venetian blinds store in a certain part of town that opens a makeshift restaurant at lunch time, catering to Mexican workers. They don’t even have a real kitchen, as such. (There was no Yelp in those days, but if there were this place wouldn’t have been listed. It wasn’t formally or legally a restaurant at all. Ah, but the food you could get for $3!)

When we just want to stuff ourselves and quality isn’t an issue, it’s Pancho’s – ┬íPanza llena, corazon contento! And so on, through the list of amazing Mexican restaurants in and around Dallas, each with its particular specialties. It’s true they all serve refried beans, and they all taste okay, but the beans aren’t the deciding factor. Ever. Half the time we don’t even eat them.

But our budding Six Sigma expert didn’t get it. Perhaps he had no taste buds, after all. Come to think of it, he had a number of traits suggesting his awareness of and connections with external reality were somewhat tenuous.

A lot of time and beans have passed since then – ha pasado mucho tiempo y frijoles. Why bring it up now? I’ve been mulling over my experiences with large-scale “agile” transformation programs at gigantic corporations. I’m on my seventh now.

Compared with numerous smaller engagements, these large ones have a distinct charactistic in common: They can’t tell the difference between the steak and the peas, the enchiladas and the beans. The thing is, the steak and the enchiladas are variable. Their quality and value depend on context. When you really want steak, even the best enchiladas in town won’t satisfy. It isn’t because there’s anything wrong with the enchiladas. That’s hard to understand. But the peas and the beans are always the same, everywhere. That’s easy to understand.

So, once the transformation program progresses beyond the initial stages, when it’s directly guided by the executives who understand the business value they’re aiming for, and it shifts into the hands of mid-level management for day-to-day “execution,” all the focus is on the beans. That’s all they can understand. Agile ceases to be about value, quality, and sustainability, and becomes all about tools, frameworks, and rules.

A number of companies have capitalized on this, and provide management tools to support it. The main players are VersionOne, CA (Agile Central), and Atlassian (Jira). As implemented and used in mega-corporations, none of these tools is in any sense “agile.” All of them have been tailored to enable traditionally-minded managers to continue working in the ways that are familiar to them, while decorating their work with “agile” buzzwords.

The tools fit nicely with the various “agile scaling frameworks” on the market. The way most of them “scale” is by re-introducing 20th-century management structures, assumptions, and procedures to coordinate multiple “agile” teams. The company brings in an army of agile coaches and trainers, and they stir the waters for a while.

Ultimately it comes back around to where it started: A linear SDLC process based on a Work Breakdown Structure and managed with some flavor of Earned Value Management, even if labeled differently. Even if the organization runs “sprints,” they’re just tick-marks on a calendar. It’s all beans and no enchiladas.

The reason a company might consider adopting “agile” methods is to improve business performance. But before they can ever achieve that improvement, their transformation program devolves into an exercise in data entry. And that’s where it dies. It’s no hero’s death, either. The patient is artificially sustained long after the brain has ceased to respond to stimuli.

Jira and similar tools, and SAFe and similar scaling frameworks, are the refried beans of “agile.” They aren’t the reason to choose this restaurant, but they’re always served, and they’re what mid-level management can understand.

Personally, I prefer steak or enchiladas. Or both. Or Thai, maybe. Sushi sounds good, too. Or maybe a good burger. Excuse me, I’m feeling a bit peckish.