Hacker News new | past | comments | ask | show | jobs | submit login
Lean Meets Wicked Problems (steveblank.com)
52 points by sblank on July 28, 2023 | hide | past | favorite | 39 comments



> The team working on the Mapuche conflict in the Araucania region of Chile, flew to Chile from London, interviewed multiple stakeholders and were back in time for next week’s class.

Jesus, for a class? Who are these people?

I question the lesson, here. A wicked problem isn't one you can solve by flying into a country for a week, developing some incredibly superficial understanding, and then start proposing solutions. Step one, here, should be approaching these issues with deep humility, and this would seem to be the opposite of that.


This is pointless armchair quarterbacking, and it's counterproductive to spread this sort of thinking.

If you read the presentation from the Mapuche conflict team, you might come away appreciating the effort and recommendations. You'll note plenty of humility and humbling in the takeaway notes from the project. And, you'll also note there were Mapuche labeled as 'willing to talk' right in the deck.

Especially here at HackerNews, I propose we generally stay positive on the idea of leaning in to solving difficult problems, and getting working on them rapidly. I believe it's better for the world.


Getting to work rapidly is guaranteed to be the wrong thing to do with any difficult problem.

The book "How Big Things Get Done" elaborates on this and gives examples of how to do it right.

1. https://www.amazon.com/How-Big-Things-Get-Done/dp/0593239512


> Jesus, for a class? Who are these people?

The people taking the class. Which they flew out for.

> A wicked problem isn't one you can solve by flying into a country for a week

A wicked problem is one whose solution can't be seen until you've started working on it.

> flying into a country for a week

Meeting the stakeholders.

> then start proposing solutions

That's sprint one.

The alternative to meeting the stakeholders before working the problem, is working the problem before meeting the stakeholders.


> Who are these people?

Future management consultants.


My exact thought.


The entire premise is absurd.

From the title, I imagined it would be a criticism of lean. A canvas describing your wicked problem must be an exercise of practical comedy, there's no other explanation.


It does seem more like satire.


This is hilarious, because it misses one of the core issues with wicked problems. Wicked problems are wicked because they are nearly impossible to theorize about.

You cannot know if a solution is viable unless you actually implement the solution. So you won't know if you're right or wrong until it is too late.

You can collect all the data, interview all the people, see all the things, and still do the wrong thing. And you won't know it until you've tried. Sometimes you won't know if you've failed or not for years afterwards.

There are no good ways to approach wicked problems in general, because each has its own complications.

So while it may feel good to feel like you've done something here, I question whether anything was actually done. Because there is no way to know if any of these proposed solutions are actually viable or effective.


Did you read the article?

It didn't miss any of that. That was the point of the article.

>> Cristobal and I wondered if we could combine the tenets of Lean (get out of the building, build MVPs, run experiments, move with speed and urgency) with the expanded toolset developed by researchers who work on Wicked problems and Systems’ Thinking.

>> Our goal was to see if we could get students to stop admiring problems and work rapidly on solving them.


Yes, but none of the solutions will be implemented. It's ultimately navel gazing.

And one of the qualities of a wicked problem is that you can't really run experiments. Finding out if (for example) a new method of teaching during early education will result in more capable college students has no experiment that you can run that doesn't involve just doing the thing. And while you can implement with "speed and urgency", the fact that this is an issue that affects people after many years means after you've implemented it, there is no more speed or urgency to be had. And you also don't get to know if you've done anything until the end. And also with all of the other confounding factors that go into the growth and development of human beings, you can't even know if what you did was the thing that made the difference.

They're hard. Full stop. We need to approach them with caution and full realization that we are essentially replacing our wings mid-flight. You don't get to be wrong without things going very wrong. So whatever you do, you have to try and do it in a way that mitigates as much possible harm as you can perceive.

And even if they did come up with a solution to one of their problems, so what. Because another quality of wicked problems is that they're fairly unique. What works for one problem is not guaranteed to work for any other problem. So even if these techniques solved even one of these problems, it does not help.

We don't need tools to solve wicked-class problems, that's about as likely as perpetual motion, we need tools to identify them.


> You cannot know if a solution is viable unless you actually implement the solution. So you won't know if you're right or wrong until it is too late.

It is certainly possible to sometimes know when a given idea cannot work.

> There are no good ways to approach wicked problems in general, because each has its own complications.

You are guessing, necessarily. That people are generally speaking not able to stop guessing, and realizing that they are guessing, is a lot bigger of a headwind imho.


The whole thing comes from a paternalistic mindset of doing things to people, rather than enabling them to see their situation, perhaps differently than before.

In wicked problems, if you, the outsider, are doing anything other than facilitating or acting as a go-between, you're part of the problem.


> I question whether anything was actually done

It sold services. There's a big market for that.


Apparently they're also solving Ukraine's need to provide medical care to people wounded in the war, and their primary contribution is to suggest using RFID bracelets. Even if you fully accept that patient tracking is crucial, jumping to RFID seems wholly unnecessary. Wartime is exactly when you do NOT mess around with fancy tech and hard-to-source materials; there is not even a hint here of why barcodes wouldn't be the solution, just weird fetishizing of RFID tech.

https://drive.google.com/file/d/12a2aLdG25YaPUzFwpVlIeO346DU...


I quickly skimmed through the presentation and you seem to be misrepresenting it.


The superficiality of the presentations reeks of consultancy. I wouldn't be surprised if all students ended up in KPMG-Deloite-Young. Even the smallest problem, something design ships something, only showed that the students had actually spoken to people, but not listened or understood, and it was smothered with a bad case of AI, the very thing called out at the start as an example of "not wicked."


One could argue, according to Cynefin, that "wicked problems" are where Lean and Agile diverge. When you're operating in a generally-known environment where what's unknown are things like product/market fit or codebase behavior, it makes sense to operate in short iterations or kanbans with rapid customer feedback, while still trying to use Lean to drive efficiencies along the way.

But if you really have no idea what's going on, there's no way to know what is "efficient," and so you have to fall back on more pure Agile, i.e. Cynefin's recommendations for the unordered "chaotic" or "complex" domains: do something small, analyze how that affected your environment, learn from that, and respond/repeat.


> do something small,

> analyze how that affected your environment

> learn from that, and

> respond/repeat.

I could be missing your point, but this is precisely a core tenet of Lean - Plan, Do, Check, Act

Agile speaks nothing of these sorts of processes(aside from the cargo cult enterprise scrum nonsense of course)


> Agile speaks nothing of these sorts of processes

Agile is less specific about the how, but is very much centered on team-led adaptation of concrete process to specific circumstances.

Lean does the same thing, but actually talks about how to achieve that.

(Lean and Agile are mostly built on the same ideas, but the Lean literature comes at the ideas from an engineering mindset, while Agile literature does it from a fuzzier and more touchy-feely mindset.)


The pissing match above has at its root a global namespace conflict around the term “Lean”.

To clarify, the underlying heirachy is:

Shewhart cycle “PDCA” (1939)

Lean manufacturing “Lean” (1988)

Agile Manifesto “Agile” (2001)

The Lean Startup [subset of lean] (2011)

Lean around here often handwaves to mean the Lean startup with Lean manufacturing under that. Technically I’d suggest Lean on its own is really a reference back to Lean manufacturing and the principals of Lean developed by Toyota in the 1960-1980’s in what they later named “The Toyota Way”


You leave out Lean Software Development (2003-, there are several works from the same authors), which is fairly directly about applying/adapting Lean manufacturing to software development. (And is situated within the Agile space, but a lot more focussed on the meta-level process of controlling/adapting process than most Agile work, which often focusses on specific process that worked specific places, and gets easily bent to support the kind of adaopting canned process that is anathema to the Manifesto.)


Interesting I’ve followed lean for many years and must admit I was not aware of this. Curious to dive in. Thanks for including this.


You seem to have missed the point of Agile.

But then, that's normal.


To be fair, its largely because the Agile literature as a whole is vaguely written on the most important points.


Yes, everything is very unclear. It's also corrupted on most communication, either on purpose or by accident.

Corrupting Agile was once a very profitable activity.


Did I?

Where does the manifesto talk about a process of any sort? It’s a set of values. It doesn’t mention anything about iterations, breaking things down, etc


PDCA, in more humanities than engineering terms, is basically the final Agile Principle “At regular intervals, the team reflects on how to become more effective, then tunes and adjusts its behavior accordingly.”

https://agilemanifesto.org/principles.html


This is PDCA:

> Welcome changing requirements, even late in development. Agile processes harness change for the customer's competitive advantage.

And the sibling has the one that states PDCA for the development process too.


I'm reminded of a recent thread: "The design thinking movement is absurd"

https://news.ycombinator.com/item?id=36893635


Steve I have read your blog posts sporadically for years and am running a startup working on a wicked problem area (fully autonomous food preparation, retail and logistics) and would be happy to provide a case study in practical approaches we have used. Incidentally I will be in CA in the next few months and we are raising.

Frankly our best strategies to date have been common ones: business-level modeling as a means to ensure strategic validity, in-depth study of predecessors and competitors, commitment to rapid iteration, brainstorming, first principles thinking, breaking down the problem, broad and dynamic base of contributors (different professional skillsets and mentalities), extensive documentation, looking at the whole problem, investing in developing required specialist skills, etc.

After seven years of hard work according to public sources we are now the world leader in our space by multiple objective metrics, from which I would suggest we have been successful technically as well as financially.

IMHO there's no secret sauce. It takes a lot of time and money and concerted effort. A lean / startup approach just allows more efficient execution.


Dude wants a wicked problem, he should figure out how to change the name of the institution he works at.

"Imperial" College? In this day and age?


> Solving Disinformation/Information Pollution for the BBC

This isn't really a complex problem; it is an intractable problem. It is very easy to solve disinformation - have a good editor who notes the quality of the evidence clearly.

The problem comes in to play because once there is a reliable brand, partisans/corporate suits/do-gooders/hacks see value and set around corrupting the brand to extract value from it. It doesn't matter how good the solution is, eventually someone will decide that reality is pulling in a direction they don't like and bam! disinformation and censorship.

The BBC is not in the business of neutrality. They are a reputable front for British propaganda. Seems likely that one of the reasons the quality is so high is someone is purposefully keeping the channel clear so they can nudge the debate more effectively. Especially after we discovered what the US agencies were doing in social media with the so-called Twitter Files - professional government disinformationists embedding in media outlets is obviously the done thing.


It turned out that was Taibbi's own invention, changing an acronym so that it pointed to a government agency when it really didn't:

> Taibbi claimed “the EIP was partnered with the government Cybersecurity Infrastructure Agency—CISA—to censor Twitter,” but the reporter mixed up the government agency with the nonprofit firm Center for Internet Security.

https://www.thedailybeast.com/msnbc-host-mehdi-hasan-makes-m...


> It is very easy to solve disinformation - have a good editor who notes the quality of the evidence clearly.

- how do you identify candidates?

- how do you know that the candidate that is selected is actually what the judges think they are?

- how do you know that the candidate is not at least partially compromised, on an ongoing basis?

- for any given editorial issue, how do you know that the candidate is performing perfectly (or, when you said "solve", were you perhaps joking)?

- is that air you're breathing right now?


> It is very easy to solve disinformation

I don't know about that, but it is very easy to make statements like this which make your credibility go "poof."


You havn't managed to provide an argument our counterexample, so it is obviously at least something of a challenge to find unsolvable disinformation. Most of the easy examples are intensely political and it is obvious that motivated reasoning or clashing values is the root cause and no issues to do with facts.

> make your credibility go "poof."

I suspect you are using credibility to assess claims rather than arguments. If that is so, you're never going to escape disinformation because credible people routinely discover they can sell their credibility and do so. And credibility has never been a measure of truth, lots of credible people are also key spreaders of disinformation (like the BBC, for example). That would make the problem intractable. It is pretty normal for people to optimise for credibility, which means the disinformation is generally pushed specifically by people who are legitimately credible.


Editing every topic manually in the face of automated generation of misinformation may not be a solution that scales


Narrator voice: it's not actually very easy to solve disinformation.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: