Archive for October, 2016

Dangerous Expectations

buttercupExpectations result from mental models and wants. When you have a mental model of a system and you want the system to behave in a way that fits your mental model, that’s an expectation.  And when you want the system to behave differently than your mental model, that’s also an expectation.  When the system matches your wants, the world is good.  And when your wants are out of line with the system, the world is not so good.

Speculation is not expectation. Speculation happens when you propose, based on your mental model, how the system will behave. With speculation, there’s no attachment to the result, no wanting it to be one way or another. There’s just watching and learning.  If the system confirms your mental model, the applicability of the model is reinforced (within this narrow context.) And when the system tramples your mental model, you change your mental model.  No attachment, no stress, no whining, no self-judgement.

When doing work that’s new, system response is unknown. Whether the system will be exercised in a new way or it’s an altogether new system, metal models are young and untested.  When it’s the first time, speculation is the way to go. Come up with your best mental model, run the experiment and record the results.  After sitting in data, refine your mental model and repeat.   If your mental model doesn’t fit the system, don’t judge yourself negatively, don’t hold yourself back, don’t shy away.  Refine your mental model and build-test-learn as fast as you can.  And if your mental model fits the system, don’t judge yourself in a positive way.  This was your first test and you don’t understand the system fully.  Refine your model and test for a deeper understanding.  [Note – systems have been known to temporarily conform to mental models to obfuscate their true character.]

When doing work that’s new, expectation gets in the way. If you expect your models to be right and they’re not, you learning rate is slower than your expectations.  That’s not such a big deal on its own, but the rippling self-judgement can be crippling. Your emotional state becomes fragile and it’s difficult to keep pushing through the work. You doubt yourself and your abilities; you won’t put yourself out there; and you won’t propose radical mental models for fear of looking like you don’t know what you’re doing.  You won’t run the right experiments and you never the understand the fundamental character of the system. You block your own learning.  If you expect your models won’t to fit the system, you block your learning from the start.  Sometimes your lack of confidence blocks you from even trying. [Note – not trying is the only way to guarantee you won’t learn.]

Within the domain of experiments, mental models and generic systems, it’s relatively easy to see the wisdom of speculations and the perils of expectations, where wanting leads to judging and judging leads to self-blocking.  But it’s not so easy to see in the domain of life where experiments are replaced with personal interactions and generic systems are replaced with everyday situations and mental models are ever-present.  But in both domains the rules and consequences are the same.

Just as in the lab, in day-to-day life expectations are dangerous.

Image credit – Dermot  O’Halloran

 

 

Celebrating Seven Years of Blog Posts – what I’ve learned about writing.

%desxToday marks seven years of weekly blog posts.  Here’s what I’ve learned so far:

When you can write about anything, what you choose tells everyone what you’re about.

Sometimes you’ve got to start writing to figure out what you have to say.

Some people think semicolons are okay; others don’t like to show off.

When you don’t want to write and you write anyway, you feel good when you’re done.

Use short sentences. Use fewer words.

Writing is the best way to learn you don’t know what you’re talking about.

Writing is a good way to have a deep conversation with yourself.

Worrying about what people will think is the surest way to write like crap.

Writing improves by writing.

When the topic comes slowly, start writing. And when the words don’t come at all, repeat.

If you don’t know what you are talking about before you start writing, no worries. You’ll know when you’re done.

When you have nothing to say it’s because what you have to say is too personal share.

For me, writing is learning.

 

Image credit David Kutschke

 

What if it works?

jumping-dogWhen money is tight, it’s still important to do new work, but it’s doubly important not to waste it.

There are a number of models to increase the probability of success of new work.  One well known approach is the VC model where multiple projects are run in parallel.  The trick is to start projects with the potential to deliver ultra-high returns.  The idea isn’t to minimize the investment but to place multiple bets.  When money’s tight, the VC model is not your friend.

Another method to increase the probability of success is to increase the learning rate.  The best known method is the Lean Startup method.  Come up with an idea, build a rough prototype, show it to potential customers and refine or pivot.  The process is repeated until a winning concept finds a previously unknown market segment and the money falls from the sky.   In a way, it’s like the VC Model, but it’s not a collection of projects run in parallel, it’s a sequential series of high return adventures punctuated by pivots. The Lean Startup is also quite good when money’s tight.  A shoe string budget fosters radical learning strategies and creates focus which are both good ideas when coffers are low.

And then there’s the VC/Lean Startup combo. A set of high potential projects run in parallel, each using Lean’s build, show, refine method to learn at light speed.  This is not the approach for empty pockets, but it’s a nice way to test game changing ideas quickly and efficiently.

Things are different when you try to do an innovation project within a successful company. Because the company is successful, all resources are highly utilized, if not triple-booked.  On the balance sheet there’s plenty of money, but practically the well is dry.  The organization is full up with ROI-based projects that will deliver marginal (but predictable) top line growth, and resources are tightly shackled to their projects.  Though there’s money in the bank, it feels like the account is over drawn.  And with this situation there’s a unique and expensive failure mode lurking in the shallows.

The front end of innovation work is resource light. New prototypes are created quickly and inexpensively and learning is fast and cheap.  Though the people doing the work are usually highly skilled and highly valuable, it doesn’t take a lot of people to create a functional prototype and test it with new customers.  And then, when the customers love it and it’s time to commercialize, there’s no one home. No one to do the work. And, unlike the relatively resource light front end work, commercialization work is resource heavy and expensive. The failure mode – the successful front end work is nothing but pure waste.  All the expense of creativity with none of innovation’s return.  And more painful, if the front end was successful the potential failure mode was destined to happen. There was no one to pick it up from the start.

The least expensive projects are the ones that never start. Before starting a project, ask “What if it works?”

image credit – jumping lab

Make it work.

square-pegIf you think something can’t be done, it won’t get done.  And if you think it may be possible, or is possible, it may get done.  Those are the rules.

If an expert says it will work, it will work.  If they say it won’t work, it might.  Experts can tell you will work, but can’t tell you what won’t.

If your boss tells you it won’t work, it might. Give it a try.  It will be fun if it works.

If you can’t make it work, make it worse and then do the opposite.

If you can’t explain the problem to your young kids, you don’t understand the situation and you won’t make it work.

If something didn’t work ten years ago, it may work now. Technology is better and we’re smarter.  More likely it would have worked ten years ago if they ran more than one crude experiment before they gave up.

If you can’t draw a one page sketch of the problem, it may never work.

If you can’t make it work, put it down for three days. Your brain may make it work while you’re sleeping.

If you don’t know the problem, you can’t make it work.  Be sure you’re trying to solve the right problem.

If your boss tells you it will work, it might.  If they tell you how to make it work, let them do it.

If none of your attempts have been fruitful and you’re out of tricks, purposely make one performance attribute worse to free up design space. That may work.

If you don’t know when the problem occurs, you don’t know much. Your solutions won’t work.

If you tried everything and nothing worked, ask someone for help whose specialty in an unrelated area.  They may have made it work in a different domain.

If you think everyone in the group understands the problem the same way, they don’t.  There’s no way they’ll agree on the best way to make it work. Don’t wait for consensus.

If you don’t try, that’s the only way to guarantee it won’t work.

Image credit – Simon Greig

 

 

Mike Shipulski Mike Shipulski
Subscribe via Email

Enter your email address:

Delivered by FeedBurner

Archives