2010-09-15

Old dogs, new tricks.

With some regularity I bump into people who work for companies that need to evolve and reinvent the way they currently do business -- or at least how parts of their business operates.  Key words are: global reach, open APIs and much, much faster innovation cycles.  For companies that have been operating in mature sectors for a number of years there is usually no deep understanding how they need to respond to the changes in premises and expectations.

Quite the contrary, there is usually a lot of resistance to challenging basic assumptions and established "facts".

The first thing that strikes me about how a lot of well-established companies go about doing "new, exciting and wild stuff" is to apply all their normal, rigid, heavy processes.  Some places there is usually a lightweight process in place -- which still is far more rigid and heavy than what is the case at more, I dread to use the term, agile companies.

This leads to expensive failures.

My advice is: don't expect too much and don't plan too much.  Just start learning by doing.  If you are an old dog, you need to learn the new tricks first.  The things you need to learn are:
  • Fail
  • Fail fast
  • Fail cheaply
  • Get up and have another go
Most Internet successes you see are the result of dozens of failures, false starts and heads being banged against dead ends.  Most mature businesses are used to a better than 10% success rate for projects they do. It took Twitter 10 years and half a dozen false starts to become an overnight success.  And those people were adaptable and entrepreneurial.  Chances are you, or your company, are not.  You will fail a lot.  Get used to it and make the most of it.  Most failures still yield some valuable results.


Once you have learned how to fail properly (or at least become mentally prepared for it) you are ready to start learning.  Work fast, iterate over both ideas and solutions.  Be prepared to change your mind and re-evaluate your assumptions.  Don't do more work than strictly needed (following the You Aren't Gonna Need It principle), but take care to do the things that matter properly.  Quality always matters.  You will never go back and fix things that seemingly work so do it properly the first time around.

Having a plan for how you want to scale also matters a lot.  Actually being able to scale to infinity (and beyond, sorry, mental tic) in the demo version, or closed/limited beta doesn't.  Not really.

However, when you go live you should be able to handle the worst case scenario:  that your stuff actually takes off.  Nothing turns users against a service quicker than sluggish performance and downtime.  (Twitter could get away with it, but you might not.  In particular if your brand has certain expectations associated with it).

Once you do launch, congratulations, you are almost half way!  Those pesky users will not do what you intended them to do.  They will find their own uses for your neat new offering.  This is Part Two of the learning experience.  Now you have to iterate to converge towards what the user actually wants and needs. (Just look at Apple's Ping service.  It more or less fell flat on its nose when it was launched.  Which isn't too strange since it missed the mark by a mile. Unless they manage to iterate some sense into Ping, it will have to die).

If you were too lazy to read the entire post, let me summarize: learn by doing, don't expect quick returns and be prepared to change your mind often.

What I've written about here isn't about any one company in particular.  Incumbent companies running the risk of being de-throned and fading into irrelevance is an area of great interest to me and about 18 months ago I specificly set out to seek out challenges of this type.

This is what happens when people tell me that something "can't be done" :-).

No comments:

Post a Comment