Tuesday, July 30, 2013

Iterative Testing VS Big Swings in Product Design

When designing a new product, you need to start somewhere.  That starting position is your first big swing.  It’s an initial launch that puts a stake in the ground that states emphatically, "This is what I want to build." It’s often rough, somewhat broken, and filled with tattered assumptions, but it’s the beginning, and that’s big.

It’s at this point you face your first choice between iteration vs taking another big swing.  The choice is never obvious.  

Lets start with iteration:

Iteration, when done both well and quickly, is always a path of progress. Small improvements on the low-hanging fruit in a product are easy and meaningful wins.  It’s a hill climbing algorithm. You’re nearly always guaranteed to arrive at your local maxima in you put your head down and grind like crazy in fast, data-driven, iterative product development.


But, the global maximum (“Your Real Goal”) might never be within your grasp with this approach. Enter the siren song of the Next Big Swing.

The next big swing will never take you to the global maximum. Ever. A big swing is like getting dropped (somewhat) randomly on the map above. If you’re lucky, it might drop you on the correct hillside such that if you start iterating again, you’ll eventually arrive at the global maximum.

Sometimes after a period of iteration, a big swing is the best next step. See:

Odeo -> Twitter

Zimride -> Lyft

Burbn -> Instagram

And sometimes just putting your head down and staying focused is the best path.  The difference between Twitter’s MVP and the product you see today is a combination of A) constantly iteration with good feedback loops in quick dev cycles B) holding on for dear life during massive back-end scaling C) making the experience ubiquitously accessible across all platforms.  They’ve accomplished a lot, but the product today is very recognizable in the context of their MVP. 

One mistake I personally have fallen victim to repeatedly is thinking that iteration = small choices.  That’s not true at all.  The feature/improvement you choose to iterate on in your product should be the one that will be most impactful in pushing you closer to your big goal. If your hypothesis of the impact from iteration feels small then either A) you’re iterating on the wrong thing or B) it might be time for another big swing. 

But be wary of the temptation of (B).  (B) is exciting; it seems so big and bold in contrast to constantly grinding away on iteration. But it’s also a bit of a restart, admitting defeat for recent iteration. Not something that should be indulged lightly.

Big swings are like exclamation marks in writing. You only get a couple… use them sparingly.


  1. srizah reblogged this from thegongshow
  2. edwardshepard reblogged this from thegongshow
  3. dlev reblogged this from thegongshow
  4. mayerzahid reblogged this from thegongshow
  5. flungster reblogged this from thegongshow
  6. dontmakecrap reblogged this from thegongshow
  7. caterpillarcowboy reblogged this from thegongshow and added:
    First time I learned about hill climbing was in my genetic algorithms course, a PhD level course that was cross listed...
  8. thegongshow posted this