Time to Bake

By | July 11, 2013

Design needs time to develop properly.  You can whip out stuff super fast, but it probably won’t be very good. Design can have logic bugs, sub-optimal flows, unintended consequences, poor usability, unhappy customers and other nasty results.  The worst case is the design hampers your business growth in the future.

What do all of the following professions have in common:  designer, doctor, electrical engineer, architect, programmer, copy writer, teacher, artist, handyman and the guy who writes bills for the senate to approve?

They all require time to do their jobs properly.

This is not to say that infinite time = perfect results.  You can give something too much time.  It is still important to finish a job.  The key question is “how much time?”  Like cooking food, there is a perfect amount of time and temperature.  Too little and it’s raw, too much and it’s burnt.  How does one determine the proper time table and resource allocation?

I find a very quick diminishing returns when you add more people to a design project.  Having 4 designers on one project (even a large one) will yield a ton of communication overhead and misaligned designed principles.  1-3 is good, but I would say 2 is ideal.  I usually “pair program” with another designer and get solid results.

In terms of time, there is a trend to do lightweight or “lean” design.  “Get it done quicker and start programming!”, they say.  Truthfully, I am not seeing this trend yield better business results than taking more time in the design phase.  It is resulting in alot of guessing without proper thought.  Good designs take twists and turns until they bake properly.  Good ideas are thrown away, improved, resurrected and changed.  If programming starts, it makes it very hard to adapt to learning during the design process.

Not all projects are created equally.  I have a general gut-feel for how long designs should take.  Usually, I try to start extremely early, before the business even commits to the project.  This helps me get a head start.

Then, I usually do a 10,000 foot design and sketch out generally how the thing will be approached.  No detail, but an architectural guideline.  This usually yields a good feel for how complicated the project will be.  Here is a general cheat sheet for size of project and how much time I think to allocate:

Trivial Size
Many designs literally need a single screen.  I usually allocate a couple of hours, although in a pinch you can draw something on paper.

Small Size
This is a basic flow of some sort.  Nothing complex, but has some detail.  This can take a few days to two weeks.

Medium Sized
OK, now we have something meaty.  This project has lots of details and several functions / flows.  A new website often fits into this category for those of you who don’t work on applications.  Give this one a month of design to get it right.

Large Sized
Major projects often take more than one person and are going to go through several iterations before it settles down.  The goal is to move towards “convergence of ideas” where the design can move into engineering before the design is done, but after it stops churning on a daily basis.  This can take many months.  One large project I recently finished that had ample design time took 9 months.  It came out beautifully.

So what is the lesson here?  Design takes time, but we all want to ship code too.  Don’t lollygag .  Give enough time to bake, and not a moment more.

One thought on “Time to Bake

  1. Ben H

    Glen, this is one of the most helpful guides you’ve put together – particularly about the timing needed to let designs fully bake… Also, I love the work you and your team are putting together!

    Reply

Leave a Reply