Design Discussions Matter!

Consider this scenario:
You have a box of Lego blocks, you have 10 people in a group. You are the moderator and you have an idea. You start by piecing something using just 5 blocks. You have to pass the block around to each individual in your group and you give them 1 instruction : "You can add 2 pieces to the assembly any more and you have to remove 1 for each additional piece. You cannot remove pieces such that the structural integrity of the assembly is comprised."

If you assess the iteration after every round you will begin to notice a pattern. At a particular point an idea is injected and as the evolution progresses the idea is transferred for person giving it to the person receiving it. At the end of a 10 person iteration you will notice a dramatic change to your initial assembly.

The key is perspective. If you super-impose this process on interface or software design you get concept of User Feedback. But most of us are used to testing what we conceived as an intuitive user interface and interaction work-flow against what a user feels about it. But what I'm trying to get to is far earlier than you've even shipped something off the drawing board. The decisions and trade-off made when you first start sketching .

That seems like a rather long preface to what I'm actually going to get into, but I felt it would be nice to add some flavor to a rather interesting design discussion at my work place.

3 Designers (User experience and Interface designers) and 1 Design pattern - Wizard flows.

What is the main objective of a Wizard Flow and what are some of the best implementations and some shortcomings in them. We got into a room, hooked the clunky Think Pad T500 to the projector, stacked up on some caffeine and started talking.

What we ended up doing is first a competitive analysis of different wizard implementations and then a rational implementation of a wizard flow for our application.

We analyzed EBay's search model, Amazon shopping model, Apple and Dell's Store purchase wizard, Geico's insurance quote system and Turbo Tax's tax filing work flow.

What did we deduce? Actually it was a series of questions.

• What are you selling?


• What are you promoting with a sale?


• How aggressively do you intend to promote the sale?


• What is your users objective?


• Do you know enough about your user to make that sale?


• How informed is the user at the end for the sale?

Lot of questions but who's going to answer them? That's precisely the challenge an experience design should address. I'm going to skip the gory details and get to the point, "What did I consider in implementing my design"

My first decision is that for a wizard flow to be successful I need to make sure that the user is actually interacting with a smart software.

I need contextual and suggestive help.

For that to be successful I need to make sure that I have enough information about the user I'm dealing with.

I need User Data Input.

I need to understand if I'm dealing with someone who is casually looking around or with some one who is serious about it. Users hate filling out forms or registering for the service just to find out what you have to offer. Based on that I need to decide when to prompt him to provide the data that I need.

I need to inject subtle data gathering at intermediate stages of the flow.

If its a large sale process that may involve 4 or more stages, I need to ensure that I retain the users attention and keep him motivated to complete the wizard.

I need to provide progress indication, state saves and casual contextual text to motivate the user to move ahead.

I had to create a sense of empowerment to the user. Allowing the user to make the decisions, be informed about how much its going to cost and at any-point feel free to leave.

I need a clear and visible affordance to provide information about items purchased, discounts applied and ensure that the progress is saved.Something like a mini-cart.

After all that I was motivated enough to actually hit the drawing board and create this Frankenstein work-flow that I think will best suit the bill.

I created a flow with 5 stages (in a crude nutshell)

• Stage 1 :- Allow user to select from a set defined list of package bundles. Then buy it or customize it.


• Stage 2 :- Get the basic data about the user is a short form that had the high-5 for my wizard to actually help customize the sale. Then buy or continue customizing.


• Stage 3 :- Inform the user that the system need just a little more information to help understand what to suggest. The user can fill out the information to see additional suggestions to understand the nature of the sale and then decide to buy.


• Stage 4 :- Entice the user to buy additional package that may add a discount at the same time suggest only what the user may need. User can skip or look for discounts.
  Stage 5 :- Get the rest of the payment information, ask if they would like to create an account and make the sale!
  
  If the user decided not to create an id and consider it as a one-off purchase, auto generate a temporary id that the user can use to if they return. The next sale will be so much easier!

Bottom line:Peers aren't around adding mass to office space. As a designer you need perspectives, some you may never even think of. It's easier to collaborate and reap the benefits than spend all those dollars on 4 rounds of iterative user testing, because you dint catch something earlier on.

PS : I had 2 rounds of user testing and the highest success rate of all the work models I've designed and tested! :)