UX, like development, is a process. Good results are an outcome of a good plan. Now, the great thing about UX is that it is far more flexible than most development methodologies and as a UX professional, picking and choosing which techniques best fit the current project is easy to do. UX can be scaled down for good, quick results or ramped up for 'knock your socks off' great results. It all depends on your particular situation.
BUT, (there's always a ‘but’) it is very hard to get really good results without following this framework:
Know Your Audience & gather data
Inception is not a term typically used by UX professionals, but it doesn’t help equate a stage of the UX process to a stage in the development process. This is the information gathering stage in which UX professional learn about the target audience. One of the biggest mistakes too many UX efforts make is in not creating and living by an Audience Matrix.
An Audience Matrix lists out all key influencers that will likely affect how a website or application is used. More than just demographic data, it specifies key factors that are germane to using the product. For instance, does gender really matter when using Microsoft Office? Not really. While things like gender can be a go-to demographic, it often leads to misleading conclusions as test results are almost always better attributed to individual technical knowledge, job experience or position.
We're pretty sure not knowing their target audience was why BIC thought producing a pen for women was a good idea, read the customer comments for a good laugh (http://www.amazon.com/BIC-Cristal-1-0mm-Black-MSLP16-Blk/dp/B004F9QBE6).
Modeling Your Data
Once you have gathered the data, the next step is to summarize what you have found. Personas and Scenarios are the best tools for quickly capturing key user types and how they interact with the website or application. Mental Model are an incredible tool to see into the minds of your target audience, but do take time to build well.
Start Building & Testing Prototypes
The next step in the process is to take your entire data models and start building wireframes, UI (User Interface) layouts, and functional flows. Low-Fidelity is the key here. This is the ‘testing hypothesis’ stage of the process. Much like scientists who have gathered the data, drawn conclusion and build models, now is the time to test those conclusion with low-fidelity (rough draft) versions of the interaction and UI.
Low-fidelity is important because you aren’t looking for graphic design feedback at this point. Whether green is better than red isn’t important here, what is important is does the target audience understand how it works or what they are supposed to do. Can the user complete the task faster, more accurately, more confidently and do they have higher satisfaction? If not, no problem, adjustments are very easy to low-fidelity prototypes, make some adjustments and try again.
Now let’s get that graphic design feedback
High-fidelity prototypes/mockups are next. Here’s where all of the previous stages are combined to produce a final product.
Time to do full-fledged UX / Usability Testing
Now it is a good time to retest any of the findings from the low-fidelity work as well as get that feedback from users on whether red or green is better. The final testing stage is where the rubber hits the road and testers are recruited from the audience matrix, given the scenarios created in Evaluation and tested on the final interactive mockups created in the previous step to get the final findings.
The UX Methodology is meant to have each step build off the previous step and produce better results. It isn’t necessary to do each technique within a step, picking and choosing which makes the most sense for your project is just fine. I strongly recommend that you do at least some from each stage to get good results.