I think it's important to set at least rudimentary benchmarks/milestones in terms of growth and brand recognition.
After that, throw a bunch of stuff on the wall and see what sticks. The few ideas with the most ROI in terms of 1. traffic and 2. user retention should be at the forefront of your future strategies.
This may not sound like much of a plan, but as long as you have a mini-strategy in place for every one of your marketing efforts, and as long as you understand and, if possible, participate in the community you are marketing to, the try-everything approach should work.
Seriously: Marketing success (like software development success) seems to be highly correlated with getting something out there and iterating as frequently as possible based on feedback.
Writing a plan is fine (if that floats your boat), but it often creates the delusion that you can actually control and predict how things are going to go.
"In preparing for battle I have always found that plans are useless, but planning is indispensable." ~ Dwight D. Eisenhower
If you are wondering about the volume of planning we do, there's no question that I've thought about marketing for a huge amount of time.
