For simple wireless design applications, there are several things that affect how you approach the project.
1) What is the volume? Unless you are making more than around 10K units, it's probably best to buy someone’s pre-made module. This is because of the time to design and agency approval issues.
2) What is the range needed? Through what kind of obstacles? How big a space do you have for your antenna? These issues affect “link margin” which determine how much power you must transmit and sensitive your receiver will need to be.
3) How reliable does the link have to be? Does it matter if a few transmissions are missed? In other words, if your wireless link closes the loop in a process system, the link must be robust, reliable under all conditions, probably will require receipt acknowledgements, and contain error correction. If the data is for a less critical use, then a missed data packet will probably not be important and latency may not be an issue. This issue will also influence the best type of modulation to use and the networking
4) Does the application require battery power? Battery operated data links cannot be ON all the time. This then requires synchronizing the ends of data link and what kind of networking layer you may need to consider.
5) How much data are you communicating in each burst? A few bytes? A few Kbytes? Zigbee/802.15.4 are good for a few tens of bytes. If you are trying to do digital audio or video, then you’re looking at something more like wireless LAN or Bluetooth.
Within Engineering tips, there is also an Antenna and Propagation Engineering Forum and a Microwave and Electromagnetic Engineering Forum.