I don't follow a particular process per say, but I do have a general guideline that I follow while making maps.
Step 1: The Idea
First I identify what the central idea of my map is. A big single fortress owned by both teams? A siege on a base from many sides? A running map with dynamic objectives? A fast-paced "fun" rather than "balanced" ctf map? A map that addreses the issues inherant with a valve map? Identifying what my goals are for the map I'm working on is often the first vital step in producing a level.
Step 2: The Theme
WHERE is it? From tried and true Desert to an outer-space rendition of 2fort to a 3-d version of the original Donkey Kong. What you envision the final "look" of the map to be (or should be) will have a large impact on your design, especially for the buildings. For example, a desert town may have delapitated wooden buildings that only provided partial cover, while an alpin semi-urban center or Spytech base may look sparkling new and have smooth defensive barricades. An alpine setting may have tall, steep rooftops to keep rain out of the buildings, while a dustbowl environment could easily be nearly totally flat and squad, the buildings showing the character of the desert.
Step 3: Draw Draw Draw!
Now that you have an idea and a theme, it's time to start sketching. Draw the layout, draw sights that players would likely see as they progressed through the stage, draw top-down and orthographic representations of your rooms, draw until your hands are sore and you have a good idea of how your map is going to be laid out before you even TOUCH hammer. Personally, for this point I normally take a stack of stickynotes and just start sketching. Oftentimes I'll have a very rough layout note in the middle, and i'll just post the scene and top-down notes at the approximate area of that note's location on the layout sketch. By the end of this stage I've already got a good idea of what my final layout is going to look like for the map, and might even have an idea of how combat is going to flow within it.
Step 4: Block it out.
Take dev textures (I stick with gray, and a bland rockwall) and start getting that layout down! Try and avoid doing too much detail work at this stage. Instead you should be focusing on getting your layout right. In this stage problems such as ceiling heights, crouching heights, and jumping heights should be addressed to avoid problems later.
Step 5: Functioning Objectives
Right, now that you've got your layout *roughly* done it's time to get the map working! Read (or review) the tutorials for making spawn rooms, objectives, and important triggers. Right now you want to get the map playable so you can playtest as soon as possible. Step 5 and 4 normally are done at the same time, at least for me, but really they are quite different steps and should be treated as such. This part is probably the easiest part of the entire stage, and once you have objectives and some basic lighting/skybox lighting in... well your at Alpha 1.
Step 6: Refine and Texture
At this point you should be playtesting your map with friends, clanmates, or here. Everything is together and working now, and although ugly your map is perfectly workable. Listen to the advice your playtesters give you and adjust your map accordingly if you even half agree with their opinions. This is when major layout changes are the easiest to make and you should take advantage of that. While you get your map closer and closer to it's final layout you should also focus on making your brushwork more detailed, as well as starting on your texturing. Once you texture an area you should also work with it's lighting, since lighting is the #2 most important thing within a map. (right after layout)
In this stage you should also place any large or important gameplay props. Such as the boulder on badlands.
Step 7: Refine and Props (and displacements)
We're still an alpha, but by this point you have went over the entire map with a paintbrush and carpenter nails. Your layout has changed significantly, and dozens of minor problems or unconceived issues have been quashed one by one in a flurry of intense testing. Now it's probably a good time to publicly announce your map and regularly bringing it up for gamedays and playtests. While the testers will hunt fiercely for any problems in your map YOU have a much harder job now: Find locations for props that add to rather than detracting from gameplay. Now that you don't need to worry about texturing so much, you can focus on making your lighting perfect, and getting just the right overlays up. Displacement work, especially displacements outside of the play area, is also ideal for this stage. You're near the end of your alpha, but before you go beta you can't just have a 'working map', it needs to be a 'damn good looking' map. When you finish this stage you are very nearly ready for beta.
Step 8: Optimize
You've read up on how func_details, areaportals, occulars, nodraw, and hint/skip brushes work right? No? go do it. Yes? Go back and review it. It's time for us to make this map run WELL. Now that the layout is mostly finalized and the lighting is solid, you can make decisions on how best to optimize without needing to worry about screwing things up later. Get it right the first time, halve your normal compile time and double your framerates! ONTO BETA!
Step 9: Refine and Texture Mrk II
Playtest, playtest, playtest! Get your map out there so people know about it, fight in lines to get your map into events, talk to as many server owners as you can! The more you test your map at this stage, the better it will be and the more likely it'll be a big hit. If people like your map, then PUSH it! If they don't, FIND OUT WHY! Adjust, edit, fix, and nudge until you are so tired of your own map that you need a week's vacation.
Then, take all of those suggestions, screenshots, and advice you got in your playtests and disappear for a month. iron out every possible problem in your map. Try and check every brush face visible to make sure it's perfectly aligned, and make sure that every detail is exactly as you want it to be. Once you are happy with it, check it again, adjust it four more times to be sure, and then release a new beta. Once you, and all of your testers, are certain there is nothing left to do.... well...
Step 10: Optimize and Final
Give that map one more final look-over for it's optimization, then throw it out the door and hope that it lives! Obviously you should still push your map, but at this point it's really quite grown up and... if it's good... well-known in the customs community, and potentially by a few people outside of it. Hell, at this point you might even suggest to Valve to take a look at it, they've made maps official with less work done than you have. Perhaps while you wait for them to include it in the next update you can focus on a new project.