mad_bear,
Regardless of whether your technique is pilotage or dead-reckoning, there certain things you can do to make VFR altogther easier.
Apart from the basic speed/heading/timing, I think firstly, one must look at the route on the chart and pick things that you think will be easey to spot and plan to use those. The trick, as you have discovered, is to pick things that are easy from the air. Don't be tempted to try and id features by going ground to map. To remain in control maintain a map to ground philosophy. The ground to map id method essentially means you are no longer confident that you know where you are and takes much more time away from flying accurate headings. Also, always orientate your map so the direction you are travelling in faces forward/up - if you don't you'll get caught out one day!
Things which I find most easy to spot are:
The shape of large lakes
Complex Road Junctions
The shape of a town
The relative co-location of small hamlets
Parallel road/canal/river/active railway line
Stacks (towers, ariels).
Single carriageway road with dual carriageway sections.
Motorways
The way roads radiate from or orbit around towns. You have probably misidentified a town with four roads to the west if the maps shows only three.
Things are can be extremely difficult to find but tempting to use:
Windfarms - the big ones are easy, but the CAA chart sometimes suggests multiple when only one itsy-bitsy windmill exists
Minor roads, rivers and disused railway lines: They can all look the same!
Smaller lakes, some are shown, some aren't.
Disused airfields. Some stand out for 10 miles. Some are just marks on the ground. Same goes for some active airfields!
Once you get tuned in, you start using the following:
Forests
Pattern of high ground
Pure timing, because eventually you'll learn to have faith that maintaing your heading accurately for the right amount of time at the planned speed always puts you roughly where you planned to be
Regardless of the above, as others say, it is difficult. But general confidence comes from getting the brain to see how features are placed in relation to each other, rather than concentrating on finding a single feature. The former guards against mis-identification and help greatly with spotting the 'pattern' before you get there.