If you have your leaflet distribution under control - that is you know they are going out to the right houses and in the right volume then - and only then, is the time to work on the actual design in detail. Well, of course you have to have some leaflet to be handing out first but here we can just take a very obvious basic design. What is it? How do I get it? How much is it? What are the benefits? How easy etc. These things with perhaps a nice graphic and all very clear and easy to read are a great first step. I'd never recommend starting with a fancy or complicated leaflet design as it just introduces too many variables. A plain "straight" looking Ad gives you a standard to work from - a result to compare other fancy ideas to - otherwise you have nothing to compare it to. There is always the great story from the insurance sales company. They spent a year comparing results from door knocking sales people. They were very interesting on the subject of Leads - as opposed to Sales. They discovered that when it came to leads it was simply the number of doors knocked (actually speaking to the occupant of course!) and it made no difference on the quality of the Sales person involved. Every 32 people spoken to (on average) produced a lead. As long as they spoke to the occupant and delivered the script they got a lead. Now when it came to closing that lead the quality of Sales Person involved made a very big difference - but that is a different subject to ours - leaflet distribution and leaflet design. The two factors that monitored lead generation where the script and the actual number of people it was delivered to. Hence step one is delivering the leaflets to a sensible house or business in volume with a clear message - then the design (our script) comes next to improve the response for the sales people.
The next 'big thing' is to change one thing at a time on our leaflet design. You have to do this as if you change two or more things you can't figure out what caused any change or if they all cancelled each other out. By trial and error you can discover what makes response go up and what makes it go down. Here a big guiding rule is to use the response as the factor - not your own preferences or take on the matter. This is one of the biggest killers - running your own hobby horse as to why people should want or use your product rather than just monitoring the volume of response and letting that be your sole guide. We presume your mission is to generate as many leads as possible for your sales team - of the right target market public - as possible. Not to show everyone how right you are or clever with Adobe Photoshop. That is a different game.
Another golden rule is to then pilot the changes to see what works and what doesn't. So, if you have a base design pulling one lead in 200 leaflets and that is viable - great. By all means look for one that does better, but don't work out a new design and replace your bread and butter design with it for a month or a week to see what happens! If it crashes it will cost you dear and you should never experiment with a working line. Instead - do as the big companies do and try the new leaflet design out in a smaller area on the side as an extra. When it has proved itself better over a few trials then and only then move it to the main line with confidence that your stats are going to go up rather than down.
You do need to run trials on a fairly regular basis - not only to keep looking for the better design but to pick up when your market changes and the old button is worn out and you need a new one. Depending on your business this could be decades or it could be weeks! You'll hopefully know which applies - carrots were likely "fresh and crunchy" in Victorian times too but a Nokia 3010 would be considered ancient!
By all means have artistic experiments and have a go at the odd mad idea - you'll never know if you don't try it. And if you get conflicting results be sure to really check on the volumes of leaflets delivered - I'm afraid that all to often it wasn't your design that was sending the stats up and down but the industry of the leaflet distribution staff.
Lastly - there are outside influences that can send stats up or down over a short period - be sure to monitor the long term where things settle down and you can more clearly see the stats of an individual design or company.