Modifications for Speed (and Efficiency)
The Aerocanard  / Cozy IV design is, in stock plans configuration, a clean, efficient design that achieves very good speeds for the power available:

PA-28R Arrow IV: 134KTAS @ 8,000ft
Stock Cozy MkIV: 188KTAS @ 8,000ft

A difference of 54kts(!) in similar capability aircraft with similar engines cruising at 75% power. Actually, the Arrow IV has 20hp more than the stock Cozy, full retractable landing gear, and a constant-speed prop!
The Arrow does have a larger baggage space (behind the rear seats) though.

But this is experimental aviation, and when building from plans, builders have great flexibility to  ruin  improve the aircraft's performance.

This page contains a collection of modifications that may improve speed and efficiency of the aircraft. Estimates of speed improvement are given where such a figure is available.

Listing of a modification is not necessarily an endorsement of the modification.

These modifications may or may not have been tested. USE AT YOUR OWN RISK.

If you see any error or have additional items or data to contribute, please contact me so I can add it.
Modification Catagories
Modifications relating to the engine itself, cooling, cowling, spinners etc.

Changes to the gear, nose and mains, to reduce drag.

Fuselage changes that may reduce drag.

Wings, fillets, fairings, etc. Extra care here, as you may affect flying characteristics.

A light aircraft is a fast aircraft!

Any speed mod that doesn't fit the other catagories.

Landing Gear
Flying Surfaces
Weight Reduction
Engine Modifications
The most obvious engine change is one that adds more power. This will make the aircraft slightly faster, usually at the expense of increased fuel flow.  More powerful engines are usually heavier, which detracts from efficiency. 

Other changes to the engine can increase power and efficiency with minimal weight penalties.

​Cowl and cooling mods are also covered here, which may improve speed and efficiency without weight penalty.

Upgraded O-360
There are numerous ways to upgrade the recommended O-360 engine for more power or economy.

Modern fuel injection and electronic ignition can provide slight increase in power (especially at altitude) and improved fuel economy. More even mixture to each cylinder can help provide even cylinder temperatures, which may help with balanced cooling and potentially reduced cooling drag. Starting would also be expected to be improved.
High-compression pistions can increase power.

There are many builder and flyers using electronic ignition and a few with modern fuel injection. Deeper mods are rarer. These changes are usually fairly low-risk.

Benefits = Small performance increase, dependent on details and flight regime. Electronic injection / ignition may improve efficiency of engine by 5-10%
Risks = Dependence on the electrical system.

Bigger Lycoming - O-540

O-540s can provide 260hp+, a significant increase in power, providing higher cruise speeds and much-improved climb rates.
At least 3 are flying with O-540 engines, apparently with good success.
Reportedly the economy is similar to O-360 aircraft at equivalent speeds, suggesting any performance hit resulting from the extra weight has been compensated for in additional efficiency, either in the engine or aerodynamically. The aircraft in question usually have many other changes, sometimes including retractable main gear, which likely explains part of this.

Benefits = Moderate speed gains, dependent on engine used.
Risks = Increased weight, centre of gravity issues.

Other Aircraft Engine
There are a few other, less common engine options available.
Several vendors are offering O-360 clones with advanced features for better performance, though you will pay a premium for these. Some have slightly higher displacement, for example.
Non-O-360 engines are available. UL Power offers some interesting air-cooled engines, though one should consult with those already using them first.
The Franklin 6-cylinder engine has been tried, but the additional power did not produce performance gains hoped for, possibly due to the increased weight.

Benefits = Small potential performance increase, dependent on details.
Risks = Small. Some engines may not meet expectations.

Converting modern car engines to aircraft use has been done for decades. By nature, these are typically more experimental and higher risk installations, and tend to be 'one-off'. There are some 'firewall forward' conversions available, with varying quality, usually at high prices (unless you are also considering a NEW aircraft engine!).

Three layers of extra complexity are typically required:
  •   Electronic ignition (and probably fuel injection, though some are still implementing carbs).
  •   Liquid cooling, for which there is no guidance in the plans (and is often poorly implemented).
  •   Gearbox (PSRU) to gear the higher RPMs from these engines to an RPM suitable for a propeller.

Subaru engines are popular, as they have a similar shape to the Lycomings they replace and have some successful history.
A few Mazda rotaries have flown, with mixed success. Cooling is a common problem, as well as hit-and-miss ECU reliability.
Other V6 and V8 installations have been built and flown, which seem to be generally successful after an extended period of debugging. This leads to the good advice: If you want to fly, install the Lycoming; if you want to experiment, go ahead with an auto-conversion engine.
Frequently auto-conversions are removed and replaced by a Lycoming before the bugs get worked out, usually without much fanfare (fair enough, we tend to celebrate successes more than failures).

I had a very good discussion on choosing an auto-conversion here . Many knowledgable and experienced folks gave their input.

Benefits = Potential significant performance and economy gains.
Risks = Gains often not 
realized for various reasons. Increased complexity.

Turbocharging / Supercharging
This can be done with Lycomings but is more commonly seen on auto-conversions.
Usually, a small amount of 'boost' is used to increase engine power (and it does), but the big benefit is the ability to produce 'rated' power to high altitudes, enabling faster climb and high-TAS cruise at altitude.
Turbos are generally preferred over geared superchargers, as they use 'waste' exhaust energy rather than engine power to drive them, thus are potentially an economy gain in some cases.
Extra complexity and weight are unavoidable when turbocharging, due to the plumbing and turbo itself, as well as an intercooler to keep the charge temperature down. The latter is necessary to realize the full benefits of the turbo, as well as to avoid detonation.

Benefits = Good performance increase, especially at altitude. Much better take-off performance at high-density altitudes.
Risks = Increased complexity, more things to go wrong. Cost.

Improved Cowls
It is becoming popular to build custom cowls from carbon fibre, which can produce a cowl that is much lighter than stock fibreglass ones from Featherlite.
The stock cowls also have bumps (to accommodate engine protrusions) and experience some flow separation.
Custom or modified cowls can reduce the separation and smooth the contour (if your engine installation permits!), reducing drag. Custom engine installations or changes to the firewall shape may require a custom cowl anyway.
'Boat-tail' cowls, in which the lower cowl does not sweep sharply upwards, have been shown to reduce drag.
The INSIDE of the cowl is also important in controlling and utilizing cooling airflow to maximize cooling without excessive drag. Extra vents or scoops will increase drag, and may or may not increase cooling.

Benefits = Reduced weight (partially offset heavier engine?), speed gains of as much as 15kts.
Risks = More expense (carbon is costly), may require several iterations and flow testing to
realize benefits.

Some of these aircraft fly with no spinner at all, with little apparent negative effect.
Small spinners that cover just the prop hub are sometimes used, their benefit appears to be purely aesthetic.
Due to the turbulent airflow behind the cowls, due to the big opening from the engine (and possibly separation on the cowls), any spinner within that flow is unlikely to provide much aerodynamic benefit.
Larger spinners, some with a flow guide extending into the cowl to the flywheel, may provide a small drag reduction.

Benefits = Ramp speed. Actual speed gains are most likely 0 - 1kt in most cases.
Risks = Reduced 
visial access to prop hub & bolts may hide potential problem if prop bolts lose torque.
Gear Modifications
Fuselage Modifications
Flying Surfaces Changes
Weight Reduction
Misc Modifications