2 audio sources @ Near .3m front & rear of mics.
https://drive.google.com/open?id=1WiA0Z8hPj1bajv3PbMV6V9IYD2dwdqEB
https://drive.google.com/open?id=1c5P-MRQvTHuTyNDwx3HCMgX6EA8YbP4k
Forward and reverse are at same volume and distance just but the sound sources just swapped for comparison.
interesting comparison 
Yeah it hard to tell if you don’t have a comparison as you never know the original level.
Any undirectional effects are mostly in the near field but that has nothing to do with the microphone. The further away you are the more sound will reflect and cause echo so that the difference of the soundwaves reaching front and back becomes much less as does the sound pressure.
The £2.50 unidirectional electret has a number of uses but firstly in confined cases with audio like the telephone handsets where often used, they can greatly reduce echo and become part of a attenuation process.
You can use a unidirectional in conjunction with Speex AEC and get some extremely good echo attenuation but unlike the digital method of SpeexDSP a unidirectional doesn’t need a reference signal its purely what is facing the back of the mics cardioid polar pattern.
The electrets I used are just in open air on jumper wires there are likely other versions with stronger polar patterns such as super cardioid as they have existing for decades and are very old technology.
I have been wondering if one of the 3D printer users might actually try to clone what is a common unidirectional microphone body (the usual series of perpendicular slots to the mic face).
Unidirectional only attenuate rear from front in a cardioid pattern but with positioning that is common they can greatly help overcome noise.
Because they are so relatively cheap it is possible to have multiples so that at least one has reasonable near voice and far noise.
You can increase that chance by having multiple mics of differing orientation 180, 135 or 90 degrees and using sperate instances of KWS to use the highest KW confidence hit to give finer resolution to your beamforming.
There is a huge gulf to any effective beamforming echo & noise suppression until you hit the cost levels of silicon DSP as omnidirectional mics without processing are as there name suggests.
Omni directional mems due to their tight manufacture tolerances work extremely well in arrays with DSP silicon but on their own without algs there is nothing you can do to effect noise & echo.
Cloning a microphone body for 3D-printing sounds like a job for someone with 3D-printing and audio skills. Maybe something for @JGKK? 