The ongoing optical time-domain astronomy surveys are routinely reporting fifty transient candidates per night. Here, I investigate the demographics of astronomical transients and supernova classifications reported to the Transient Name Server in the year 2019. I find that only a tenth of the transients were spectrally classified. This severe "bottleneck" problem should concern astronomers and also funding agencies. The bottleneck will get worse by a factor of 20 (or more) once LSST comes on line. We need to fundamentally rethink the purpose of surveys for transients. Here, after undertaking a detailed investigation of this issue I offer some solutions. Going forward, astronomers will employ two different methodologies: (1) multi-band photometric method which is well suited to the study of very large, many tens of thousands, samples of faint transients; (2) spectral classifications of thousands of bright transients found in shallow and nightly cadenced wide-field photometry surveys and transients associated with galaxies in the local Universe. The latter program, in addition to unearthing new types of transients and offering astronomers opportunities to undertake extensive follow up of interesting transients, is needed to set the stage for the former. Specifically, I suggest a globally coordinated effort to spectrally classify a complete sample of bright supernovae ( 19.5 mag) and transients within the local Universe (< 200 Mpc). The proposed program is within reach -thanks to the on-going widefield surveys, the development of novel spectrographs tuned for classification, great improvements in throughput of spectrographs and the increasing availability of robotic telescopes.