Dealflow is a memory problem before it is a process problem
Most of what a fund sees will be a pass, and passes still matter
A fund that looks at a thousand companies a year will invest in a handful. If the other companies simply vanish from the firm's memory, the fund has thrown away its most valuable asset: an informed view of a market and a set of founders who already know it. The pass is not the end of a relationship. It is usually the start of one.
This is why a revisit date matters more in venture than a close date. A company passed on because the metrics were early should surface again in nine months with the original notes attached, so the partner who passed makes an informed call rather than an embarrassed one. Almost every fund intends to do this. Very few have a mechanism that makes it happen.
Sourcing quality is measurable, and it is rarely what people assume
Ask a partner where the best deals come from and you get an anecdote. Tag a referrer on every company for a year and you get something better: introductions made, how many reached a first call, and how many reached committee. In most funds a handful of portfolio founders and co-investors account for most of the quality, and they are rarely the sources getting the most attention.