For every 1000 estimated sexual assaults (defined broadly), only 370 victims report the crime to police. From those reports, only six perpetrators ever do jail time. Presumably some get probation or other penalties. Let's assume a total of 10 people get penalized, probably representing a much larger portion of sexual assaults than we might guess--those who rape, rape repeatedly. So, apart from wrongful convictions, we might guess that, say, 50 of 1000 victims actually get some degree of justice, although they might not know it.
However, if we have a 2%-8% false reporting rate, we would presume that in addition to 370
That leaves 360 accusations, plus or minus, that are in the "insufficient evidence to convict" category. Some are genuine crimes, some are mistaken identity (this is my best guess on the Ford/Kavanaugh controversy now), and some are flat out false accusations where the police can't quite prove that they are false.
Part of this makes sense; sexual assault almost by definition occurs without extra witnesses, and delays in reporting prevent physical evidence from being collected and impair the accuracy of reports. The huge prevalence of intoxicant abuse among both perpetrators and victims of sexual assault takes a toll as well.
But with that known, let's understand what that 2-8% false accusation rate really means. Again, it means that false accusations are as common as prison sentences for sexual assault, and if we can't improve that ratio, people can and will lose confidence in the system for punishing these crimes.