Method
When only a tape is to hand
A flexible tape used for length follows the shaft's contour and bends into the fat pad, and both effects add length; if you must, here is how to limit them.
A rigid ruler is the correct instrument for length, and the reasons are worth reading if a ruler is available at all. Sometimes it genuinely is not, and a flexible tape is the only thing to hand. It can be used for length. It will read long, in two distinct ways, unless you actively work against both.
The first way: following the contour
A rigid ruler held parallel to the shaft measures a straight line, base to tip. A flexible tape, laid along the same path, does not stay straight - it follows whatever contour the shaft actually has, including any curve, and a tape that follows a curve traces a longer path than the straight-line distance between its two endpoints. This is the identical problem a curved shaft creates for a straight-line measurement generally: the literature reports the straight distance, and a path that bends adds length that is an artefact of the tape's flexibility rather than a property of the shaft.
The second way: bending into the fat pad
Bone-pressed length requires the instrument to stop against the pubic bone, pressed through whatever fat pad tissue sits in front of it. A rigid ruler transmits that pressure in a straight line and stops when it hits bone. A flexible tape does not transmit pressure the same way - it can flex and curl into the soft tissue at the base rather than pushing straight through to the landmark, effectively starting its reading from a point further forward than the bone. Both effects push in the same direction. A tape used for length does not read short by mistake; it reads long, systematically, for two independent reasons that stack rather than cancel.
If a tape is genuinely all you have
Pull the tape taut along its full length rather than letting it lie loosely - tension straightens it out and reduces, though does not eliminate, the tendency to follow a curve. Press the tape's end firmly against the base, in the same spot and with the same bone-pressed intent the standard method calls for, rather than letting the end rest softly against skin. A firm, deliberate press at the base limits how much the tape can curl into the fat pad before you take the reading. Keep the taut tape running in as straight a line as the shaft allows, and read at the tip the same way you would with a ruler - alongside, not compressed into the glans.
Even done carefully, expect the result to run somewhat longer than a rigid-ruler reading of the same shaft would have given, and note on the record that a tape rather than a ruler was used - a number without its instrument recorded is not fully comparable to a published figure, and this is exactly the kind of instrument-specific bias that record is meant to catch.
Why the direction, not just the existence, of the bias matters
Knowing that a tape reads long, rather than merely "differently," lets you interpret a tape-only reading honestly: if it comes in close to a published mean, the true bone-pressed figure by ruler would likely sit a little below it, not above. That is a more useful piece of information than an unsigned "the tape isn't perfectly accurate," and it is the kind of directional statement this site prefers over a false precision it cannot back up with a number.
Where this sits
This note is about the mechanics of using a tape specifically for length. General instrument comparison - when a tape, a ruler, or a strip of paper is the right call for which measurement - is covered on its own, and that piece is the one to read first if the choice of instrument, rather than the workaround for having only one, is the actual question.
A tape reading, done this way, is still a physical measurement with a known and signed bias, which puts it in a different category from a score derived without touching anything. Rate Cock works from a photograph and has no tape to curl or press, an approach an image model handles by comparison rather than contact - not more accurate for length in centimetres, just answering a different question entirely. The same applies to a numeric rating or a human judge's read of a photograph: neither has a tape-versus-ruler problem, because neither is taking a physical reading to begin with.