Molding around a metal insert puts two materials with nothing in common into one part. Five things go wrong, and all five are decided in the tool rather than at the press.
A finished component — usually metal — is placed into the tool before the shot, and plastic is molded around it. The insert was made somewhere else, by some other process, and it arrives complete.
A previously molded plastic part becomes the substrate for a second shot, either in a two-shot tool or by transferring it into a second mold. Both layers are molded, one after the other.
They are confused constantly, and the practical difference is where the risk sits. Insert molding puts the risk on holding a foreign object still against injection pressure and on two materials that shrink at different rates. Overmolding puts it on bonding one polymer to another and on the first shot surviving the second.
Melt arriving at speed pushes on whatever it meets. An insert that is not positively located and clamped will shift, and at small scale the shift that matters is smaller than the eye catches — a pin leaning a fraction of a degree puts its far end out of position. Location is a tooling problem, solved in steel: a pocket that traps the insert on more than one axis, and a core pin that holds it against the flow rather than beside it.
That mismatch is the whole mechanical relationship. A little of it is what grips the insert; too much of it is a hoop stress locked into the plastic around a rigid core, waiting for a thermal cycle or a solvent to turn it into a crack. Wall thickness around the insert is the lever, and resins with high or anisotropic shrink need more of it than the datasheet number alone suggests.
Flow splits at the insert and rejoins on the far side, and where it rejoins the polymer chains are not entangled the way they are elsewhere. That seam is weaker than the bulk material by a margin that depends on melt temperature, on how far the two fronts travelled before meeting, and heavily on whether the resin is filled — fibres do not cross a knit line. Gate position decides where that seam lands, so it decides whether the seam sits on a loaded face.
Metal conducts heat away far faster than steel tooling at mold temperature does, so the skin against a room-temperature insert solidifies early. On a thick wall that shows up as a cosmetic line. On a thin one it can stop the fill reaching past the insert at all, which reads as a short shot and gets chased with pressure that will not fix it. Pre-heating the insert is the usual answer, and it is a cycle-time cost that belongs in the quote rather than in a surprise.
Plastic does not bond to clean metal in any way worth relying on. What holds an insert is mechanical: a knurl, an undercut, a groove, a flat that stops rotation. Pull-out and torque-out are different loads needing different features, and a knurl that resists twisting may do nothing against a straight pull. Decide which load the part actually sees before choosing the feature.
Wondering how much wall your insert needs, or where the knit line will land? Send the part. If the resin is the question rather than the geometry, the shrinkage calculator shows what its shrink range does to a dimension before anything else does.