The bottom one or two layers are visibly wider than the rest of the part. Run a straightedge along a vertical face and there is a small flare at the base; stand the part on a flat surface and it rocks, or a hole near the base will not accept a pin.
Distinguishing it takes seconds. If only the base is oversized, this is your page. If every dimension is oversized, that is over-extrusion and it wants flow calibration instead.
The bulge is the price of the bond
There is no fault here in the ordinary sense. A first layer is deliberately printed with less vertical clearance than the nominal layer height, because squashing molten plastic into the plate's texture is what makes it stick. That displaced material has to go somewhere, and sideways is the only direction available.
Two more things then work on it. The bed is holding the base above the material's softening point for the whole print, so the lowest layers never fully rigidify. And everything printed above presses down on them. A soft column under load spreads, slowly, for as long as the print continues — which is why a tall part shows more of a foot than a short one built from the same file.
Where opinion genuinely divides
Ask five people and you will hear "it is your Z offset" and "it is your bed temperature" with equal confidence. Both camps are describing real observations.
The offset case is easy to demonstrate: lower the nozzle further and the flare grows, on a cold plate as well as a hot one. The temperature case is just as easy: keep the offset identical, raise the bed by 10 °C, and a PLA part that was clean develops a foot. They are not competing explanations. Squash sets the initial width; heat and load decide how much it keeps spreading afterwards.
That matters practically, because it means a machine can have a correctly set offset and still produce this on a hot bed.
The fixes, best first
Chamfer the bottom edge in CAD. Half a millimetre of 45-degree chamfer, on the model, once. It is exact, it costs nothing at print time, and it never affects anything else. If the part is yours to edit, stop here.
Use the slicer's elephant-foot compensation. Typically 0.2 mm. Worth knowing what it actually does: it shrinks the outline of the lowest layers only, which is a distortion of your geometry rather than a correction of the printer. That is fine for a bracket and wrong for a part that mates precisely along its full height.
Drop the bed temperature once the first layers are down — but check the material first. On PLA and PETG, taking the bed down by 5 °C after the first few layers firms the base with no consequences. On ABS, ASA or polycarbonate, doing the same thing is how prints come off the plate at hour two, because those materials need the bed hot for the whole job to keep contraction stress under control. See print detaching mid-print for what that trade costs.
Raise the Z offset a fraction. Effective and the easiest to overdo: too far and you are into first layer too high, where the part does not stick at all. Move in single 0.02 mm steps and confirm the first layer still merges into a solid sheet.
Measure above it, not through it
A great deal of imaginary over-extrusion has been diagnosed by putting calipers across the widest part of a base flare. When checking a part's dimensions, measure at least a few millimetres up from the plate, and take the reading away from the seam as well. If the flare is the only thing out of tolerance, the machine's flow is probably fine.
When to ignore it
If the base is glued down, hidden, or gets faced off anyway, a small foot costs nothing. It is worth fixing when parts need to sit flat, when the bottom face carries a press fit, or when several parts stack — in which case the error accumulates once per interface.
Why a soft base flares and a rigid one does not
PLA, PETG and TPU, and for the same reason in each case: a low softening temperature relative to the bed underneath them. TPU adds its own contribution, since a soft elastomer under compression spreads more readily than a rigid polymer does. Materials printed on a bed far below their glass transition — the polyamides, polycarbonate — rarely show much of a foot at all, because their bases are genuinely solid within a few layers.