Containment is a strategic option that is frequently advocated for dealing with invasive plants. It is often presented as the fall-back option when eradication is deemed unfeasible, notably when eradication attempts are abandoned. However, containment confronts the same needs for detection, delimitation and destruction of plants as eradication. Its main advantage is that the area to be managed is smaller. Its main disadvantage is that the time over which management is required is infinitely longer, assuming that eradication is successful and the containment effort is not abandoned. We argue that a containment program should be built around clearly defined containment units, consisting of an occupied zone and a surrounding buffer zone, at a scale that aligns with the plant's dispersal capacity. There will always be a probability >0 that some propagules will be dispersed beyond any practical buffer zone designed to cover the seed shadow of plants occupying the containment unit. This requires that, under a containment strategy, some resources should be invested in both the occupied zone, to reduce propagule pressure, and the area beyond the buffer zone to deal with the consequences of longdistance dispersal events. Containment is not always easier than eradication or the most cost effective alternative to it.