Einstein's equations were derived for a free massless spin-$2$ field using universal coupling in the 1950-70s by various authors; total stress-energy including gravity's served as a source for linear free field equations. A massive variant was likewise derived in the late 1960s by Freund, Maheshwari and Schonberg, and thought to be unique. How broad is universal coupling? In the last decade four $1$-parameter families of massive spin-$2$ theories (contravariant, covariant, tetrad, and cotetrad of almost any density weights) have been derived using universal coupling. The (co)tetrad derivations included 2 of the 3 pure spin-$2$ theories due to de Rham, Gabadadze, and Tolley; those two theories first appeared in the $2$-parameter Ogievetsky-Polubarinov family (1965), which developed the symmetric square root of the metric as a nonlinear group realization. One of the two theories was identified as pure spin-$2$ by Maheshwari in 1971-2, thus evading the Boulware-Deser-Tyutin-Fradkin ghost by the time it was announced. Unlike the previous 4 families, this paper permits nonlinear field redefinitions to build the effective metric. By not insisting in advance on knowing the observable significance of the graviton potential to all orders, one finds that an arbitrary graviton mass term can be derived using universal coupling. The arbitrariness of a universally coupled mass/self-interaction term contrasts sharply with the uniqueness of the Einstein kinetic term. One might have hoped to use universal coupling as a tie-breaking criterion for choosing among theories that are equally satisfactory on more crucial grounds (such as lacking ghosts and having a smooth massless limit). But the ubiquity of universal coupling implies that the criterion doesn't favor any particular theories among those with the Einstein kinetic term.