We clarify inflaton models by considering them as effective field theories in the Ginzburg-Landau spirit.In this new approach, the precise form of the inflationary potential is constructed from the present WMAP data, and a useful scheme is prepared to confront with the forthcoming data. In this approach, the WMAP statement excluding the pure phi^4 potential implies the presence of an inflaton mass term at the scale m sim 10^{13}GeV. Chaotic, new and hybrid inflation is studied in an unified way. In all cases the inflaton potential takes the form V(phi) = m^2 M_{Pl}^2 v(phi/M_{Pl}), where all coefficients in the polynomial v(x) are of order one. If such potential corresponds to super symmetry breaking, the susy breaking scale is sqrt{m M_{Pl}} \sim 10^{16}GeV which turns to coincide with the GUT scale. The inflaton mass is therefore given by a see-saw formula m sim M_{GUT}^2/M_{Pl}. The observables turn to be two valued functions: one branch corresponds to new inflation and the other to chaotic inflation,the branch point being the pure quadratic potential.For red tilt spectrum, the potential which fits the best the present data and which best prepares the way for the forthcoming data is a trinomial polynomial withnegative quadratic term (new inflation).For blue tilt spectrum, hybrid inflation turns to be the best choice. In both cases we find a formula relating the inflaton mass with the ratio r of tensor/scalar perturbations and the spectral index ns of scalar perturbations: 10^6 m/M_{Pl}= 127 sqrt{r|1-n_s|} ;(the coefficient 127 follows from the WMAP amplitude.Implications for string theory are discussed.