The Large Hadron Collider will be the first accelerator to directly and fully explore the Fermi energy scale, thereby offering the opportunity of a real understanding of its dynamics. The relevance of this task is only comparable with the challenge of extracting new information from the LHC data and of finding the correct signal interpretation out of many theoretical possibilities. In fact, existing data already provide important hints on likely and less likely new physics models and constrain many appealing cases whose LHC signals would clearly stand out; a more realistic scenario is that discoveries will rely on the ability to trigger on the right quantities out of a complicated background. This highlights the possibility of a complementary role between existing data and the LHC ones in the search for new physics effects. Specifically, the former data may be used to literally guide the definition of the most promising LHC search channels. This of course requires the prior definition of a calculable theory framework, so that it can actually be compared with data themselves. I will consider certain classes of models and study whether applying the existing constraints allows indeed to identify the LHC channels where these models have best chances to be unveiled. This program proceeds through the following main questions: a) how loose constraints on a given set of theory parameters may be and which channels may be most sensitive to them; b) what level of background is to be expected in the search channels found by point a; c) what is the `discriminating' power that these search channels have on the theory under consideration, i.e. the chance that a positive signal may be due to something else than that theory. Crucial is the choice of the considered classes of models. I will follow two parallel strategies, that should be complementary in the balance between generality and predictivity, and that should span a large set of `representative' theories.