We propose a novel approach to designing RDF-stores with the goal of improving the consistency and predictability of query performance. When designing these systems, three properties are commonly desired: support for the full range of SPARQL query features (Q), support for widely varying RDF datasets in terms of structuredness and size (S), and high performance (P). We develop the empirical SPQ conjecture which states that it may be impossible to achieve all the three desiderata simultaneously. We present a strong case for its plausibility based on our experimental results. The tradeoffs among the three and design guidelines based on the SPQ conjecture are also discussed.