Continental extension is accommodated by normal faults, and fault growth is achieved by increases in length (L) and displacement (D). When normal fault length and displacement were initially studied in the 1970’s and 80’s, it was assumed that faults maintain a constant D/L ratio from initiation to cessation, growing via sympathetic increases in displacement and length (the propagating fault model). Since the advent of 3D geophysical seismic data, it has been suggested that faults reach their maximum length before accruing significant displacement (the constant-length model). Several uncertainties remain, such as (i) how normal fault displacement and length relate to each other in different settings and whether a universal scaling law can be used, (ii) how different geological factors affect fault geometry and growth. To address these issues, I first assemble and interrogate a global database of normal faults that includes displacement and length, as well as tectonic setting, host rock lithology, fault maturity, and D/L through time where available. I next present a case study that uses growth faults imaged in 3D seismic data from offshore NW Australia to track fault lengthening, throw, and changes in slip rate through time. Observations on fault growth from the global database and seismic-based interpretations were then tested using physical analogue models, designed to investigate fault displacement/throw and length through time in higher resolution than seismic data allows. Results demonstrate that ‘one-size-fits-all’ D/L scaling laws are not accurate because the fault geometry and growth is affected by fault size, host rock lithology, fault reactivation, and fault maturity in particular. Universal scaling laws are inherently problematic because the relationship between length and displacement/throw is ever-changing throughout a fault’s life. Individual normal faults follow the following growth pattern: a lengthening stage (10-30% of a fault's active life), displacement stage (30-75% of a fault's active life), and possible tip retreat stage (final 25% of a faults’ active life). On a fault array scale, faults follow a cyclical growth pattern, where faults alternate between the lengthening and displacement stages.