Thanks for all the input. I guess I'm not being clear that (and this is 3rd hand info) the ONLY thing changed from the same supplier was thinner nickel and below the IPC limit of 3um, which stopped the cracking . I'm reluctant to agree to an ENIG process outside the IPC-4552 nickel thickness window without understanding the mechanism of why a thinner nickel over copper might make the tracks more forgiving/ductile.From what has been said so far I get the impression that no-one else has decided that reducing the nickel on an ENIG flexi is helpful in improving its ability to be bent.  

Like others have said though, there are other ENIG flexi/flexi-rigid with zif ends around and don't see a problem - some with stiffener some without. The finished circuits do get formed quite tightly in the product and this isn't normally a major problem.

What I was also keen to know was if we did allow this what impact would we expect if we, say, spec'd the low end of the IPC range (ie to be typically only 3um) to have on other performance characteristics, eg solderability, compared with the standard 3-6um range allowed.