You should inbreed, rather than back cross. Inbreeding preserves recessive traits from your photoperiod parent better than if you were to backcross to the autoflower. Also, you start to see autoflowers at the F2, so your journey is a lot closer to being finished than you think! And backcrossing... many marijuana "breeders" have propagated this bad advice for years - the only situation it is really useful is if you are trying to recreate a clone only strain. However, it leads to badly inbred lines because both alleles for a given gene only come from one parent - opening up many genetic deformities. Also, bad "breeders" have seen success with this method because they have to search less for recessive traits. I actually saw an old "breeding" article from the 90s where a respected guy said that "if you don't see a trait come up in the first generation then it isn't worth breeding for". I think he was talking about blueberry smell or something, but he showed a lack of basic Punnet Square knowledge - and due to his ethos, the bad advice was never challenged.
Anyways, there is one benefit to back crossing - you will see more autoflowers in your F2 than if you were inbreeding. If you backcross you will see 50% auto in the F2, but you will be 3/4 the original auto and 25% the photoperiod. This makes for smaller gains in aromas and potency with each generation, and means you have to look harder and harder to find the most potent.
If you INBREED them, which means a male and female (brother and sister) of your F1 you are growing now, you will have 25% auto at the F2, BUT you're still half and half between the two parents. So lets say you're crossing a 12% auto to a 20% photoperiod. The potency percentage in the offspring will show up as a bell curve distribution - with most of the offspring having a median THC content(12 and 20 are the extremes, so the median is 16%), and a few offspring that are super potent and a few offspring that are schwag.
If you backcross to the autoflower parent (12%) with one of your F1s (median THC content 16% - could find one more potent, but the further from the median number, the less frequent they appear), you now only have a range of possible potencies from 12% to 16%, with most of your offspring at 14%, and a few extremes that are higher or lower. Because of this, there is no reason to back cross to the less potent parent, ever.
The best way to go about the situation (for potency and flavors) is to inbreed your F1 to make F2. Only 25% of the F2 will autoflower, but you will see definite increases in potency and aroma by the F3 compared to if you were to backcross to the autoflower parent. Of those F2 that autoflower, breed a male and a female together and by F3 you will be fully autoflower. You will still have to stabilize plant structure, but at this point you're more potent than if you had taken an F1, backcrossed it to an auto, and bred the resulting autoflower offspring. Both scenarios take 3 generations, but the difference being that the backcrossing shortcut will give you 50% auto compared to 25% auto at the F2, but inbreeding gives you more potency and aroma, even though you see 25% autoflower at F2.
NOW, once you have a fully autoflower male from the F2 - branch the line off. Breed two fully autos from the F2 together to have a fully auto F3 but ALSO use an autoflowering male to backcross to the photoperiod female. You would then have a 3/4 cheese and 1/4 auto. If your F2 male was average at 16% (based on the hypothetical numbers above) and your photoperiod was 20%, this line of seeds would have a median potency of 18%, but you would have to inbreed to F2 to see any autoflower.