I is some testing regarding this feature on strelok with 2 factory ammo in last 2 years.
First of all it works very well IF YOU INPUT CORRECT DATA.
This app is doing an linear interpolation, for this needs at least 2 valid / verified values to interpolate in between.
My advice is not to use any arbitrary % or out of internet factor.
Is well known the fact that powders are not affected by temperature in a linear way, from 0 C/ 32 F to 10 C/ 50 F you can have a very stable temp with low muzzle velocity change but over 15 C / 59 F to have some spikes also over 25 C / 77 f to have another jump, this is the reason you need valid data for all temp range you are shooting in.
If you use a data book and record muzzle velocity or drops in different environmental conditions is easy to get a muzzle velocity for a specific temp.
Ideally is to have a measurement at each 5 C / 40 F change work between your temp extremes an input in strelok only this who make sense.
For example if you get
850 m/s at 0C
853 m/s at 5C
859 m/s at 10 C
870 m/s at 15 C
884 m/s at 20 C
898 m/s at 25 C
928 m/s at 30 C
931 m/s at 35 C
notice this spikes, jumps in speed use them in Strelok as references, intermediary data will be close enough, is no need to input 850 at 0 C also 857 at 10 C because if you use 850 at 0 C and 870 at 15 C and you interpolate for 10 C (strelok will do this for you) you will get 863 m/s close enough from 859 m/s (what you got in the field), closer the measured values, closer to reality results.
From my testing I got to the conclusion this app needs at least 4 known values to be able to get correct results, maybe with stable powders needs less.
Mine is like this, yes I know very unstable powder, is Lapua factory ammo.