A specified value cannot be infinitely precise and has some inaccuracy, be it a voltage, current, resistance, capacitance, inductance, temperature or anything else. The inaccuracy on a value, the specification of its deviation from the intended value, is usually specified as a tolerance and may be given in absolute units (+/-1 V, +2 V/-3 V etc.) or as a percentage (+/-2%, +10%/-6% etc.).
Looking for tolerances on every value and using them is the normal view in electronics engineering and many mistakes are made when they're disregarded or ignored. For example, the UK mains voltage is currently specified at 230 Vac +10%/-6%, so a load plugged into it must be designed to operate continuously and successfully from any voltage between 216 Vac and 253 Vac. (I still think to 240 Vac +10% = 264 Vac but that's just me being old!) Designing equipment around just 230 Vac could lead to nasty, smouldering results. Another is designing an RC oscillator, where the oscillator frequency must be calculated for highR and highC and also for lowR and lowC (R and C at their highest and lowest values from their tolerances), to obtain the max/min frequency.
If it sounds like ongoing aggravation then just see it as part of engineering life :-) It's really that the view of every value as precise ("a 10K resistor is 10,000 ohms") is fake and leads to mistakes, problems and unreliable equipment.
Returning to your specific case, you can obtain a supply voltage tolerance from your supply load device's manufacturer. If the supply you have, your battery, can't meet this, you could consider using a +15 V to +15 V DC-DC converter. Look at the specifications for these, you should find one with a wide input tolerance for your battery and a tighter output tolerance for your load. They're not particularly cheap, though, and they introduce switching noise that you must design for or design out.