Each number in scientific notation has two parts that can be handled separately. To multiply (3×104)(2×103), regroup: (3×2)×(104×103)=6×107. The coefficients multiply as ordinary numbers, and the powers of ten follow the product rule, so the exponents add. Division mirrors it: 4×1028×106=48×106−2=2×104.
Sometimes the coefficient comes out too big or too small. (5×103)(4×102)=20×105, which has the right value but is not standard, since 20 is not less than 10. Rewrite 20=2.0×101, so the answer is 2.0×106. Every place the coefficient's decimal point moves is balanced by changing the exponent by 1 in the opposite direction.
Addition looks similar but works differently, and that is the trap. In ordinary numbers, 4×103+5×104 is 4,000+50,000=54,000. Adding coefficients and exponents would give 9×107, which is 90 million. To add, first make the powers match, 4×103=0.4×104, then add the coefficients, 0.4+5=5.4, and keep 104. The power of ten acts like a unit, and you can only add amounts measured in the same unit.
Think of it like this. Adding 4×103 to 5×104 is like adding 4 dimes to 5 dollars. You can't just say 9 of something; convert first. 4 dimes is $0.40, so the total is $5.40, just as the sum is 5.4×104. Coins only explain adding, though: when multiplying, the powers of ten combine on their own.
Core idea. Combine the number parts with the matching operation while the powers of ten follow the exponent rules.