Zeros can be used as (insignificant) place holders to the left of significant digits if the number is a decimal. For example, a mass of 42 g has two significant digits. Expressed in kilograms, the mass of 0.042kg should still have two significant digits. The zeros to the left of the 4 are place holders and not significant:
 

If the last significant figure(s) is (are) zero, life gets a whole lot more complicated. Suppose I used a metric ruler with millimeter markings to measure the width of a skateboard. The skateboard, a precision model, measures exactly 20 centimeters, and I report the width as 20cm 0.02mm. The uncertainty in this number is in the fourth digit (the hundredth's place), so it has 4 significant digits, not just 1. If I write out the 4 sig-digs, the width is 20.00cm. Since the zeros to the right of the decimal place are not necessary as place-holders, their inclusion indicates they are significant.
 

If the last significant digit is zero and to the left of the decimal place, we have a problem. Our standard way of writing numbers has no real way of differentiating between 4500 to 2 significant digits, and 4500 to 3 or 4 significant digits.  Let's go back to and look at the earlier example when you counted football fans.  You might want a better count than your rough estimate, so you could decide to count filled rows in the stadium.   Each row holds 10 people, and most rows were packed.  When rows weren't completely filled, you could combine them and find the equivalent number of filled rows.  When you were done, you found 450 filled rows.  Your counting might have overlooked a few empty seats, but you are fairly sure the number of people in the stadium is between 4490 and 4510.  So when you report 4500 people attended the game, you really have three significant figures.  If you had counted individual people as they entered the stadium and found that exactly 4500 people attended, you might also report the attendance as 4500.  We need a way to distinguish these three cases apart.

Enter scientific notation. Scientific notation, sometimes called exponential notation, allows us to express all significant zeros to the right of the decimal place by multiplying by factors of ten. For example, we could write the attendance counted at the gate as 4.500 x 10². This has four significant digits, as it should. For the row-counting scenario with three significant digits, the attendance is 4.50 x 10².  And the rough estimate could be reported as either 4.5 x 10² or as 4500.

Scientific notation will be used to indicate significant zeros, so zeros at the end of a number and to the left of the decimal place will be considered insignificant place holders. This convention, however, is not uniformly followed, so you must be careful using it.