Round a number to N significant figures (scientific and lab precision)

Why fixed decimal places fail significant-figure rules

ROUND(x, 2) rounds to two decimal places regardless of scale: 12,847 stays numerically 12,847, while 0.004567 becomes zero. ROUND does not itself set the displayed number of decimal places.

INT(LOG10(ABS(x))) finds the order-of-magnitude power of 10 (4 for 12,847; -3 for 0.004567). Passing SigFigs - 1 - INT(LOG10(ABS(x))) as ROUND's second argument rounds 12,847 at -2 (hundreds -> 12,800) and 0.004567 at +4 (0.0046).

Zero-safe N significant figures formula

Always guard A2 = 0 first because LOG10(0) is mathematically undefined (#NUM!).

=IF(A2=0, 0, ROUND(A2, B2 - 1 - INT(LOG10(ABS(A2)))))

Why INT beats TRUNC for sub-1 decimals

For 0.004567, LOG10 is -2.34: INT floors to -3 (giving 4 decimal places), whereas TRUNC would chop to -2 and lose a significant digit.

=INT(LOG10(ABS(A2)))

Inputs and interpretation

Require a positive integer significance count within the spreadsheet’s numeric precision (typically no more than 15 decimal digits). This rounds a number; it does not preserve trailing-zero significance in its display. Use appropriate scientific formatting for values such as 12,800. The diagnostic LOG10 expression is undefined at zero; the main formula guards that case. Very large/small numbers and powers-of-ten boundaries remain subject to floating-point limits.

Further reading

• Microsoft Support — LOG10 function