Hypothesis test binomial distribution in Excel I have a list of invoices and a list of payments. (both with order ID)
I want to test if the amount of unpaid invoices is significantly less than 10% using binomial test in Excel 
How do I do this?
 A: Suppose your spreadsheet indicates you have $X$ unpaid invoices out of $n,$ where $n > 100$ or so. Then a normal test should be fine. Formally, you want to test $H_0: p = 0.1$ against $H_a: p < 0.1.$ (Some authors would say $H_0 \ge .1,$ but
in either case the 'hypothetical' value of $p$ is $p_0 = 0.1.)$
Let $\hat p = X/n$ be the estimate of $p.$ You want to reject $H_0$ if $\hat p$ is sufficiently far below $0.1.$ The normal test uses the statistic
$$Z = \frac{\hat p - p_0}{\sqrt{p_0(1-p_0)/n}}.$$
You would reject $H_0$ in favor of $H_a$ at the 5% level of significance if
$Z < -1.645.$ 
If Excel has a built-in procedure for this test, it may give you a p-value.
In that case, reject at the 5% level if the p-value is less than 0.05.

Here is output from such a test using Minitab statistical software. The data
are 9 unpaid invoices out of 135.
 Test and CI for One Proportion 

 Test of p = 0.1 vs p < 0.1

 Sample  X    N  Sample p  95% Upper Bound  Z-Value  P-Value
 1       9  135  0.066667         0.101980    -1.29    0.098

 Using the normal approximation.

Here the $Z$ statistic is $-1.29 > -1.645$ (and the p-value is $0.098 > 0.05),$
so you would not reject $H_0.$ The estimate $\hat p = 0.67$ is smaller
the $0.1,$ but not enough smaller to be 'statistically significant'.
(The interpretation of the one-sided confidence interval is that the
true value of $p$ could be as large as 10.2%.) 

In the sketch below the bars are probabilities from $\mathsf{Binom}(135, .1),$
which is called the 'null distribution'. The blue curve is the best-fitting
normal density curve. The p-value given above is the area under the normal
curve to the left of the dotted red line. (The exact p-value is the
total heights of the binomial bars to the left of the red line.)

