=1*(DATEDIF(A1,NOW(),"y") & "." & DATEDIF(A1,NOW(),"ym"))

I don't think that will work correctly. Here's the first problem: the formula

gives the same fractional part for both 1 month and 10 months, i.e. 5y1m =

5.1, 5y10m=5.1

But correcting the formula to give 5.01 and 5.10 will not work either.

Let's say you have these time intervals: 11y11m, 13y11m, 5y11m, 4y9m. If you

generate the decimal numbers 11.11, 13.11, 5.11, and 4.09, the sum is 33.42,

and 33.42/4 = 8.355. What does that mean? If the terminology is consistent, it

should mean 8 years and 35.5 months, or 10 years and 11.5 months.

But if you add the years and months separately, you get 33 years and 42

months, which is equivalent to 33 years + 3 years + 6 months = 36.5 years, and

the average is 9.125 years, or 9 years and 1.5 months.

Some suggestions for solutions:

1) generate the elapsed time in decimal years

=DATEDIF(A1,NOW(),"y")+DATEDIF(A1,NOW(),"ym")/12

2) generate the intervals in months, i.e. =DATEDIF(A1,NOW(),"m"), average the

months, then convert the average to years and months with =INT(X/12) and

MOD(X,12)

3) generate the interval as

=DATEDIF(A1,NOW(),"y")+DATEDIF(A1,NOW(),"ym")/100

(note the above gives 5.01 for 5y1m and 5.10 for 5y10m)

Then use the DOLLARDE function (from the ATP) to translate the numbers to

decimal years so they can be averaged correctly, and the DOLLARFR function to

convert the result back to the yy.mm format.