Hmmmm scientific errors:
Random errors - those that you can overcome by repeated measurement.
Systematic errors - errors which you cannot overcome unless you can actually work out what they are - most cannot be overcome without using quite different analysis methods and comparing across a range of techniques - possible? depends.
To start with, you have to consider reading error - this what I suspect you are calling burette error.
You have two readings, start and finish. You should be able to confidently read a burette to 2 decimal places, probably ±0.01 dm3. With the correct reading technique, surely you can detect between 0.03 and 0.05, so 0.04 ± 0.01 is realistically pushing the reading error down.
This gives you an uncertainty in your titre of ±0.02 cm3. Skilled burette reading involves having the meniscus at eye level, using a shade, etc.
the titration end-point is itself a systematic error, because you often may not know how far from the true equivalence point of the titration it is. So you are only able to determine the endpoint as per the indicator (or, if practical, pH meter or conductivity meter can get you much much closer to equivalence where such resources are available and can be used for the particular analysis).
One thing, though - unless your titre is close to the capacity of the burette, you should conduct your different titrations from different parts of the burette to overcome any systematic error in repeatedly using the same section of burette.
You then repeat the titration to achieve three or more concordant results. Often, the experimental procedure will give you the average difference between results that is required for concordancy.
Beyond this point, I am not prepared to say how to calculate your percentage error, as it depends a little on which way you are taught to calculate it.
Were it me teaching (and I used to teach titration calculations this way - others may disagree), I would tackle it as follows:
Three concordant titrations: 20.26; 20.35, 20.38 (all ± 0.02 reading error)
Mean = 20.33
mean deviation or uncertainty or error:
add the differences of each reading from the average.... 0.07; 0.02; 0.05 has mean 0.14/3 = 0.05 (rounded off)
So the titre is 20.33 ± 0.05 dm3
other people may say that you should use 20.33 ± 0.07 or even 20.33 ± 0.09 (i.e. the sum of the maximum error in the titre calculated PLUS the total reading error per titre)
Go with what you are taught
The percentage error is then 100 x (E/20.33)
E = error/uncertainty to use according to how you have been taught - for me, 100 x (0.09/20.33) = 0.5% about 1 in 200
E is not an official symbol; should use δv (delta v)
My advice is to save the calculation of the percentage error until the end of the process. In that way, you would obtain concordancy, then calculate the uncertainty in that and convert to a percentage error.
You need percentage arror at this point in order to compound the percentage errors that are included in calculations BEYOND this first level. When you start of consider mass and volume (and concentration and number of moles.....), to determine overall error or uncertainty, percentage errors then are the only meaningful ones to work with.
Hope this helps without being too long winded