I used the 2000 years mark if the word looked like a direct descendent of Proto-Daghestanian reconstruction;
1000 years mark was used for words that looked close to Lezgic words but different from the reconstructions for other language groups;
600 years (since Islamization) mark was used for the borrowings connected to religion
200 years was used for Avar borrowings since the close contacts with Avars are roughly this old
100 years mark was used for modern words borrowed from Russian in Soviet times.

clearly borrowed I used this when the word is mentioned as borrowed in Kibrik et al. (1977a), or in other sources (like Starostin’s database or my own knowledge) and I found the word in the dictionaries of the contact languages.
probably borrowed. I used this when I could not establish the source word in the following cases:
it is mentioned as borrowed in Kibrik et al. (1977a);
the phonetic structure of the word indicates it is probably not an original Archi word (like root resonants or a pharyngealised vowel in the root without an uvular);
there are no similar words in Lezgic languages.
perhaps borrowed was used for one word, ʁʷˤalqˤi ‘fog’ which does not look like a Lezgic word (Proto-Lezgian for fog is *k:om:ol). There may be a link to Proto-Dargwa *ɣerɣʷa, but it’s too speculative.
very little evidence for borrowing. This was mostly used for words which have very close equivalents in neighbouring languages, and it is impossible to decide whether they are borrowed or just common words (the neighbouring languages are from the same language family). In all other cases I wrote a comment on why I decided for this field.
no evidence for borrowing – clearly Lezgic or Daghestanian word (established using the North Caucasian etymology database at http://starling.rinet.ru/main.html)