Highlighting achievements is a good thing. The conference can signal which papers is considers very good, authors can take deserved credit. ECRTS moved from Best paper only to Outstanding papers to serve these goals better.
In my experience, there is rarely a single best paper, clearly above all others. Typically there is a small group of very good papers, the decision which is the best is often a consequence of the criteria used, rather than authors’ achievement only.
Candidates for best papers are typically nominated during the evaluation process, the actual decision often taken at the conference, i.e., after the proceedings have been completed.
At ECRTS, we introduced Outstanding Papers, a small number of very good papers, decided upon at the TPC meeting. This allows including them in the proceedings, make the award visible with the papers after the conference. So, more authors get credit with higher visibility.
For the actual best paper, we formed a very small committee to decide during the conference, taking presentations into account. It included technical program committee chair for knowledge of papers, technical committee chair as best paper also represent the conference.
We put the outstanding papers in a special session of the program for added visibility (and put it at the end of the conference to counter attendance erosion). Make the awards visible via conference web site etc, an easy way to support authors and community.