R2 Institution, more like (though not quite) R1-ish department:
Publications: Minimum of 2-3 per year, but field and quality considered - for example, a developmental person might be fine with 2, or 1-2 per year, but social might need solid 2-3, if not 3-4 per year. If all papers are in high impact factor journals, then a solid 2 per year, if not 1-2 per year, would be just fine.
Impact: At least some publications in higher-impact journals (i.e., minimum IF over 2-3), and evidence that work is being cited at reasonable rate (in particular, work that has been out for 1+ years).
Independence: Some publications from data being collected on-site in lab, or in "field" (e.g., homes, classrooms, etc.) but reflect new data since starting (i.e., can't all be from existing data sets with doctoral or post-doc mentors), but collaboration with new and previous collaborators fine, if not encouraged in some areas. Likewise, publishing from existing data sets is just fine (i.e., faculty appreciate that good data should be published, it just can't be the only thing going for a candidate).
Grants: Minimum of 1 submission, though more encouraged (if not strongly). Faculty realize that the grant climate has become much more competitive than in the past, which is why pre-tenure faculty are not required to obtain a grant. But, funded applications could outweigh slightly less publication productivity, and applications that are scored, but not funded, are taken as a positive sign of future potential. Of those that have gone up recently, 4 out of 6 had one or more funded grants, and it may have been 5/6 or 6/6 (only know about 4 for sure off top of head), and at least half had submitted multiple applications either as PI, Co-PI, or Co-I. NSF and NIH grants are valued slightly more than foundation funding, but not enough to affect a decision.
Conference presentations: talks are valued as they tend to be more competitive, and all are valued in the sense that it demonstrates that one is getting their work out there. Thus, no presentations would be point of discussion, but 1-2 year (or more) would be fine, but could not outweigh previous factors
Invited talks: valued, as this is taken as evidence of visibility and impact, but probably wouldn't affect a decision if they were absent
Teaching: load is 2/2. Above factors reflecting scholarship/impact are more important, but mean ratings should range between a 1 and 2.99 (1 and 2 are "excellent" and "good" categories, respectively). A few courses that are "blips" in terms of lower ratings wouldn't even be mentioned; consistently poor ratings wouldn't outweigh good/excellent research/scholarship, but could affect a decision if scholarship is so/so.
Supervision: want to see graduate students making reasonable progress (if not, will usually be able to distinguish if it is a student or mentor issue), and graduate and undergraduate students involved in research activities. If candidate had consistent trouble recruiting graduate students, this could be a point of discussion (but, not sure how much this would affect decision).
Service: 1-2 department level service activities in the last couple of years generally expected, but none in the first 3 years. Highly unlikely service would affect a decision.
External Evaluations: Valued, and could affect a decision one way or another if other factors place a candidate more "on the fence". Otherwise, probably would not affect a decision.
Hope this helps, but keep in mind that at least here, while these are the "hard" items considered, each candidate is unique, and variations across these categories, as long as in range, would still result in same decision. For example, it would be possible for someone to receive tenure here if they had 1 pub. per year, with stellar other factors, though this would probably generate quite a bit of discussion.