Two main issues often arise in collections. The first is coherence. Simply put, this refers to how well the essays fit together. Pieces in a collection have to be more than simply stand-alone articles. When combined, they should be greater than the sum of their parts. What this means is that you as editor have to exercise a very firm editorial hand. Essays that don't fit must be revised so that they do or cut completely. You also need to pay careful attention to how the chapters are ordered, and those essays that are complementary should be placed close together. You should examine them all to see if they can be broken into groupings, or thematic sections, which might correspond to parts in the book. The best collections show a logic, both in the presence and the order of the articles.
Another thing you will need to do is write a good introduction. This should address the overall themes being dealt with in the book, as well as how all the chapters are linked to these themes. It will provide a rationale and a framework. Avoid writing a summary of all the chapters as a separate section and instead build these chapter descriptions into your discussion of the themes and issues, methods and arguments that ground the volume.
You should encourage the chapter authors as much as possible to make reference to the other works in the collection (e.g., "As John Brown notes in his chapter earlier in this volume, etc... My own approach will differ as I examine X from the perspective of Y and Z"). A general conclusion to the volume, which draws together the various chapters, and suggests direction for future research, is also strongly recommended.
Coherence is enhanced when the chapters all employ the same type of referencing. You would be surprised how often editors submit manuscripts in which some chapters have endnotes and other have footnotes, some have a humanities style, and others a social science style. Please make sure to follow the Chicago Manual of Style, 16th edition. You can also consult UBC Press's website documentation samples.
The second main issue concerns quality. The essays should all be of the same basic calibre. Weak ones should be strengthened or dropped completely. Uneven quality is one of the most common complaints levelled against edited collections in peer review, and it is always best to try to deal with this issue in advance so that readers don't have to comment on it. Again, a firm editorial hand will be required here.