Six operations put a set of objects where they belong, and the order they run in isn't a matter of preference. Run them out of order and you get a slide that's technically aligned and still looks wrong. It's the most frustrating state to debug, because every individual check passes.
What are the six operations, in order?
- Set the span. Decide which columns the group occupies, and place the first and last object's outer edges on those column lines. This fixes the total width before anything inside it is negotiated.
- Equalise size. Give every object in the set the same width, and the same height where the set is a row. Objects that should read as peers must be the same size, because a size difference is a claim about rank.
- Align the leading edge. Put the left edges (or top edges, for a column) on their intended lines.
- Distribute. Spread the objects so the gaps between them are equal.
- Cross-align. Align the other axis: the tops of a row, the left edges of a stack.
- Lock or group. Freeze the result so the next drag doesn't undo it.
Advertisement
Why must you equalise size before you distribute?
Step 2 before step 4 is the rule people break. It's easy to break, because distribute feels like the obvious first move. Distribute computes equal gaps between whatever objects it's given. If those objects have different widths, you get equal gaps between unequal boxes, and the eye reads the width difference as deliberate emphasis you never intended. The four KPI cards on the opening slide are the case in point. Their widths are 2.30, 2.28, 2.35 and 2.30 inches, a standard deviation of 0.0299 inches and a range of 0.07 inches. Their gaps are 0.13, 0.25 and 0.22 inches, a standard deviation of 0.0624 inches and a range of 0.12 inches. That's 8.64 points of inconsistency between the widest gap and the narrowest. Their tops are 1.90, 1.88, 1.92 and 1.90, a range of 0.04 inches or 2.88 points.
Run distribute on that set without equalising the widths first, and the gaps become equal at 0.20 inches while the widths stay unequal. Now the slide has a defect that's harder to see and harder to fix, because the obvious check passes.
Rebuilt properly, the four cards are 2.933333 inches wide each, the three gaps are all 0.20 inches, and the set spans 12.333333 inches. The hand-placed version occupied 9.83. Nobody decided the rebuilt cards should be wider. Four three-column cards on this grid are simply that width, and the grid was the decision.
Should you align to the slide or to selected objects?
One more control deserves a name, because it silently causes about half of all alignment failures. It's the choice between aligning to the slide and aligning to the selected objects. Align-left with "align to slide" selected pushes every object to the slide's left edge and ignores your margin entirely. Align-left with "align to selected objects" pushes them to the leftmost object's edge, which is correct only if that object is already where it belongs. So check the setting before every align, or fix the leftmost object first and then align to selection. It takes a second, and it saves a lot of undoing.