Why a burndown alone misleads
A burndown that flattens looks identical whether the team stopped delivering or someone added work. Those need opposite responses.
The ambiguity at the heart of a burndown
A burndown plots work remaining. When the line flattens there are exactly two explanations: the team stopped finishing things, or someone added work. The chart shows neither.
Those call for completely different responses — one is a delivery conversation, the other is a scope conversation — and the chart most teams look at every day cannot tell them apart.
What a burnup adds
A burnup plots work completed as a rising line, and total scope as a separate line above it. Delivery and scope become two visible things rather than one ambiguous one.
When the ceiling moves, you can see that it moved. The conversation becomes about the change rather than about whether the team is trying hard enough, which is both more accurate and considerably less demoralising.
Never rebase the ideal line
When scope grows, it is tempting to redraw the ideal line from the new total. Do not. That makes a growing project look like it is tracking perfectly, which defeats the entire purpose of drawing a reference line.
Added work should make the chart look worse, because the situation is worse. A plan that absorbs unlimited scope while continuing to report itself as on track is not a plan.
Scope reduction is not progress
The mirror image is worth naming. Removing work makes a burndown drop, and it looks exactly like delivery. Nothing was built.
On a burnup the scope line comes down to meet the delivery line, which is honest and much less flattering. That is generally a point in the burnup's favour.
Neither is a management tool
Both are information radiators for the team doing the work. Used as a performance measure they stop being honest, because the easiest way to improve a burndown is to change what gets counted — and that will happen, quietly, within about two sprints of anyone being judged on it.