Both views read the same five stages and the same stale rule: a deal is stale when it has not changed stage within that stage's threshold — 10 days at initial offer, 5 at accepted, 3 at subject removal, 2 at deposit, 7 at completion — because a subject-removal deal sitting three days is a bigger problem than an offer sitting a week.
The board wins, and it should be the default.
Stalling is a property of an individual deal, not of a stage. On the board every deal is a discrete object with an amber left rule and a day count, so a manager scans five short columns and lands on the specific address that needs a phone call — then acts on it, because the card is the thing you drag to the next stage. The stage header still carries the count, the aggregate value and the threshold, so nothing is lost.
The funnel answers a different question — where value is concentrated and where it leaks between stages. It is the better artefact for the weekly leadership review and for a printed page, because it fits five rows and totals into one glance. But it aggregates the stalls into a number ("3 stalled"), which means the manager has to click through to find out which deals those are. That extra step is exactly the friction you do not want at 8am.
Recommendation: board as the working view with the "Stalled only" filter in the scope bar, funnel as a widget on the executive dashboard. Same data, two jobs.
Thresholds are per stage and shorten as risk rises: 10 d at initial offer, 5 d accepted, 3 d subject removal, 2 d deposit, 7 d completion. A stalled card shows the amber rule, the badge, and the reason in plain words — “no stage change in 6 d, subjects lapse Aug 19”. Thresholds are set per office by the managing broker.