2ndOpinion.FYIEN
genius.wiki

#492 2020 · Connecticut Department of Labor · Government / public administration

Connecticut's unemployment call center wasn't drowning in claimants — it was drowning in the same claimants calling back out of fear they'd lost their place in line

问题

A service channel appears overloaded by raw demand, and the obvious response is to add capacity or make people wait longer, when a large share of the apparent load is actually self-inflicted — generated by the system's own failure to reassure people that their request is still being handled

背景

When COVID-19 lockdowns hit in March 2020, Connecticut's unemployment claims volume rose roughly fivefold almost overnight, and the state Department of Labor's call center — sized for ordinary claim volume — was suddenly fielding around 26,000 calls a day with hold times stretching for hours. The standard response to that kind of surge is to add phone lines, hire more agents, or simply let the queue grow longer, treating every incoming call as a discrete, necessary unit of demand that has to be served one way or another.

But a large share of that call volume was not 26,000 distinct claimants each calling once — it was a much smaller number of anxious claimants calling, waiting on hold with no visibility into their position, giving up, and calling straight back because a dropped or abandoned call gave them no assurance their place in line still existed. Every hang-up-and-redial cycle added a fresh call to the day's total without adding a single new person actually needing help, inflating the apparent size of the problem the department was trying to solve.

换别人会怎么做

The standard response to a fivefold surge in call volume was to add phone lines and hire more agents, treating every incoming call as a discrete, necessary unit of demand that had to be served one way or another, since that's the natural response to a queue that appears overloaded.

他们看到了什么

The Department of Labor saw that a large share of the 26,000 daily calls wasn't 26,000 distinct claimants, it was a much smaller number of anxious claimants calling, waiting on hold with no visibility into their position, giving up, and calling straight back because a dropped call gave them no assurance their place in line still existed, meaning every hang-up-and-redial cycle inflated the apparent demand without adding a single new person actually needing help. Rather than continuing to expand raw call-taking capacity to serve what looked like genuine demand, the fix was targeting the redial behavior directly, deploying a callback system that let callers request a scheduled, guaranteed callback and hang up, removing the anxiety that was driving the repeat calls in the first place.

那一手

Rather than continuing to expand raw call-taking capacity, the department deployed a callback system that let a caller request a scheduled callback and hang up instead of waiting on hold, with the system tracking each caller's phone number to guarantee they would not lose their spot and would not need to call back to check. The fix targeted the redial behavior directly: a caller who trusted they had a genuine, tracked reservation for a callback had no remaining reason to hang up in frustration and immediately dial again.

为什么管用

Guaranteeing a tracked callback reservation meant a caller had no remaining reason to hang up in frustration and immediately redial, since they now trusted their place in line was real and preserved regardless of whether they stayed on hold or hung up, eliminating the exact behavior that had been inflating the call volume beyond genuine demand. Because the intervention addressed the actual cause of the apparent overload, anxiety-driven repeat contact, rather than adding capacity to serve a volume that was itself partly manufactured by the system's own lack of reassurance, daily call volume fell from roughly 26,000 to about 11,000 within about a week, a drop of well over half achieved without hiring a single additional agent or adding a single phone line. This is why the callback approach became a standing part of Connecticut's claims process rather than a one-time pandemic patch, the fix solved the actual bottleneck at its source rather than compensating for a symptom that would have kept recurring under any amount of added capacity.

值了多少

Within about a week of the callback system going live, the department's daily call volume fell from roughly 26,000 to about 11,000 — a drop of well over half achieved without adding agents or phone lines, because most of the eliminated volume had been repeat, anxiety-driven redials rather than new claimants. The Connecticut Department of Labor continued expanding the scheduled-callback approach into 2021, adding an online call-scheduling feature to its Consumer Contact Center as a standing part of its claims process rather than a one-time pandemic patch.

什么时候会失灵

The mechanism depends on the apparent overload genuinely being inflated by anxious repeat contact rather than reflecting real, distinct demand, a queue where every call truly represents a different person with a different unresolved need would gain nothing from a callback reservation system, since there'd be no redial behavior to eliminate in the first place. It also depends on the callback guarantee actually being trustworthy and reliably honored, a callback system that itself failed to call people back on time, or lost track of reservations, would recreate the same anxiety and redial behavior it was meant to eliminate, just shifted to a different failure point. And this fix specifically addresses demand inflated by lack of reassurance about queue position, it does nothing to reduce genuine underlying demand or speed up the actual resolution of each real request, meaning an organization still needs adequate capacity to handle the true volume of distinct requests once the artificial inflation from repeat contact has been stripped away.

后来呢

The episode is cited in call-center and public-sector service-design literature as a clear example of demand inflation caused by a channel's own lack of reassurance — a pattern generalized well beyond unemployment offices to any high-anxiety queue (benefits hotlines, emergency services non-emergency lines, customer support after an outage) where a large share of apparent overload is self-generated repeat contact rather than genuine new demand.

资料来源

  1. [1]The Secret to Reducing Repeat CallersMindful (callback technology vendor case study), 2021getmindful.com
  2. [2]DeLauro Statement: New CT Department of Labor Contact CenterOffice of Rep. Rosa DeLauro, 2020delauro.house.gov

收下它

同一路聪明