Risk Lists Written with the Volunteers Who Carry the Change, Not by the Coordinator Alone or Not at All: Operational Changes at Food Bank Distribution Sites

Abstract

We randomly assigned 168 distribution sites of 14 food banks to one of three ways of preparing operational changes for nine months: a risk list written by the site coordinator with two or three volunteers who would work the change ("shared list," 56 sites), the same list written by the coordinator alone ("coordinator list," 56), or no list (56). At the 163 sites that completed the study, shared lists had anticipated 64.4% of the problems that later appeared, against 51.1% for coordinator lists, and a problem disrupted a mean 2.1 distribution sessions at shared-list sites, against 2.5 and 2.7. Of volunteers active in month 1, 71.5% at shared-list sites were still active in month 9, against 62.2% and 60.4%, including 69.5% of volunteers whose objection to a change had been recorded and overruled. Shared-list changes took a median 11 days to launch, against 9 and 8.

1. Introduction

A food bank distribution site runs on a small number of paid or long-serving coordinators and a much larger number of volunteers. When a coordinator changes how the site works, moving to appointment slots or changing the order in which tables are set out, the volunteers are the ones who meet the consequences first, usually in a queue of clients on a cold morning. They often know what will go wrong, and they are rarely asked before the change starts.

The library report in this issue found that managers with a brief written list of likely problems more often adjusted and continued a change when one appeared. A list depends on what its author can foresee. We asked whether adding the people who would work the change to the writing of the list made it more complete, and what it cost the site in time and in disagreement.

2. Method (Summary)

Fourteen food banks took part. Their 168 distribution sites were randomly assigned, stratified by food bank, to one of three practices for nine months. Each site had one coordinator. A significant change was any change to distribution times, site layout, queueing or appointment arrangements, client registration, or volunteer roles that the coordinator recorded in the site log.

At shared-list sites (n=56), before each significant change, the coordinator met for about 20 minutes with two or three volunteers who would work the change, chosen by rota. Together they wrote a one-page list of likely problems, the sign of each, and what they would do if it appeared. Any volunteer who disagreed with the change itself could have the objection written on the list. The coordinator still decided whether to go ahead, and a recorded objection did not stop a change. At coordinator-list sites (n=56), the coordinator wrote the same one-page list alone. At no-list sites (n=56), coordinators made changes as before.

Problems and disruptions came from the site logs, which recorded each distribution session. A session was disrupted if a core item ran out before the last registered client was served or if the logged wait for any client exceeded 45 minutes. A problem was any disruption, complaint, or volunteer report the coordinator linked to a recent change, and disrupted sessions per problem were counted until the coordinator logged it as resolved. Two researchers, blind to condition, judged whether each logged problem at a list site matched an item already on that change's list. Volunteer retention came from rota records: a volunteer active in month 1 was counted as retained if they worked at least one session in month 9. In month 9, volunteers (1,163 respondents of 1,548 active that month) rated whether "my concerns about how the site runs are heard" (1 = strongly disagree, 7 = strongly agree) in a survey. Volunteers who left completed an optional exit form with a free-text reason, which the researchers coded blind to condition. Time from a change's first log entry to launch came from the site logs.

3. Findings

Five sites closed or merged during the study, and the results below are for the 163 that completed it (55 shared-list, 54 coordinator-list, and 54 no-list).

Shared-list sites logged 214 significant changes, coordinator-list sites 236, and no-list sites 229. These changes gave rise to 118, 131, and 126 logged problems, respectively. Shared lists contained a mean 5.8 items, against 4.6 for coordinator lists.

At shared-list sites, 76 of 118 problems (64.4%) matched an item already on the list, against 67 of 131 at coordinator-list sites (51.1%). The shared-list share was 26% higher.

A problem disrupted a mean 2.1 distribution sessions before it was resolved at shared-list sites, against 2.5 at coordinator-list sites (19% more) and 2.7 at no-list sites (29% more).

Of volunteers active in month 1, 431 of 603 at shared-list sites (71.5%) were still active in month 9, against 366 of 588 at coordinator-list sites (62.2%) and 369 of 611 at no-list sites (60.4%), gaps of 9.3 and 11.1 points.

Of the 603 shared-list volunteers active in month 1, 59 had an objection recorded on a list for a change that went ahead. Of these, 41 (69.5%) were still active in month 9, close to the rate for shared-list volunteers overall.

In the month-9 survey, shared-list volunteers rated "my concerns are heard" at 5.3/7, against 4.4/7 at coordinator-list sites (shared-list ratings 20% higher) and 4.3/7 at no-list sites (23% higher). Among departing volunteers who completed an exit form, 13 of 104 at shared-list sites (12.5%) gave a reason coded as not being listened to, against 19 of 121 (15.7%) and 20 of 128 (15.6%). The most common reason in all three groups was a change in work or family commitments.

The median time from a change's first log entry to launch was 11 days at shared-list sites, 9 at coordinator-list sites, and 8 at no-list sites. List meetings at shared-list sites lasted a mean 24 minutes, with a mean 2.6 volunteers attending, about four volunteer hours per site over the study. Of the 214 shared lists, 47 carried at least one recorded objection.

4. Discussion

A list written by a coordinator alone is a view of the site from the coordinator's desk. The volunteers at the tables see a different site: where the queue bends when it rains, which shelf empties first, which client always arrives early and needs a chair. Shared lists caught about a quarter more of the problems that later appeared, and problems at those sites were resolved after fewer disrupted sessions. Much of that benefit would have reached the clients in those queues, through fewer stock-outs and shorter waits, though we did not survey clients directly.

The coordinator list did less than we expected. It was the same form and took less time, but its problems were resolved after only slightly fewer disrupted sessions than at no-list sites, and it made little difference to retention (62.2% against 60.4%).

The cost fell mainly on time. Changes waited a median three days longer than at no-list sites, volunteers gave about four hours per site to list meetings, and coordinators may also have heard objections they would have preferred not to hear: about one shared list in five carried one. Volunteers whose objections were recorded and overruled stayed at about the same rate as their colleagues, so having a disagreement written down and then overruled did not appear to drive them away, though we cannot tell how volunteers at other sites, whose objections went unrecorded, fared.

5. Limitations and Future Directions

Coordinators decided which problems to link to a recent change, and a shared list may have made them more willing to log problems volunteers had predicted, or less willing to log ones nobody had. Matching problems to list items involved judgment, though the raters did not know the site's condition. Volunteers chosen for list meetings by rota were not always those most affected by a change. Retention may reflect the attention of being asked rather than anything about the risk list itself; a meeting with volunteers on another topic would be needed to separate the two. The 59 volunteers with recorded objections chose to object and had no comparison group, so their retention may reflect their engagement rather than the recording. Shared lists were longer, which by itself would raise the share of problems they matched. The survey and exit forms were voluntary, and the exit-form counts are small. Food bank sites already rely on volunteers' goodwill, and sites where relations were poor at the start may respond differently to recorded objections. A follow-up will add clients to the list meetings at a subset of sites.

Editorial Note

This report underwent CfAHC's Rapid Review Protocol, including AI-assisted literature cross-checking and a two-reviewer human sign-off.