Decide It Once, or Decide It Two Hundred Times

A pantry opens at nine. By ten past, somebody at the sign-in table has been asked a question they were not expecting, and they have answered it. Not badly — reasonably, kindly, on the spot, with four people waiting behind. By the end of the session they will have done that thirty or forty times.

None of those answers is written down anywhere as a decision. That is the whole problem, and it is not a records problem. It is that a question your program answers at the counter is a question your program answers differently, session to session, volunteer to volunteer, depending on who is asking and how the morning is going.

The same question, two mornings apart

Here is the shape of it. One question, asked on two different days, answered two different ways — both times by someone doing their best.

Read that second answer again from the other side of the table. Someone who has already made the harder decision to come and ask for food is now being questioned about it — and the person behind them, an hour later, is not. Nobody chose that. It emerged, from having no answer prepared.

And here is the part that should be uncomfortable: the people who get asked are rarely random. Memory is not applied evenly. It attaches to who looked like they might be pushing it, who came in a decent car, who did not say thank you last time. That is not a character flaw in your volunteers; it is what happens to every human being asked to make the same judgement forty times before lunch with nothing to lean on.

What follows is six decisions worth moving from the right-hand column to the left. They came out of building workbooks for this sector, and they are in rough order of how often programs told us they were still making them at the counter.

1. How soon can somebody come back?

This is a policy, not a law. It is not handed to you by the USDA and it is not the same at the pantry across town. Some programs set thirty days. Some set fourteen. Plenty set nothing at all, deliberately, and that is a legitimate and defensible choice — a program that has decided everyone may come whenever they need to has made the decision, which is the point.

What is not defensible is not having decided. Then the rule is whatever the volunteer at the table thinks it might be, and it varies.

Two things make a written rule work in practice. The first is that the check happens at the moment of the visit, not in a report three weeks later — a formula that looks backwards at that household's previous visit and says so on the row. The second is the wording. Our distribution log flags a too-soon visit as “earlier than your rule”, never as ineligible or refused, and that is deliberate. Programs make exceptions constantly and for excellent reasons: a fire, a new baby, a household that came on the wrong day and travelled two hours to get there. A pantry with zero exceptions is either turning people away rigidly or not recording the exceptions it makes.

2. What is the safe range, and what happens when a reading is outside it?

Every temperature log we looked at gets the first half right and drops the second. Somebody writes 41°F in a box, twice a day, for years. The log fills up. Nobody reads it.

The reading was never the point. Here is what a monitoring visit actually turns on:

The reading is recorded. What somebody did about it, at 6:41, is not.

That blank is the finding. It is not a data-entry lapse — it is a record that says nobody knows what happened to the food in that cooler. And it is entirely possible that the right thing was done and simply never written down, which is worse in a way, because the program has done the work and cannot show it.

So the decision to make in advance is not the number. It is: what is the person standing in front of that cooler supposed to do, and who do they tell? Write that down, and the reading becomes the beginning of a record rather than the whole of it. Our cold chain log ships no ranges of its own — those come from your food safety plan and your health department, and a template that supplied one would be making a food safety decision for a program it has never seen. What it does instead is refuse to let an out-of-range reading sit there alone.

3. How many do we cook?

Meal programs report waste. Almost nobody reports the other number.

Planned, produced, served. Collapse them into one and you lose the only comparison worth having. Produced minus served is waste, which everybody tracks because it is money. Planned minus produced is the number nobody tracks, and it is people: a service that ran short is a service where somebody who came did not eat.

Those two failures pull against each other, which is exactly why they have to be looked at together. Optimise waste alone and you will run short more often; the kitchen that never wastes a portion is usually the kitchen that runs out. A program that knows both numbers can have the real conversation — how much over-production it is prepared to buy in order to stop turning people away — instead of quietly drifting toward whichever failure is less visible. Over-production is not simply waste; some of it is the price of not running out.

4. Who needs which training?

This one is not discretionary in the same way — where a food assistance program requires civil rights training, the requirement is real and annual. But the part programs get wrong is not whether. It is who gets which.

Depth of training follows what a person actually does. A volunteer who interacts with applicants and participants, or handles personal information, or determines eligibility, needs the full training. A volunteer who does none of those — who unloads pallets and goes home — needs less. That distinction is about the role, not the job title, and a register that records “trained: yes” cannot show it at all.

Which produces the specific failure worth designing against: the person who was given the shorter training at orientation and then, three weeks later, was put on the sign-in table because somebody called in sick. Every box on their row is ticked. They are now deciding who gets food, without the training that decision carries. In our register that reads WRONG LEVEL, in red, above every other status — because in a monitoring visit it is the single hardest thing to explain away, and because it is a real gap rather than a paperwork one.

5. Whose insurance responds?

If your program has volunteers driving their own cars — meal delivery, senior transport, a food run — this is the decision on this page with the largest gap between how consequential it is and how rarely it is written down.

In a personal vehicle, it is normally the volunteer's own auto policy that responds. Many organisations carry nothing that covers a volunteer driver at all. And volunteers routinely assume the opposite: they are driving for you, so surely they are covered by you. People deliver meals for years on that assumption.

The chain looks complete. The missing link is the one nobody thinks to check for.

The decision to make in advance is a single question to your insurer, in writing: does our policy cover a volunteer driving their own vehicle, and if so, how does it sit relative to theirs? Then tell your drivers the answer, whatever it is. If the answer is that their own policy is primary, they are entitled to know that before they take a route — and some of them will want to talk to their own insurer, which is a conversation that should happen before an incident rather than after one.

Two related decisions belong in the same sitting. What liability limits do you require, if any — that figure comes from your insurer or your funder, not from a template. And how often do you re-check a driving record, given that one checked at intake and never again is a record about a person who no longer exists.

6. Who confirms an identity, and who signs?

The last one is narrower but it is the cleanest example of the whole pattern, because the failure is completely invisible from the outside.

At a free tax preparation site, volunteers sign a standards of conduct agreement. That agreement is not valid on the volunteer's signature. It becomes valid when the site coordinator, sponsoring partner, instructor or IRS contact has confirmed that person's identity against a government-issued photo ID and signed it themselves.

What happens in practice: orientation ends, forty people sign, the forms go in a binder, and the binder looks finished. Nobody counter-signs. The site finds out during a review.

Everything done except the two things that make it count: who confirmed the photo ID, and when they signed.

The decision in advance is trivially small — who at this site confirms identities, and when in the week do they sit down and do it — and it is the difference between a compliant site and a site that believes it is compliant. Our tracker starts every row at NOT VALID and keeps it there until both the confirming name and the counter-signature date exist, however many other boxes are ticked. The sample row we ship is deliberately complete in every other respect and still reads NOT VALID, because that is what a binder of signed forms looks like from the outside.

Three things you cannot decide in advance, and should not try

This essay could easily be read as an argument for turning a human service into a flowchart, which would be a bad outcome. So, plainly, the limits.

The exception. A rule exists so that departing from it is a decision rather than a mood. It does not exist to prevent departure. A household that came on the wrong day after a two-hour journey should get a box; what the rule buys you is that somebody chose that, on the record, and would choose it again for the next person in the same position. Never automate the exception away — and be suspicious of a report showing zero of them.

The concern at the door. No file alerts anybody. When a delivery volunteer reaches a door nobody answers, the response happens in the next ten minutes, by phone, by a person. What you can decide in advance is who that person is — a name and a number, known to every driver before they take a route. What you cannot decide in advance is what is wrong. The record's job is only to make it visible tomorrow morning when the call did not happen.

Anything about a survivor. Which brings us to the one decision that is already made for you.

The one place a spreadsheet is the wrong answer entirely

If you are a victim service provider receiving funding that prohibits entering client-level data into a Homeless Management Information System, you are required to use a comparable database that meets the applicable data, privacy and security standards. A spreadsheet is not one, and cannot be made into one. No workbook on our site — free, branded, well-intentioned or otherwise — is an acceptable place for client-level information about survivors.

That restriction is easy to lose behind the acronyms, so here it is without them. Information identifying somebody who is fleeing violence, sitting in a file that can be emailed, copied, synced, and opened by anyone who ends up with it, is a safety problem first and a compliance problem second. The requirement exists because of what happens when the wrong person finds the file.

More generally, for every program: where a file touches information about people asking for help, four decisions are worth making before the file exists rather than after. Where does it live. Who can open it. How long do you keep it. How is it destroyed. Use a household or client reference instead of a name wherever your program allows, and record no more than the decision in front of you actually requires. Every workbook in our set says this on its own first sheet, because the file outlives the page it was downloaded from.

Frequently asked questions

Is a visit-frequency limit required?

Not by us, and not by any rule we can state for you. It is a policy, and it differs by program, by state distributing agency and by partner food bank. Plenty of programs set no limit at all. What matters is that the answer is your program's rather than the sign-in table's, and that it is applied the same way to everybody.

Do we have to train volunteers who never talk to anybody?

Where a food assistance program requires civil rights training, the shape of it is that frontline volunteers — people who regularly interact with applicants and participants, determine eligibility, or handle personal information — receive the full training annually, and volunteers who do none of those receive a lighter version appropriate to their role. The exact definitions and curriculum come from your state agency and partner food bank. The practical point is that the level follows the job somebody actually does, and it changes the day their job changes.

Our insurance broker says we are covered. Is that enough?

Get it in writing, and get it specifically about a volunteer driving their own vehicle rather than about volunteers generally. Then tell your drivers what it says. The failure mode here is not usually a broker being wrong; it is a verbal reassurance that nobody wrote down and that the volunteers never heard at all.

Can we use these workbooks as our official program records?

Only if your program says so. None of them is an official form and none should ever be submitted in place of one. They are your own working records, which is the part that is genuinely hard to keep straight.

What does this cost?

Nothing. No account, no email gate, no watermark. We build background-check software and sell that separately; none of these six is a screening product.

The six workbooks

Free, branded, live formulas, no email required. Every interval, safe range, liability minimum, training interval and retention period is a cell you fill in from your own agreement — a number we invented would be trusted, and it would be wrong somewhere.

• Food Pantry Distribution & Visit Log — your frequency rule, applied by formula to every household
• Cold Chain & Food Safety Log — your ranges, and a flag that will not clear without a corrective action
• Community Meal Service Log — planned, produced and served as three numbers, with RAN SHORT counted
• Food Program Civil Rights & Training Register — the level derived from the role, and retention from your fiscal year end
• Volunteer Driver Eligibility & Route Log — whose insurance responds, and the door nobody answered
• Free Tax Site Volunteer Certification Tracker — NOT VALID until somebody confirmed a photo ID and signed

All six are on the human services toolkit page, which explains what each one holds open and why.

One last thing

Every decision on this page can be made in an afternoon. Not researched, not consulted on, not committee'd — decided, written in one place, and told to the people who will be asked about it at nine in the morning with four people waiting.

What you get for that afternoon is not efficiency. It is that the answer somebody receives at your counter stops depending on which volunteer is standing there and how the morning is going. For a program whose entire purpose is to treat people decently when they are having a hard time, that is not an administrative improvement. It is the thing itself.