Twelve names, three lists, and the same two questions asked one person at a time. What a farm believes about its own standing is an opinion about an opinion, and it is not evidence of anything. The reading is the written answer of somebody who was asked on their own, laid next to the other eleven, grouped by subject, and ordered by how many times the same subject came back.
The dust off the access road had been bothering the neighbor for six months before the farm heard about it, and it heard from a third person at a barbecue. The buyer who took a load every year stopped calling, and nobody knows why, because nobody asked. The man who quit gave the real reason to a workmate in the shed and a different one at the gate.
When a certifier asks whether there is a channel through which somebody outside can complain and get an answer inside a stated time, the farm says the door is always open, which is neither a channel nor a time. Each of those arrived late, and each arrived through the one route that already has a deadline and a counterparty on the other end.
An afternoon of calls and visits, plus one hour at the table afterwards, produces the sheet. Setting a standard, watching what happens, comparing the two and correcting is what the four functions of management call control, and the sustainability axis is where the farm’s standing with the people around it stops being a feeling and becomes a page. The method is the one already used inside the gate for the goal, which is asking each person separately, turned around to face the other way.
Why your own read on the farm’s standing is not evidence
Because it is a reading of you, taken by you, and the error does not always run in the direction you expect. Booth and colleagues, writing in the Journal of the Royal Society of New Zealand in 2024, treat what a community grants a business as an unwritten contract, meaning the informal social expectations that a local community holds for a business. They ran two online surveys in New Zealand in 2023, one for the public and one for farmers, and the farmer half drew twenty-five answers. Twenty-five answers cannot describe farmers anywhere, and the authors say so where they compare the two halves.
What survives that limit is the shape of the finding. Where the two groups lined up most closely on what good farming means, the public rated farmers as a more trustworthy source of information than farmers rated themselves. Guessing wrong in either direction costs the farm the same thing. A farm that assumes the neighbors think worse of it than they do keeps quiet to avoid a fight it is not having, and a farm that assumes the opposite keeps quiet because it sees nothing to discuss.
Both end the year with no sheet, and the sheet is the only part of this that anybody else can read. So write your own guess down before the first call, in a line or two, and keep it beside the finished list, because the gap between the two is the reading and it exists on the page only if both halves are on it.
Who goes on the three lists
Everybody the farm’s activity lands on, which is a longer list than the one you would write from memory. The FAO published SAFA in 2014, a framework for assessing sustainability in food and farming businesses, and the definition it works from is the one Freeman wrote in 1984: a stakeholder is anybody who “can affect, or is affected by, the actions of the enterprise”. SAFA separates those who hold power over the business from those who merely absorb what it does. Under the sub-theme it calls Stakeholder Dialogue, a business that meets the objective “is able to list all stakeholders”. A business that cannot describe how it finds or engages them is recorded as unacceptable, not as incomplete.
| List | Who is on it | What only this list can tell you |
|---|---|---|
| Over the fence | Four neighbors along a boundary or the access road | What the farm does to the ground, the air and the road that nobody inside notices any more |
| Down the chain | Four out of the buyers, haulers and suppliers who deal with the farm | Whether the farm is easy or expensive to deal with, which the price never shows |
| Inside the gate | Four of the people who do the work, family included | What gets said in the shed and never at the office |
The partner institution goes on the middle list. Where a university or an institute runs a trial in one of your fields, a signed partnership term already names a person and a phone number on both sides, which makes it the one relationship out of the twelve where working out who to call is not part of the job.
A farm with no hired hands still fills the third list. GRASP, the social practice add-on that GLOBALG.A.P. auditors assess, settles that case by treating the core family members working on the farm as the workers. Where four neighbors, four buyers and four people in the shed are a larger share of everything the farm has, one unanswered complaint weighs more and not less.
Why twelve conversations, and why they are not a survey
Somewhere between six and fourteen conversations is where new subjects stop appearing, and where in that range it lands depends on how alike the people are. Guest, Namey and Chen, in PLOS ONE in 2020, measured how fast a set of interviews runs out of new material, and found the first six or seven interviews carrying most of the subjects in a group of people alike enough to be treated as one, with no new subject after eleven conversations on their looser count and after fourteen on their stricter one in two of their three sets of data.
Their third set held more variety among the people interviewed, and it needed more conversations and still ended covering a smaller share of the subjects. Three lists of four are that third kind of set on purpose, because the three groups see different things, and the price of the variety is the one they measured: twelve mixed conversations do not run the subjects out the way twelve alike ones would.
Twelve conversations measure nothing about a population, and the sheet must never be shown as if they did. Those were long research interviews run by people trained to run them, not four minutes at a gate. What carries across is the shape of the curve and not the number on it: the complaints run out faster than anybody expects, and the ones that keep coming back are the ones worth counting.
The same question, asked alone, written before you answer back
Two questions, the same words every time. What does this farm do that gets in your way, and what does it do that helps. Ask each person on their own, because a second listener turns the round into a rehearsal and the first confident answer becomes everybody’s. Write the answer while it is being said, and do not correct it, because correcting it on the spot is what ends the collection. SAFA lists the ordinary ways of doing this and none of them are exotic, naming “surveys, meetings, interviews and focus groups”.
Recording the words instead of your summary of them is what makes the sheet worth rereading. Booth and colleagues warn that the two groups may use the same term and mean different things by it, giving as their own case sustainable farming, which the public and farmers may read differently even where the core ideas expressed are similar. A summary written in your own words converts their sentence into yours, and the disagreement you spent an afternoon buying disappears in the act of writing it down.
Grouping and counting, which is what makes it a reading
Put like with like, count, and order the lines by the count. Twelve answers are twelve stories until the repeated subject is visible, and repetition is the only thing in this exercise you can count without judging anybody. Lay the sheets out, mark what each answer is about, and let the same subject raised by three different people become one line with a three beside it. The order of the lines is the output. The top line is the one the farm has to answer first, whether or not it agrees with it.
Written is what makes any of it hold. SAFA asks a business to show how its decisions changed because of what it heard, and to hold “evidence (minutes, notes, source documents) of the impact” rather than a recollection of the conversation. It also names the failure this sheet exists to prevent, marking down the business that “routinely fails to inform stakeholders of the outcome of engagement”.
Listening well and never reporting back is a failure with a name in the standard. Read the finished sheet against the community support record, which holds what left the farm for the community and who authorized it, because the two pages answer each other and the second one contradicting the first is the ordinary result rather than the odd one.
The three lines at the bottom of the sheet
Who receives the next complaint, how many days the farm takes to answer, and where the complaint gets written down. None of the three is an invention of this article, and the requirement does not stop at the payroll. Under the sub-theme it calls Grievance Procedures, SAFA asks that everyone the business touches, “in or outside of the enterprise”, reach a complaint procedure “without a risk of negative consequences”, and its indicator asks the farm to describe that procedure for each group it deals with, how it is publicized and how much it actually gets used.
GRASP requires the complaint process to state “time for resolution, who will answer” and to be usable “free of any retaliation or penalty”. It puts a number on the deadline for hired workers, thirty days from filing, or the last day of employment if that comes sooner, and it asks the farm to keep a summary record of every complaint of the past twenty-four months, showing which were resolved and how long each took. Those are audited under a private standard that reads the same in every country, which is why the deadline is a number rather than a local rule.
An open door fails the first line, not the third. The United Nations Guiding Principles on Business and Human Rights, endorsed in 2011, put it in one sentence: a mechanism serves its purpose only if the people it is meant for “know about it, trust it and are able to use it”. The same principle asks the procedure to carry an “indicative time frame for each stage” and to be known to every group it was set up for. GRASP adds the awkward half nobody volunteers: at least one place to file should be independent of the supervisory staff, because the person being complained about is often the person standing at the only door.
What you do with the line at the top
Answer it, and answering is not the same as agreeing. GRASP counts a complaint as dealt with when it gets a positive or a negative answer, or no resolution at all with the reason for that written down. That removes the excuse that kills most of these sheets. You do not have to be able to fix the dust off the road in order to close the line about the dust off the road. You have to answer it, on a date, to the person who raised it, and write down what you said.
The top line needs an owner, and the owner has to be a box on the drawing rather than whoever was standing nearby. That is what the farm org chart is for, and it is also the test of the first of the three lines: if nobody’s sheet says they take complaints from outside, the deadline has no owner and the line is decoration. Write the date of the next count before you file this one, because a single count is a photograph and two counts a year apart are a reading.
Where to start
Twelve names, an afternoon of calls and visits, and one hour at the table afterwards.
The second count is the hard one, because by then the top line is a subject the farm already promised to answer, and the sheet stops being a listening exercise and becomes a record of what got done about it. That is the reason to write the date of the second count on the day of the first. A neighbor who complains to you is a neighbor who still expects an answer from you. The one who stopped complaining to you is the one whose complaint is already somewhere else.