Thomas Gaultier
Menu

Grievance Data as Early Intelligence for Management

Thomas Gaultier 9 min read

The most useful management report I have ever produced was one page long, and one line on it mattered. The line said that a large share of the quarter’s grievances traced back to a single contractor’s operations. The general manager acted within the week. The contractor was summoned, a corrective plan demanded, a hold placed on its next payment milestone. Within a quarter the category had halved. Nothing on that line was new. The same fact had been sitting in the case database for months, spread across dozens of records that nobody had read together.

That is the condition of most grievance data. It is collected, because the standard requires it, and filed, because nobody has asked what it is for. A register read case by case tells you about cases. A register read as a dataset tells you where the operation is generating harm and which teams and contractors are generating it. It tells you which settlements have stopped trusting you, and what will be blocking the haul road in three months. The UN Guiding Principles ask a mechanism to be a source of continuous learning. In practice that means somebody has to read the data, and somebody with authority has to act on it.

This article is about turning a grievance register into early intelligence for site management.

Three things worth measuring

A measurement system that tracks only volume and closure rate produces a number that looks like performance and means nothing. Three dimensions are needed. Operational performance asks whether the mechanism processes cases as designed: acknowledgment-on-time rate, average time to close by severity level, cases past their deadline, and the backlog. Outcome quality asks whether the results are fair and accepted: complainant satisfaction, cases reopened, cases escalated to mediation or appeal, and remedies verified as delivered. Systemic impact asks whether the mechanism is changing the operation. Are recurring categories falling over time, are operational changes traceable to grievance data, and what is the trend in the settlements with the worst history?

Every one of these should be disaggregated before it is reported. By gender of complainant, by settlement, by category, by the department or contractor concerned. An average is a way of not knowing. A satisfaction rate that is high overall and low among women, or high near the gate and low behind the hill, is two facts. The second one is the intelligence.

The lender and standards frameworks point the same way without saying how. IFC Performance Standard 1 requires the company to monitor and measure the effectiveness of its management programme and to act on the findings. The IRMA standard requires that mechanism performance be monitored and evaluated over time. Neither tells you what to look at. The three dimensions do.

Start small if the register is thin. Five indicators, tracked honestly for four quarters, teach more than forty tracked badly. The acknowledgment-on-time rate, cases past deadline, satisfaction at close-out, reopened cases, and the top three categories by settlement will carry most sites for the first year. Add the rest when someone is actually reading the page.

Reading the register for patterns

Patterns are where the register turns into intelligence, and a handful recur on almost every site. A rising count in one category, dust or blasting or water, points to an operational change upstream: a new haul route, a changed bench sequence, a pump drawing harder. A cluster in one settlement points to a relationship problem, often a broken commitment, that will surface in the next round of negotiations whatever the cases are nominally about. A cluster against one contractor points to a supervision failure that the contractor’s own reporting will not show you. A rise in cases reopened points to remedies that were closed on paper. A fall in intake with no operational change points to a mechanism people have stopped using.

Timing matters as much as category. Grievances that arrive in the weeks after a blasting schedule changes, a contractor mobilises or a compensation round closes are telling you about that event. It does not matter which box they were filed under. Plot the register against the project calendar once a quarter and the causes usually show themselves.

The last one deserves emphasis because it is read backwards so often. A quiet register beside an operation that blasts, hauls and draws down the water table is not good news. It is a trust reading. Before congratulating anyone, check that the channels are alive. Ask people in the furthest settlements whether they know the mechanism and would use it. Re-run the barrier analysis for the current project phase. Then expect the cure to look like deterioration. A mechanism that starts earning trust sees intake rise for a year or two as people surface what they had been holding back. Volume falls for the right reasons only later, when root-cause work removes the sources.

These patterns are also the earliest warning the operation will get. I made the case in Early Warning Systems for Mining-Community Conflict that the signals of an escalation are visible months ahead. The grievance register is the cheapest sensor in that system, because the community is filling it in for you. It is also the only sensor that tells you what people think, rather than what the monitoring equipment measured. In this work, what people think is the thing that closes roads.

From pattern to action

Data collection without a decision attached is filing. Each pattern should have a standing response. A rising category triggers a root-cause review with the operational department that owns it, and a change or a documented reason why not. A settlement cluster triggers a visit by someone senior, a review of open commitments to that settlement, and a conversation about what is actually wrong. A contractor cluster triggers a corrective action plan, and where the contract allows it, a hold on payment until the plan is delivered. A reopened-case rise triggers an audit of close-out practice. A silent register triggers the trust diagnostic above.

The consequence should also travel back to the community. When grievance data drives a change, a revised blasting schedule, a new watering regime on the haul road, a rerouted access track, tell the people whose complaints produced it. Nothing recruits the next complainant like a remedy the whole village can see. And nothing kills a mechanism faster than a community that files, sees nothing change, and concludes that the register is where complaints go to be counted. I have heard it put more bluntly in a village meeting: the company has a book for our problems, and the book is full. The same logic runs through Building Trust Through Data Transparency: shared data builds trust only when the sharing leads somewhere.

Reporting so that someone acts

Measurement matters only if it reaches a person with authority, in a form that provokes a decision. Three tiers are enough, and each should fit on the time its audience actually has. A weekly line to the grievance team itself: open cases, deadlines this week, anything Level 3. A monthly page to site leadership: the three dimensions, disaggregated, with the patterns named and an action against each. A quarterly review with senior management and, where relevant, lenders, with trends over four quarters and the operational changes made.

Three rules govern every tier. Never report a raw indicator without context: an eighty percent resolution rate means nothing without the severity mix, the backlog and the satisfaction score beside it. Always end with actions, named and owned, because a report that ends with data is administration. And keep it short. The one-page monthly report is the discipline that makes the data usable. The line that moves the general manager is usually a single sentence about one contractor, one settlement or one category. Two numbers belong on the general manager’s own dashboard whatever else is there. The first is the acknowledgment-on-time rate, which moves weeks before complaints reach the office. The second is the count of cases past their severity deadline, with the three oldest named.

Data the board should see

Grievance data is also social risk data, and boards are beginning to ask for it. Consider a register that shows recurring water complaints in a settlement on the access road, or a rising trend against the security contractor. That is a leading indicator of exactly the disruption Franks and colleagues documented as translating social risk into business cost. A board that sees only lagging indicators, the blockade after it happened, the lawsuit after it was filed, is governing in the rear-view mirror. The monthly grievance page, summarised to a quarter’s trend and a handful of named patterns, is a leading indicator a director can read in five minutes. The wider set of relationship measures I describe in Social Performance Indicator Tracking belongs alongside it.

Build the monthly page from the register you already have

The Grievance Data Intelligence Framework sets out, in four sections, what to measure and what each pattern should trigger. The sections are core indicators across the three dimensions, disaggregation cuts that expose hidden gaps, recurring patterns with the action each calls for, and reporting tiers with their audiences. It is a two-column reference rather than a scored checklist: read the indicator or pattern on the left, and apply the disaggregation, action or audience on the right. Use it to design the monthly one-page report, and then use the report to test whether the mechanism is a source of learning or a filing system. Download the Grievance Data Intelligence Framework

Intelligence is only useful if it changes the conversation

A grievance register read as intelligence tells management where the operation is hurting people before the hurt becomes a confrontation. That protects the households behind the cases and the investment behind the operation at the same time, which is the only kind of protection that lasts in this work. But the data cannot do the talking. A settlement cluster or a contractor pattern is a reason to convene, not a reason to send a memo. The conversation it calls for is a structured one: the affected people, the department or contractor concerned, and someone independent enough to keep it honest.

That is the role the Social Accord Architecture gives its Feedback Loop. Grievance indicators are named sensors with owners and a rhythm. Their readings go to a council where company and community both sit. A pattern that the operation cannot resolve on its own goes to the mediation pathway rather than into a drawer. Dual accountability makes the loop work: the company is accountable for acting on what the register shows, the community for using the register rather than the road. A mediator brought in on the strength of a pattern, before the blockade, is the cheapest intervention in the whole system.

One thing to do this month. Take your last quarter’s register, sort it by settlement and by contractor, and write the two sentences that a general manager would need to read. If you cannot write them, the data is not being read. If you can, put them on one page and send it. Then do it again next month, because the value is in the trend, and a general manager who has seen three pages starts asking for the fourth. To design a grievance intelligence system or to convene the conversation a pattern is calling for, write to me at thomas@thomasgaultier.com.

I am Thomas Gaultier. I mediate company-community disputes in mining, oil and gas, and energy projects, and I have resolved more than 2,000 claims across Sub-Saharan Africa and Europe. More about the work, or tell me what you are dealing with.

If the relationship around a project is under pressure, or heading there, let us talk.

A 30-minute call is enough for me to understand the situation and tell you honestly whether I can help. Active opposition, a social license problem, an FPIC process that needs an independent facilitator, a negotiation a community has to prepare for, or a conflict that is not public yet but will be. It makes no difference which side of the table you are on. No obligation, no generalist pitch.

Every first conversation is confidential.

Cover of Legacy Grievances and Historical Claims by Thomas Gaultier

New release, and the last one

Legacy Grievances and Historical Claims

The tenth and final volume completes the library. It is the field guide to harm that outlived the operation: transitional justice adapted to corporate disputes, successor liability when the asset changes hands, and the pathway behind more than two thousand resolved claims.