# Stakeholder consultation: deciding without false consensus Stakeholder consultation: a practical method to gather input, preserve disagreement, and make traceable decisions without false consensus. - Canonical URL: https://www.harmate.com/en/blog/stakeholder-consultation-deciding-without-false-consensus - Author: Harmate Team - Published: 2026-09-06 - Updated: 2026-09-06T18:07:30.313839+00:00 - Language: en ## Content # Stakeholder consultation: deciding without false consensus Stakeholder consultation is a structured dialogue with people affected by a decision. It is not a disguised vote or a tally of opinions. Its purpose is to improve a proposal before a decision is made by exposing arguments, disagreements, and participation limits. The essential contract is simple: state what can still change, who will decide, and how contributions will be considered. Without it, stakeholder consultation quickly becomes a participatory alibi with no credible path to action. False consensus appears when a synthesis turns reservations or silence into collective agreement. ## What stakeholder consultation means The French [National Commission for Public Debate](https://www.archives.debatpublic.fr/glossaire/338) defines *concertation* — structured stakeholder consultation — as a process that gathers the views of stakeholders or the public before a decision. The authority retains the final choice but is expected to explain it in light of the process. The objective is not automatically consensus. What matters is a clear decision process and an honest account of how contributions can influence it. | Approach | What participants do | Who decides? | Honest output | |---|---|---|---| | Information | Receive a decision or proposal | The organiser | Shared understanding | | Consultation | Provide feedback, often once | The organiser | Collected input | | Stakeholder consultation | Examine a problem, exchange reasons, and improve options | The decision owner named at the outset | Better-defined options, agreements, and documented disagreements | | Negotiation | Seek an agreement between different interests | Parties authorised to conclude an agreement | A compromise or recorded disagreement | | Co-decision | Formally share the power to choose | Actors defined by the process rules | A joint decision | These labels are not a ladder of moral value. They describe different distributions of decision power. A clearly bounded consultation is more honest than “co-creation” with no real room for change. ## Start with the decision, not the survey A useful stakeholder consultation begins before choosing tools. The [OECD Guidelines for Citizen Participation Processes](https://www.oecd.org/en/publications/oecd-guidelines-for-citizen-participation-processes_f765caf6-en.html) recommend defining the problem, expected outcome, participants to involve, and their potential influence before selecting a method. ### Write the decision question Draft a question the decision owner will have to answer. For example: > How should we redesign the onboarding journey to reduce early drop-off without making it longer? This is stronger than “What do you think of our onboarding?” It gives the dialogue a practical object while leaving causes and possible solutions open. Check three conditions: - there is a real decision to make; - more than one option remains possible; - the people invited have experiences or knowledge that could change the choice. If everything is already fixed, inform people and explain why. If the only goal is to measure preferences between stable options, run a one-off preference survey. Stakeholder consultation is worthwhile when dialogue can still change the problem definition, the options, or their implementation conditions. ### Separate constraints from room for influence Write down what will not change: the maximum budget, applicable rules, timeline, available resources, or final accountability. Then name what is open: priorities, practical arrangements, acceptance criteria, and risks to address. This protects participants from a false promise. It also protects the decision owner: rejecting a proposal that conflicts with a disclosed constraint is not the same as ignoring the process. ### Identify affected groups without claiming representativeness A large list of participants does not become representative by default. Map the perspectives needed to understand the decision: people who use the service, deliver it, bear its effects, fund it, or hold critical expertise. For each group, record the invitation channel, number approached, responses received, and known absences. This describes participation. It does not certify statistical representativeness. ## Explore experiences first, then measure stable dimensions Early questions should let unanticipated reasons emerge. The [Pew Research Center’s guidance on survey questions](https://www.pewresearch.org/writing-survey-questions/) shows that open and closed formats can produce markedly different answers: listed options make some ideas more salient, while open questions can reveal responses missing from the list. Useful prompts include: - Tell us about the last time the current process helped or blocked you. - What seems unworkable in the current proposal, and why? - Under what conditions would this option become acceptable? - Who might bear a cost or risk we have not named? - What experience contradicts the diagnosis we presented? - What must remain unchanged? - If we could test one improvement, which one would you choose? A well-framed [open-ended question](/en/blog/open-ended-questions-get-actionable-data-without-bias) asks for a specific experience, reason, or condition. It is not automatically neutral, but it preserves respondents’ words, exceptions, and perceived causal links better than a predefined list. Closed questions belong later, once important dimensions have been checked against real experiences. They can compare how often a barrier occurs, rank options, or test whether people understand a category consistently. They should not erase the responses that were used to build the categories. Choose channels according to the evidence needed. A [focus group](/en/blog/focus-group-guide-method-agenda-and-20-open-ended-questions) exposes interactions between participants; an asynchronous survey gives people more time and reduces some group influence; an interview explores one journey in depth. A stakeholder consultation may combine these formats if each one has a clear role. ## Analyse input without smoothing out disagreement A credible synthesis does not search for an average sentence. At minimum, distinguish: - findings supported by several independent experiences; - disagreements, including their reasons and affected groups; - edge cases that change the conditions of an option; - proposals, which are not decisions yet; - missing information that still requires verification. Every important theme should remain linked to its source responses. A contextualised [verbatim quote](/en/blog/verbatim-how-to-use-exact-quotes-without-distorting-their-meaning) can make a tension understandable, but it does not establish frequency or representativeness. State who is speaking, in what situation, and why the quote matters without exposing identity unnecessarily. A simple decision table is often enough: | Option | Supporting arguments | Reservations or conditions | Documented signals and input | Blind spots and limits | Review status | |---|---|---|---|---|---| | Split onboarding into two stages | Reduces the density of the first contact | Requires a prompt follow-up | Experiences from new hires and buddies | Few manager responses | Test | | Add a mandatory group session | Creates shared reference points | Restricts schedules and may inhibit some questions | Support-team feedback | Effect on drop-off is unknown | Investigate | This structure prevents two shortcuts: treating a frequent theme as truth and treating an attractive suggestion as a feasible option. ## Example: consulting stakeholders before redesigning onboarding An organisation observes early drop-off in an onboarding journey. It could ask, “Would you like more meetings?” The question already contains a solution and reduces the issue to a preference. A stronger stakeholder consultation follows this path: 1. The owner states that they will decide on the new journey. Total duration and budget are fixed, while sequence, formats, and responsibilities can change. 2. New hires, buddies, managers, and support teams describe concrete moments of clarity, confusion, and drop-off. 3. The analysis separates shared barriers from situations specific to a role or location. Minority positions remain accessible. 4. Two options are framed alongside their implementation conditions. A closed question can then measure the distribution of reservations or preferences without erasing qualitative nuance. 5. The owner selects a test and publishes the arguments considered, proposals not chosen, and date for review. The output is not manufactured consensus. It is a decision whose scope, evidence, and disagreements can be inspected. ## Close the feedback loop Reporting back is not a decorative summary. The OECD recommends explaining how input was used and giving reasons when recommendations are not adopted. The French government’s [participation-process guide](https://www.modernisation.gouv.fr/files/2024-03/Guide_CIPC_concevoir_une_demarche_de_participation_citoyenne.pdf) likewise treats reporting as the transition between participants’ work and the organiser’s next steps. A useful report answers five questions: - what did we hear; - where are the agreements and disagreements; - what did we decide, and who owns that decision; - which contributions were accepted, rejected, or deferred, and why; - what observable step comes next? Publish the limits as well: missing groups, uneven participation, a constraint discovered too late, or insufficient evidence. Trust does not come from a frictionless summary. It comes from a checkable connection between what people said and what will happen. ## Mistakes that turn stakeholder consultation into theatre ### Asking after the decision If nothing can change, inform people and explain the reasoning. Simulated openness creates expectations that no report can satisfy. ### Confusing volume with diversity One hundred similar responses do not compensate for the absence of a directly affected group. Track who was reached and disclose blind spots instead of using a total count as a quality certificate. ### Starting with a list of solutions A closed list can measure preferences, but it rarely reveals unexpected causes. Explore experiences and conditions first, then stabilise categories. ### Removing minority positions A well-supported disagreement may reveal a rare but decisive risk. Keep it as evidence and explain whether — and how — it affects the decision. ### Failing to respond to participants Without a feedback loop, participation becomes extraction. A date, owner, and status for each option make the follow-through accountable. ## Where a tool such as Harmate can help The method works with interviews, workshops, and spreadsheets. A tool becomes useful when several stages must remain connected: building an [actionable questionnaire from the decision](/en/blog/actionable-questionnaires-start-with-the-decision-not-the-questions), tracking participation, analysing open responses with their source excerpts, keeping disagreements visible, and assembling a decision record. In Harmate, this workflow can connect the decision question, collection, source responses, disagreements, and a decision record. The tool does not decide for the owner, guarantee representativeness, or turn consultation into co-decision. ## Frequently asked questions ### What is the difference between consultation and stakeholder consultation? In common English usage, “consultation” covers a broad range of input processes. A one-off survey mainly collects opinions; the process described here adds structured dialogue, examination of reasons, option development, and feedback before a decision. The organiser should define the level of influence explicitly instead of relying on the label alone. ### Must stakeholder consultation produce consensus? No. It may produce partial agreements, shared conditions, and persistent disagreements. Those differences remain useful when they are linked to their arguments and considered in the final decision. ### How many people should be invited? There is no universal number. It depends on the affected groups, engagement channels, timeline, and breadth of perspectives required. Record who was invited, who participated, and which perspectives are missing. ### Can an online questionnaire support stakeholder consultation? Yes, when it belongs to a wider process: the scope is stated, questions let arguments emerge, responses can be explored, and participants receive feedback. A standalone form that only collects preferences is better described as a survey or one-off consultation. ### How can we tell whether stakeholder consultation was useful? Check whether it changed the understanding of the problem, the options, their conditions, or the final decision. Response count is secondary if contributions have no visible follow-up or essential groups could not participate.