A stakeholder engagement assessment matrix, or SEAM, is a grid: one row per stakeholder, one column for each of the five engagement levels, and two marks on every row. C for where somebody is now, D for where the work needs them. The gap between the two is the entire output.
One row per stakeholder, two marks per row
The grid is deliberately unglamorous. Down the side, every stakeholder the work has. Across the top, five levels in a fixed order: unaware, resistant, neutral, supportive, leading. In each row you place a C in the column that describes where that person or group sits today, and a D in the column where the work actually needs them.
That is the whole instrument. What makes it worth the hour is the second mark. Plenty of stakeholder analysis stops after describing the present: this group is powerful, that one is hostile, this one has gone quiet. Describing the present generates sympathy for your own situation and no actions. Recording a target next to it converts each row into a question with an answer: is there a gap here, and if so, who is closing it and by when.
The gap is the output, not the levels. A row where C and D sit in the same cell is finished work, and it should be visibly finished rather than quietly dropped. A row where they sit three columns apart is a piece of work nobody has scheduled yet. Read the grid by looking at distances, and the rows sort themselves.
The five levels, from unaware to leading
| Level | What it looks like | What moves them up |
|---|---|---|
| Unaware | Does not know the work exists. | Being told, by someone they already listen to. |
| Resistant | Knows, and is against it. | A conversation about the objection. |
| Neutral | Knows, holds no useful view. | A reason to care that is theirs, not yours. |
| Supportive | Wants it to work. | Something concrete to do. |
| Leading | Spends own capital to drive it. | Nothing. Protect it. |
The scale runs one way, which matters more than it appears. These are not five boxes to sort people into, they are an ordering, and ordering is what lets a gap have a size. Two columns of movement is roughly twice the work of one, and a row that needs to travel from unaware to leading is not an engagement action, it is a campaign.
Two of the levels get misread routinely. Unaware is not a criticism, and on a large piece of work it is the correct and expected starting position for most people; the error is leaving somebody there whose objection would have been cheap to hear early. Neutral is the one that looks safe and is not. A neutral stakeholder has heard about the work and has not found a reason to care, which means that when a competing demand arrives, nothing holds them.
Reading a filled grid
Filled in, the grid stops being a description and starts being a queue. The IT department is two columns from where the work needs it and everything else is one, so the IT department is where the next fortnight of engagement effort goes. Nothing about that conclusion required judgement in the room; it fell out of the marks.
Two refinements are worth the trouble. Add a date to the assessment, because the value is in comparing one grid against the last one. And consider a third mark for the level reached at the previous review, which turns each row into a trajectory: a stakeholder who has gone from resistant to neutral is responding to something, and a stakeholder who has not moved in three months is telling you the plan for that row is not working.
Not everyone needs to be leading
The most common way to waste this tool is to put every D in the last column.
It feels rigorous and it is the opposite. Leading means a stakeholder spends their own political capital to drive the work, and most stakeholders should not: an external vendor at leading is a supplier writing your strategy, and a regulator at leading has stopped being independent. For a great many rows the correct D is supportive, and for some it is neutral. A vendor who knows what is expected, has no objection and does the job is exactly where they should be, and moving them further costs you attention and buys nothing.
Setting every target to leading also destroys the tool’s one real function, which is prioritisation. If every row shows a gap, no row stands out, and the grid degrades into a list of everyone you have ever met with an implied obligation to court all of them.
Ask of every D you write down: what would this stakeholder do at this level that they would not do one level lower, and does the work actually need that thing? If the answer is vague, the target is too high. The discipline is in justifying the desired level, not in assessing the current one, which is usually obvious to anyone paying attention.
Closing a gap is a conversation, not a column
Here is the limit, and it is the same limit every grid has. The matrix locates work. It does not do it, and it does not tell you how.
That gap is widest on the resistant rows, which are the ones people most want a procedure for. Moving somebody from resistant to supportive means finding out what they are actually objecting to, which is frequently not what they said in the meeting. A PMI congress paper on the subject argues that overcoming resistance is not so much about persuading stakeholders as about incorporating and orchestrating them, building a dialogue rather than mounting a case. No column on a grid contains that. What the grid does is make sure the conversation is scheduled and that somebody owns it.
A grid can tell you a row has not moved in three months. It cannot tell you that the objection was never about the thing you kept explaining.
There is a related cost worth naming, because it is invisible in the artefact. The grid records judgements about named people, in writing, in a document that circulates. “Resistant” is a reasonable shorthand in a planning session and an accusation when it reaches the person it describes, and a matrix that leaks does more damage than the engagement it was meant to plan. Some teams handle this by recording roles rather than names, some by keeping the grid off the shared drive. What does not work is writing it as though nobody will ever read it.
Run the assessment with two people who see different parts of the organisation, not alone. Current engagement level is judged from observed behaviour, and one person’s view of a stakeholder is one set of meetings. The disagreements are the useful part: a stakeholder one of you reads as supportive and the other as neutral is almost always the row you have been getting wrong.
What the grid quietly assumes
Three assumptions sit underneath, and all three are usually fine until they are not.
The first, and the only one the grid cannot help you with at all, is that the rows are complete. A matrix says nothing whatsoever about a stakeholder who is not on it, and it says nothing loudly: the page looks finished, every row has two marks, and the omission is invisible precisely because the artefact is tidy. The rows that go missing are predictable. Regulators nobody has needed yet, the team that will inherit the thing after handover, a neighbouring department that finds out at go live, the community group with a rightful claim and no route into your governance. None of them are in the room when the grid is drawn, which is exactly why they are not on it. Building the row list from an organisation chart guarantees this failure, because the chart only contains people the organisation already knows it has.
The second is that a stakeholder holds one position. Groups do not. “End users” as a single row averages a department that contains enthusiasts and objectors, and the average is a fiction that satisfies the grid while describing nobody. When a row keeps refusing to move, the most likely explanation is that it is two rows.
The third is that the assessment is current. Positions drift for reasons that have nothing to do with your engagement plan: reorganisations, budget rounds, a new director with a different view. The UK government’s project delivery standard is explicit that stakeholder attitudes should be assessed, updated and validated throughout the work, which is a polite way of saying that a matrix built once at initiation and filed is worse than none, because it carries the authority of a documented decision while describing a world that has moved on.
None of the three argues against the tool. All three argue for keeping it cheap: a grid you will redo every quarter is worth far more than a beautiful one you will defend. If you have also run a salience model assessment, the two answer different halves of the same question: salience tells you whose claim commands attention, and the engagement matrix tells you what to do about the ones that do. The stakeholder engagement plan is what comes out of the second, and it is only as good as the gaps that fed it.
Common questions
- What is a stakeholder engagement assessment matrix?
- A stakeholder engagement assessment matrix, often shortened to SEAM, is a grid with one row per stakeholder and one column per engagement level. Each row carries two marks: a C for where that stakeholder is now and a D for where the work needs them to be. The output is not the marks but the distance between them, because that distance is what tells you who needs attention and roughly how much.
- What are the five levels of stakeholder engagement?
- Unaware, resistant, neutral, supportive and leading. Unaware means they do not know the work exists. Resistant means they know and are against it. Neutral means they know and have no view worth acting on. Supportive means they want it to succeed. Leading means they are actively driving it, spending their own capital to make it happen. The scale runs one way, so it is an ordering rather than a set of categories.
- How do you create a stakeholder engagement assessment matrix?
- Four steps. List the stakeholders, mark where each one currently sits on the five level scale, mark where the work genuinely needs them to sit, then read off the gaps. The third step is where the thinking happens, because the honest answer is often that the current level is already enough. Anything with no gap needs no plan, which is the point of doing it in this order.
- What do C and D mean in a stakeholder engagement assessment matrix?
- C marks the current engagement level and D marks the desired one. Both sit on the same row, so a row with C and D in the same cell needs nothing from you, and a row where they are three columns apart is asking for real work. Some teams add a third mark for the level reached at the last review, which turns the grid into a record of whether the engagement plan is doing anything.
Filed under Stakeholders
A payment agreed in advance and released when a piece of work reaches an agreed end state, conditional on the outcome rather than on the contribution made along the way.
A stakeholder holding power, legitimacy and urgency at once, whose claim takes priority over every other claim on the work.
Where a stakeholder sits on the scale from unaware to leading, judged on observed behaviour rather than on stated position.
Read next