Mendelow's Matrix Explained: Mapping Stakeholders on the Power Interest Grid, With a Worked Example

Mendelow's matrix is a two-by-two grid that sorts the parties affected by a decision along two axes: how much power they hold over the organisation, and how much interest they take in the particular issue. Each quadrant carries a prescription for how much management attention the group should receive.

It comes from a 1981 conference paper on environmental scanning, and the thing that has kept it in use for four decades is its economy. Stakeholder lists are easy to produce and almost useless, because a list implies every name deserves the same response. The power interest grid converts the list into a resource-allocation decision.

Mendelow's matrix explained in one sentence

Stripped to its claim, Mendelow's matrix explained is this: attention is finite, influence is unevenly distributed, and the sensible response is to spend the attention where the influence and the engagement coincide. Everything else in the model is bookkeeping for that sentence.

The two axes, defined precisely enough to use

Power is the capacity to affect whether the decision proceeds or survives. That includes formal authority, but also the ability to withhold something the organisation needs — capital, a licence, labour, a supply, planning consent. The test is not seniority but consequence: what can this group actually stop?

Interest is how much the group cares about this specific issue, not how much they matter in general. The distinction is where most classroom applications go wrong. A pension trustee has enormous power over a company and, on the question of which brand of forklift to buy, no interest at all.

Both axes are relative to the decision in hand. The same regulator is a key player on a licensing question and minimal effort on a marketing question, which is why a single grid drawn for "our stakeholders" is worthless. The grid is drawn per decision.

The four quadrants of Mendelow's matrix

High power, high interest: key players

These groups can stop the decision and are paying attention. They must be engaged before positions harden, involved in shaping options, not presented with conclusions, and given a route to raise objections that does not require them to escalate. Most of the management effort belongs here, and the commonest failure is engaging them after the option set has already been narrowed.

High power, low interest: keep satisfied

Powerful, currently indifferent, and capable of becoming key players the moment the issue touches something they care about. The management task is to keep them from being surprised. Enough information to prevent a surprise, not so much that you manufacture the interest you were trying to avoid — the balance most often got wrong in the other direction, by drowning a regulator in detail until they start reading it.

Low power, high interest: keep informed

Groups who care a great deal and cannot compel anything. They are the natural allies of a decision and the natural source of its public opposition, and the cost of keeping them informed is low. The reason to do it is not courtesy: an informed group with high interest and low power will often supply the argument that changes a key player's mind.

Low power, low interest: minimal effort

Monitor and no more. The quadrant's real function is permission to stop spending attention, which is what makes the grid useful at all. The risk is that a group is placed here because nobody asked, and the position is never revisited.

Placing a stakeholder without guessing

The weakness of most applications is that placement is asserted. Two habits fix that.

The first is to state, for each group, the single thing they can withhold. If nothing comes to mind, the group is low power, and the honest placement follows. "The local authority can refuse the licence" is a placement argument; "the local authority is important" is not.

The second is to date the grid. Positions move, and they move fastest on the interest axis. A group that was keep-satisfied in the planning stage becomes a key player the week the consultation opens. A grid without a date implies a permanence the model does not claim.

A worked Mendelow's matrix example on the power interest grid

Mapping power interest grid stakeholders is easier to follow on a case with real constraints than in the abstract. A regional water company plans a four-year capital programme to replace ageing mains, funded by a tariff increase requiring regulatory approval.

The economic regulator. Can refuse the tariff and therefore the programme. Interest in this particular submission is moderate and not high, because it is one of dozens under review. High power, moderate interest: keep satisfied, with early sight of the submission so that nothing in it is a surprise at determination.

The majority shareholder. Can replace the board and controls the equity injection the programme needs. Highly engaged because the return profile changes. Key player, and the one group whose objections have to be resolved rather than managed.

Residents on the affected streets. No formal power over the programme, and very high interest because the works will close their roads for months. Keep informed, generously and early. They cannot stop the works, but an organised residents' group supplies precisely the material a councillor needs to raise it with the regulator.

The trade press. Low power and, at this stage, low interest. Minimal effort, with the position reviewed if the programme slips, because a delayed capital programme is a story and interest is the axis that moves.

The value of a Mendelow's matrix example worked to this level of detail is that it produces four different actions rather than one plan labelled "stakeholder engagement". That is the whole of what Mendelow's matrix offers, and it is enough.

What the grid does not tell you

Three things sit outside the model and are regularly attributed to it.

It does not rank the quadrants by importance to the organisation's purpose. Key players get the most attention because they can stop things, which is a statement about leverage and not about whose claim is strongest. A community with no legal standing may have the better case and still belong in keep informed.

It does not say what to communicate, only how much. Two organisations can place the same group in keep satisfied and behave entirely differently — one sending a quarterly summary, the other offering a standing meeting — and the grid endorses both. The choice of channel and register is a separate judgement the model is silent on.

It does not handle coalitions. Power interest grid stakeholders are plotted as independent points, yet the events that change outcomes are usually two low-power groups combining, or a high-power group being persuaded by a high-interest one. A grid drawn at a single moment shows the pieces and not the alliances, which is the strongest practical argument for redrawing it at each milestone instead of filing it.

What the matrix cannot tell you about coalitions

The grid places each group on its own, and influence in practice is rarely exercised alone. Three effects follow, and a case study that mentions any of them is asking for more than a placement.

Low-power groups combine. Individually dispersed stakeholders — customers, residents, junior staff — sit in the low-power quadrants until something gives them a common channel: a regulator, a union, a campaign, a platform. Their claim does not change; their ability to impose a cost does, and it changes in steps and not gradually.

High-power groups borrow interest. A group with leverage and no stake in a particular decision can be recruited by one with a stake and no leverage. A supplier that cares nothing about a site closure will act on it if a customer it depends on does, which is how a low-power group reaches a decision it cannot influence directly.

Position is partly a product of what the organisation does. A supplier base chosen for its weak bargaining position stays weak only while the choice holds; a group kept informed, not consulted may organise precisely because it was kept informed. Mapping treats the distribution as given, and a strategy that lasts a few years is one of the things shaping it.

The practical instruction is to draw the map for the position you expect rather than only the one you have, and to say which movement would be most expensive. A map of today, handed in as a finding, describes a moment that the decision itself will change.

Where the model is criticised

The most substantial objection is that the grid is static, and describes positions at a moment while the decisions it supports run for years. Practitioners handle this by drawing it twice, now and at the point of the next major milestone, and treating the arrows between the two as the actual output.

The second is that it flattens groups that are not homogeneous. "Employees" spans people whose jobs are at risk and people for whom nothing changes, and a single point on the grid conceals that difference. The remedy is to split a stakeholder until each entry has one coherent interest, which usually means more points than a slide comfortably holds.

The third objection is ethical rather than analytical. The grid allocates attention by power, which means a group with a legitimate stake and no leverage is correctly placed in minimal effort. The model is a description of how influence works, not a defence of it, and an assignment that notes this distinction tends to read better than one that does not.

Drawing the grid a second time

The single change that most improves an application is drawing the grid twice: once for the position today, once for the position expected at the next milestone, with an arrow for every group that moves.

On the water company, the arrows are the interesting part. The regulator moves right as the determination approaches and the submission becomes one of the files actually being read instead of one of the files received. The residents move up, not because their formal power changes — it does not — but because an organised group with a councillor behind it acquires the capacity to delay through the planning process. The trade press moves right the moment the programme slips. Only the shareholder stays where it is, which is itself worth stating, because a stable key player is the group whose engagement plan can be set once.

Arrows also expose the placements that were never argued. A group that cannot be shown moving under any foreseeable event is usually a group that was placed by reflex, and going back to the withholding test tends to move it.

Using it in an assignment

A Mendelow's matrix example is expected in almost every stakeholder question, and reproducing the grid is worth nothing on its own. The marks are in the placement arguments, one sentence each, and in the four different actions that follow. A grid with four labelled quadrants and no stated action for any of them has stopped halfway through the model.

Two additions raise a competent answer. Name one group whose position you expect to move and say what would move it. And state the decision the grid was drawn for in the first line, because a reader cannot judge any placement without it.