GRPI Model

What the GRPI model is

The GRPI model is a framework for finding what is holding a work team back. It arranges the causes of that breakdown in four layers: goals, then roles, then processes, then interpersonal relationships. The four letters are the initials of those words in English.

The model is neither a performance appraisal framework nor a measure that produces a score. What it offers is an order of examination: where to look first when people say a team “does not work together”, and what is left unexamined until the layer before it has been settled.

The order is the whole claim of the GRPI model

Take the order away and the model is left as a list of four words nobody disagrees with. What makes it a tool is its instruction: start at the top, and do not move down until the layer above has been settled.

The reason for the order is that a fault in a higher layer shows up as a fault in a lower one. Conflicting goals held by two departments show up as hostility between two people. Overlapping roles show up as “a recurring misunderstanding”. A process nobody has defined shows up as lapses in following up.

Anyone who deals with what shows and leaves what lies behind it spends effort on signs that come back. One form this takes is a meeting held to improve the relationship between two people whose responsibilities overlap. The meeting ends in mutual goodwill, the overlap remains, and the disagreement returns weeks later, by which time both people believe the problem lies in them.

A word of caution is needed here: this order is a logical argument, not a measurement drawn from data. In the sources we reviewed we found no published figure for how the causes of team breakdown are distributed across the four layers, and we draw no conclusion, for or against, from the absence of that figure.

The first layer: goals

An agreed goal is one a team member can be wrong about when describing it. If five members are asked for their team’s goal and give five different answers, the first layer has not been settled, and there is no point examining the layers after it.

A fault in this layer takes one of two forms: a goal that is missing, or a goal that is present but conflicts with another goal the same member is held accountable for. The second form is the more dangerous, because the team appears to agree in the meeting and then each member acts according to what they are actually measured on. A single question tests for this second form: what number is each member held accountable for at the end of the year?

Consider a team that brings together one person measured on speed of delivery and another measured on the number of errors found after delivery. Each works to what they are held accountable for, and every disagreement between them looks like a disagreement about seriousness or standards. At root, each such disagreement comes from a conflict built into the first layer before the two ever met. A conversation does not resolve it, and neither does clarifying a process. What resolves it is a decision about the goals themselves: which of the two comes first when they collide, and who holds the authority to decide.

The second layer: roles, and a calculation showing how quickly overlap can grow

This layer has only two faults: an overlap, which makes one piece of work belong to two people, and a gap, which makes it belong to nobody. Both are discovered after the damage is done, not before, unless who does what has been written down.

The number of points where an overlap could arise grows faster than the number of members, because it depends on the number of pairs, not the number of people. In a team of n members, the number of pairs equals n multiplied by n minus one, divided by two:

• A team of 4: 6 pairs.
• A team of 6: 15 pairs.
• A team of 8: 28 pairs.
• A team of 12: 66 pairs.

Doubling a team from four to eight therefore multiplies the number of pairs by more than four and a half, from 6 to 28. It follows that what can be left to unspoken understanding in a team of four cannot be left that way in a team of twelve, not because people have changed but because the number of boundaries between them has grown.

The pair count shows how the number of boundaries grows. It does not set an optimal size for a team, or any particular number of members.

One tool for this layer is the RACI matrix, used here not to document the current arrangement but to expose the overlaps and gaps in it. Roles in a team are built on job analysis, and a member’s role in the team should not contradict that member’s job.

The third layer: processes

Processes are the way work moves between members: how a decision is made when they disagree, when they meet and about what, how work passes from one person to another, and what counts as done.

A fault that can arise in this layer lies in the rule for settling decisions. One team decides by consensus without ever having agreed to, so a single objection is enough to hold a decision up for weeks. Another team assumes its leader decides, while the leader is waiting for the team to decide. Each of these faults shows up as a slowdown that gets put down to “weak commitment”.

This layer also covers the choice of leadership style according to the team’s situation and maturity, which is the subject of situational leadership. A process that suits an experienced, independent team does not suit a team that has only just been formed.

The fourth layer: interpersonal relationships

Relationships are examined last, not first: trust, the way people disagree, and what is said in the meeting as against what is said outside it through informal communication. Placing this layer last does not make it less important. The position means that what gets attributed to relationships can originate in the layers above.

If the first three layers are settled and the fault remains, then the fault does lie in this layer, and the remedy is of the same kind: a direct conversation, continuous feedback given at the time, or a change in who is on the team if nothing short of that works. The difference is that the effort is then spent in the right place.

Four situations in which the GRPI model is used

• When a new team is formed. Here the model is used to build rather than to examine: the four layers are written down before work begins, which spares the team months of discovering them by mistake.
• When a team has been reorganised. Roles change and habits do not, so people go on working to boundaries that no longer exist.
• When a visible failure has occurred, such as a missed deadline or work that had to be redone. The value here lies in preventing a jump to blaming one particular person before the first three layers have been examined.
• When a team spans several departments. Members are held accountable in their own departments, not in the team, so a conflict in the first layer can be part of how the team is set up rather than an exception.

By contrast, the model is not used on a functioning team where nobody has raised a complaint. The examination itself raises questions about boundaries that practice has already settled, and it may create a disagreement where there was none. That is a cost with nothing to show for it when there is nothing to fix.

Three things the GRPI model is not

• A scoring measure. The model produces no number to compare teams with or to track from one period to the next. Anyone who turns it into a questionnaire with four headings, each rated on a scale of one to five, removes the thing that sets the model apart, which is the order.
• Organization development. That is a broad field of planned interventions across a whole organisation, whereas the GRPI model is a framework for examining a single team, used in one or two sessions.
• A stage of team maturity. The four layers are not phases a team passes through in sequence. They are points of examination that exist together at all times, and every breakdown is examined starting from the top one.

What undermines a GRPI examination

• Starting at the bottom. This can happen because the fourth layer is the one in plain view, and it is the one people complain about in their own words.
• Using the model where the breakdown is not. A team that lacks the information, the tools or the time will not be fixed by four layers. Its problem lies in resources, not in organisation.
• Letting a party to the disagreement run the examination. Whoever holds the decision on goals finds it hard to see a fault in them, so the examination settles on the lower layers without anyone intending it.
• Ending a session with nothing written down. An examination that does not end in a few agreed lines leaves the team feeling good for two days, and then nothing changes.
• Confusing the examination with an assessment of individuals. What is examined is how the work is arranged, not how individuals perform. Once the session finds its way into anyone’s record, nobody answers frankly in a second session.
• Stopping at the first layer. Rewording the goal at every session can become a ritual that lets a team avoid confronting the overlap in its roles.

The definition of the GRPI model does not establish that the model suits every case. The cases set out above, a team where nobody has raised a complaint and a team short of resources, are ones where it does not.

Before the GRPI model is used in a session

Four questions are asked in order, and the answer to each is written down before moving to the next: what goal the team is held accountable for, who does what at each boundary between two members, how a disagreement is settled when one arises, and what remains once those three are answered.

The answers are then reviewed together, with one question in mind: how many of them had no agreed answer before the session? Where the first or the second had none, the fault has been found, and what was attributed to relationships was an effect, not a cause. If the first three genuinely were settled, the matter lies between particular people, and at that point it is not something a framework resolves.

This is an explanation of the concept, not legal advice.

Related terms

• Organizational Politics
• Competencies
• Informal Communication
• Performance Standards
• Peer Appraisal