Leadership Team Alignment Is Not Agreement: The Parallax Test
Most leadership teams I work with describe themselves as aligned, and they can point to the evidence. The offsite ended with a strategy everyone signed. The weekly meeting rarely produces open conflict. When the CEO asks “are we all on the same page?”, every head nods.
Then a decision lands between sales and delivery, and three people act on three different understandings of who was supposed to close it. Nobody lied in the meeting. Each of them described the team accurately, from where they sit.
That is the problem with how leadership team alignment usually gets checked. It gets tested in the room, out loud, with the CEO present, which is the one setting guaranteed to produce agreement. Agreement and alignment are different things. Agreement is what people say together. Alignment is whether, asked separately, they describe the same team: the same members, the same priorities, the same owners, the same routes for a dispute. You can measure the second one in an afternoon, and in my experience it is the one that rarely gets measured.
- Leadership team alignment is whether the members of a leadership team, asked separately, describe the same decision system. Agreement in a meeting is a different measurement, taken in the one setting most likely to flatter it.
- The Parallax Test: five questions, answered alone and in writing by the CEO and every executive before anyone discusses them, then laid side by side. Where the answers diverge, the structure behind that question exists only in people's heads.
- In the research behind Senior Leadership Teams, Hackman reported that almost every senior team thought it had set unambiguous boundaries, yet when members described their team, “fewer than 10% agreed about who was on it”.
- Donald Sull and Rebecca Homkes surveyed 11,000 senior executives, leaders and managers in over 400 companies: only about a third could list their firm's top three priorities, and among those involved in developing the strategy the figure was only slightly more than half.
- Each of the five questions maps to one part of the decision system: membership, priorities, Decision Clarity, Escalation Discipline and Decision Authority Dependency. The question with the widest spread tells you where to work first.
- Close the gap in writing, not in a workshop: turn the widest-spread question into a one-page decision, circulate it, and run the same five questions again next quarter.
What does leadership team alignment actually mean?
Leadership team alignment is the degree to which the members of a leadership team, asked separately, give the same description of how the team decides. It is a property of the structure, observable in written answers, not a mood you can read from the room.
Most definitions on the first page of a search treat alignment as shared commitment to a vision, a strategy or a set of values. That layer is real, and offsites, strategy documents and values workshops already address it. The layer underneath is plainer: if the strategy says “grow enterprise accounts”, who decides whether a particular enterprise deal gets the custom delivery terms it asks for? If the head of sales and the head of delivery disagree, where does that go, and how long does it get to sit there?
A leadership team can share the vision completely and still hold six private versions of the answers, each built from that person's experience of past decisions and each feeling like a fact rather than an interpretation. The CEO's version is usually the system as designed, or as the CEO remembers designing it. The executives' versions are the system as they experience it every week. In my reading, the gap between those columns is where execution leaks.
This is why I treat alignment as structural rather than relational, the same line I draw between relational and structural team diagnostics. A team can score well on trust and cohesion and still be misaligned on who decides what, because nobody ever wrote that down.
Why does an offsite produce agreement but not alignment?
Because the room measures the wrong thing, and it measures it under the conditions most likely to distort it.
Agreement is produced socially. Somebody, usually the CEO, states a position, and if nothing seems worth objecting to, the others nod. A nod in that setting means “I have no objection I want to raise in front of this group right now”. It does not mean “my understanding of who decides this is identical to yours”, and nobody can tell the difference, because nobody has stated their own understanding. They have only reacted to someone else's.
Richard Hackman and his colleagues documented how wide that hidden difference can be in senior teams. In the research behind Senior Leadership Teams, Ruth Wageman, Debra Nunes, James Burruss and Hackman studied more than 120 top teams around the world. Describing that work in a 2009 interview with Diane Coutu, Hackman said that almost every senior team they studied thought it had set unambiguous boundaries, yet when members were asked to describe their team, “fewer than 10% agreed about who was on it” (Coutu, 2009).
The most basic fact about a team, its membership, was not shared, and the teams believed it was. The sample is senior teams gathered internationally for a book published in 2008, and I would not carry the proportion into a sixty-person company in Cluj. What I take from it is narrower: in that sample, the teams' confidence about their own boundaries ran well ahead of the fact.
The same pattern shows up one level down, in priorities. Donald Sull and Rebecca Homkes surveyed 11,000 senior executives, leaders and managers across more than 400 companies, in research run through London Business School; it is not a Romanian sample. Only about a third could list their firm's top three priorities. Among the managers involved in developing the strategy, the share was only slightly more than half (London Business School, 2015).
The second figure is the one I keep coming back to. These are the people who wrote the strategy. If authorship did not produce a shared list for them, I would not expect a strategy offsite to produce one either. It produces a document everyone agreed to, which is a different object from a description everyone holds.
The Parallax Test: five questions, answered alone
Parallax is what happens when you look at one object from two positions: it appears to sit in a different place against the background. Nobody is wrong about the object; they are standing in different places. The test is designed to make that visible instead of arguing it away.
Start with the mechanics, because they are what keeps this from turning back into a meeting.
- Everyone answers alone, in writing, before any discussion. The CEO included, in a separate column.
- Answers are specific. Names, not functions. One named decision, not “strategic decisions”.
- Nobody sees anyone else's answers until all are in. Otherwise the first answer becomes the anchor and you are measuring agreement again.
I give each person a fixed ten minutes to write. That is a design choice, not a finding: I want the answer they work from day to day, not a position drafted for the exercise. The five questions:
- Who is on the leadership team? Write the names. Not the roles, not “the exec team”, the names.
- What are this quarter's top three priorities, in order? Ranked, because three unranked priorities leave the team to discover the ranking in the first conflict between them.
- Who closes [a named recurring decision]? Pick one real decision that comes up every month: a discount above standard, a hire into an existing team, a change to a delivery date a client has already seen. One name per answer.
- When sales and delivery disagree about [a named situation], where does it go, and how long can it sit there? Use the two functions that actually collide in your company. The second half of the question matters as much as the first.
- Which decisions does the CEO still make that, on paper, belong to someone else? Everyone answers this one, including the CEO.
Then lay the answers side by side, question by question, with the CEO's column separate from the rest.
Here is a pattern illustration of what typically comes back. It is a composite of what I see in engagements, not one company's result and not a measured outcome. Question one returns more than one list: someone includes the head of finance, who attends most meetings; someone leaves out the head of people, who attends when invited. Question two returns the same three priorities in different orders. Question three is split, with two executives naming each other. Question four returns “the CEO” from most executives and “they should sort it out between them” from the CEO. Question five is where the CEO's column stands apart from everyone else's.
None of the answers is a mistake. That is the point of calling it parallax: the fix is not to correct anybody's perception, it is to design the thing they were each estimating.
What do the gaps in the answers tell you?
Each question reads one part of the decision system, so the spread on each question points to a different repair.
Question one, membership, reads the team's boundary. If the members do not share a list, every other answer is unstable, because people are describing different teams. This is the gap Hackman's research points at, and it is cheap to close: someone with the authority to decide the membership writes it down and says who attends by invitation only.
Question two, priorities, reads whether direction was translated into an order. Disagreement on the three items is a strategy communication problem. In my experience, agreement on the items with disagreement on the order is the more common result, and it matters more, because the ranking is exactly what a leadership team needs when two priorities compete for the same people.
Question three, the named closer, reads Decision Clarity. If two people name each other, or three names come back for one decision, the decision has no designed owner and is being closed by whoever moves first. The decision rights layer is where this gets fixed, one recurring decision at a time.
Question four, the dispute route, reads Escalation Discipline. Two failure shapes show up here. The executives all name the CEO while the CEO expects them to resolve it, which means the route exists only in the executives' practice. Or nobody names a time limit, which means a dispute can sit indefinitely without anyone having refused anything. Cross-functional decision rights are where the seams between two functions get an owner and a clock.
Question five reads Decision Authority Dependency. In the Leadership Architecture model this is a moderator, scored alongside the five dimensions rather than inside them, because it describes how far the real flow of decisions depends on one person, whatever the design says. It is the question where I most often see the CEO's column differ from everyone else's: the CEO lists one or two decisions, and the executives list several more, often the same ones. That difference is not a disagreement to be resolved by vote. It is information about how authority actually moves, and the CEO is badly placed to see it, because the decisions arrive at their desk by courtesy rather than by request. The Delegation Envelope is the repair: written limits, so a delegated decision does not drift back.
The full Leadership Architecture OS diagnostic formalises the last step. It scores the CEO and the executive team separately on all five dimensions and reports the gap on each dimension, with their mean as the Alignment Gap Index, alongside the team score. The Parallax Test is the paper version of the same idea. It will not give you a score, and it does not need to: the spread on five questions is enough to tell you where the work is.
How do you close an alignment gap without another offsite?
In writing, one question at a time, starting with the widest spread.
- Pick the question where the answers diverged most. Not the one that caused the most heat in the debrief. The one where the written answers were furthest apart.
- Name who decides the answer. Membership is decided by the CEO. A closer for a recurring decision is usually decided by the CEO with the two functions involved. Say this before the conversation starts, or the conversation becomes a negotiation over who gets to settle it.
- Write the answer on one page. The member list with who attends by invitation. The three priorities in rank order. The named closer for each recurring decision on the list. The dispute route with a time limit attached. The decisions the CEO is handing back, with the limits written into each.
- Circulate it and let it be challenged once. One round of written objections, with a date. After that date, the page is the answer and the next disagreement goes to the person who owns that question.
- Run the same five questions again next quarter. Same questions, same method, alone and in writing. You are not looking for perfect agreement. You are looking at whether the spread on the question you worked on narrowed, and whether a new gap opened somewhere you did not look.
Picking one question means leaving four untouched for a quarter. Do it anyway. I would rather have one page the whole team holds than five pages negotiated in a day and remembered differently by the end of the week.
The two-question decision rights audit checks a single line between two people. The Parallax Test asks whether the whole team holds one picture of itself. If you suspect one line, start with the audit; if you suspect the picture, start here.
So, the provocation for this week. Before your next leadership meeting, ask every member to write the names of the people on the leadership team on a card, alone, and hand it in. If the cards match, you have earned the nod in the room. If they do not, the agreements that team reached this year were reached by people who were not describing the same team.
Frequently asked questions
What is leadership team alignment?
Leadership team alignment is the degree to which the members of a leadership team, asked separately, give the same description of how the team decides: who is on it, which priorities come first, who closes each recurring decision and where a dispute between functions goes. It differs from agreement, which is what members say together in a meeting. Alignment can be checked in writing; agreement can only be observed in the room.
How can you tell if your leadership team is aligned?
Ask each member, including the CEO, to answer the same five questions alone and in writing before any discussion: the team's membership, this quarter's top three priorities in order, who closes one named recurring decision, where a named cross-functional dispute goes and how long it can sit, and which decisions the CEO still makes that belong to someone else on paper. Lay the answers side by side. Divergence on any question shows a part of the decision system that was never written down.
Why do leadership teams think they are aligned when they are not?
Because the check usually happens out loud, in a meeting, with the CEO present, where a nod means no objection rather than a shared understanding. In the research behind Senior Leadership Teams, Richard Hackman reported that almost every senior team studied thought it had set unambiguous boundaries, yet when members described their team, fewer than 10% agreed about who was on it. In that sample, the teams' confidence in their own boundaries ran well ahead of the fact.
What is the Parallax Test?
The Parallax Test is a five-question exercise for leadership teams. The CEO and each executive answer separately and in writing, the answers are compared side by side with the CEO's column kept apart, and the question with the widest spread shows where to work first. Each question maps to one part of the decision system: team membership, ranked priorities, Decision Clarity, Escalation Discipline and Decision Authority Dependency.
How do you fix misalignment in a leadership team?
Work on one gap at a time, in writing. Take the question where the written answers diverged most, name who decides the answer, write it on one page, allow one round of written challenge with a deadline, then treat the page as the answer. Rerun the same five questions the following quarter to see whether that spread narrowed. I would not reach for another offsite: it produces agreement, and agreement was never the missing piece.
Sources
Wageman, R., Nunes, D. A., Burruss, J. A. & Hackman, J. R. (2008), Senior Leadership Teams: What It Takes to Make Them Great, Harvard Business School Press
Coutu, D. (2009), Why Teams Don't Work, interview with J. Richard Hackman, Harvard Business Review, May 2009
London Business School (2015), Two thirds of senior managers can't name their firm's top priorities, reporting the survey by Donald Sull and Rebecca Homkes
Client examples are anonymized composites; no figures are invented.
If your team answered these five questions alone tonight, how many versions would come back?
Five minutes. No account. A structural read on whether your decision system, not your team, is the real constraint.