In this guide
- What catchball is for
- How catchball degrades
- Why the record matters later
- What a good catchball record contains
- A worked example: three rounds on one target
- Vertical and horizontal catchball
- Catchball after planning
- Rules for a useful thread
- Catchball anti-patterns
- What catchball needs from the software
- Catchball in Profit.co
- Questions this article answers
Key Takeaways
- →Catchball is the negotiation step lean brought into hoshin planning. It produces an agreed target and the reasoning behind it, and keeping only the number is what makes a missed target unarguable later.
- →It degrades in three stages: real negotiation, feedback window, announcement. By the third, only the vocabulary is left.
- →A good record has six parts: the proposal, the response, the assumptions, the agreement, any dissent and a revisit trigger.
- →Horizontal catchball is where plans fail quietly, because a target that depends on another function’s work gets agreed by people who don’t control that work.
- →The record has to live with the objective. Keeping it in email is a goal setting software gap, not a discipline problem.
- →Two or three rounds over about four weeks. Anything unresolved after that gets escalated, not another round.
Catchball produces two things: an agreed target and the reasoning behind it. Most organizations keep the first and lose the second, which is the part they need when the target is questioned six months later.
What catchball is for
Catchball is the back-and-forth between levels of an organization that turns a leadership intention into commitments people can actually deliver. Leadership proposes an objective or target. The level below responds with what it would take, what it can commit to and what it would need. The proposal comes back revised. The name comes from the image of a ball passed between players until both are satisfied with where it lands.
Done well, catchball does three things. It tests targets against the knowledge of the people closest to the work. It surfaces the assumptions behind each target while they can still be changed. And it creates ownership, because people commit to numbers they helped shape.
All three depend on the exchange itself, not just its outcome. That’s the part that usually goes missing.
How catchball degrades
Under time pressure, catchball tends to slide through three stages.
| Stage | What it looks like | What’s lost |
|---|---|---|
| Real negotiation | Proposals and counter-proposals with reasons, often over several rounds | Nothing yet, but the reasoning lives in meetings and email |
| Feedback window | Targets are circulated with a deadline for objections | Most objections never get raised; silence becomes agreement |
| Announcement | Targets arrive as final, with a meeting to explain them | The method is gone; only the vocabulary remains |
Even at the first stage, the outcome is usually recorded and the exchange isn’t. The X-matrix shows 36%. It doesn’t show that the region argued for 30%, that 36% depended on a new contract model landing by Q2, or that everyone agreed to revisit it if it didn’t.
Why the record matters later
The reasoning behind a target becomes valuable at three predictable moments.
- When the target is missed. The review needs to know whether the assumptions failed or the execution did. Without the record, the conversation becomes a contest of memories.
- When an owner changes. A new regional lead inherits a number with no explanation. They either accept it blindly or reopen the whole negotiation.
- At the next planning cycle. Knowing which assumptions held last year is the best input to this year’s targets.
A target without its reasoning is an instruction. A target with its reasoning is a hypothesis someone can check.
What a good catchball record contains
- The proposal. What was proposed, by whom, and the reasoning offered.
- The response. What the receiving level said it could commit to, and what it would need.
- The assumptions. Conditions the target depends on, stated so they can be checked later.
- The agreement. The final target, who accepted it, and on what terms.
- Any dissent. Disagreement that remained after agreement, recorded rather than smoothed over.
- A revisit trigger. The condition under which the target will be reopened.
None of this needs to be long. A thread of four or five short messages usually covers it. What matters is that it stays attached to the objective it shaped, rather than in an inbox or a meeting note filed somewhere else.
A worked example: three rounds on one target
Here is what a useful catchball thread looks like for one objective: raising services to 36% of sales. It runs between the division head and the North America regional lead.
| Round | Who | Message |
|---|---|---|
| 1 | Division head | Proposing 36% services share for North America by year end. Based on group trajectory to 40% in 2028 and the new service centers opening in Q1. |
| 1 | Regional lead | We can commit to 30% with the centers alone. 36% needs the new multi-year contract model, which legal hasn’t approved. Without it, renewals stay annual and attach rates stay flat. |
| 2 | Division head | Agreed that the contract model is the dependency. Proposing 36% if the model is approved by end of Q1, 31% if not. Adding the contract model as a priority on the matrix, owned by commercial. |
| 2 | Commercial lead | Can get the model to legal by February. Approval by end of Q1 is realistic but not certain. |
| 2 | Regional lead | Accept: 31% committed, 36% if the model is live by 31 March. Need two additional service sales roles from Q2 in either case. |
| Close | Division head | Agreed: 31% committed, rising to 36% if the contract model is live by 31 March. Two roles approved from Q2. Revisit at the April review. |
Illustrative thread. Six short messages, three people, one recorded agreement.
Notice what the thread captures that a final number wouldn’t: the dependency on the contract model, a new priority created to address it, a horizontal commitment from a third function, the resources agreed, and a dated trigger to revisit. If the target is missed in September, the review knows exactly which assumption to check first.
Vertical and horizontal catchball
| Direction | Between | Typical question |
|---|---|---|
| Vertical, downward | A level and the level below | What can you commit to, and what do you need? |
| Vertical, upward | A level and the level above | Here is what our plan delivers; is it enough? |
| Horizontal | Peer functions or sites | Our target depends on your work; can you commit to it? |
Most catchball guidance covers only vertical exchanges. Horizontal catchball is where many plans fail quietly, because a target that depends on another function’s work gets agreed by people who don’t control that work. In the example above, the commercial lead’s commitment is what made the 36% target credible.
Signs of healthy catchball
- Some targets changed. If none moved between the draft and the locked plan, the levels below were informed, not consulted.
- Dependencies were named. Threads mention other functions and what they committed to.
- Revisit triggers exist. Agreed targets include the condition under which they reopen.
- Threads are short. Most settle in two rounds and fewer than ten messages.
Catchball after planning
On paper, catchball is a phase that ends when the matrix is agreed. In practice, the same negotiation is needed during the year: when the review finds outcome lag and has to decide whether to retire a priority, when an external shock makes a target unreachable, or when a correlation grade is disputed.
Running mid-year decisions through the same catchball thread keeps the history in one place. A year later you can read a target’s full life: proposed, negotiated, agreed, questioned, revised.
Rules for a useful thread
- Keep participants few. The proposer, the receiver and anyone whose commitment is needed. A thread with fifteen people becomes a broadcast.
- Write the reason, not just the number. “30%” is a position; “30% without the contract model” is information.
- Move to a meeting when it stalls, but record the outcome back in the thread.
- Close it explicitly. The last message states what was agreed, including any revisit trigger.
Catchball anti-patterns
- The comment period. Targets are circulated with “any concerns by Friday”. Silence is treated as agreement, so the people with the most at stake are the least likely to respond.
- The negotiation by proxy. Delivery leads send their objections through a chief of staff. The reasoning arrives summarised and the dependency details are lost.
- Splitting the difference. Group proposes 36%, the region offers 30%, and everyone settles on 33% without discussing why. Neither assumption gets tested, and the target means nothing.
- The endless thread. Twenty messages and five participants with no agreement. After two rounds, escalate for a decision.
- Catchball in December only. Treating it as a planning ritual, then announcing mid-year target changes without negotiation.
How long it should take
During planning, two or three dated rounds over about four weeks; see Running catchball rounds during planning. Mid-year, a single objective’s catchball should usually settle within one review cycle.
What catchball needs from the software
The six parts of a good record are also a specification. Whatever runs your catchball process has to carry each one, and the differences between products show up here rather than on a feature list.
| What the record has to hold | What that requires of the tool | How to test it in a demo |
|---|---|---|
| The proposal and its reasoning | A thread attached to the objective itself, not a comment on a dashboard | Ask to see a target’s full thread a year after it was agreed |
| The response | Each participant able to counter their own number as structured data | Ask whether a counter-proposal is a field or free text |
| The assumptions | A place for conditions that survives the agreement and can be checked later | Ask where an assumption lives once the target is locked |
| The agreement | Explicit acceptance from every participant before targets commit | Ask whether an administrator can commit targets without it |
| Any dissent | Recorded alongside the agreement rather than resolved away | Ask what happens to an objection that is overruled |
| A revisit trigger | A dated condition that resurfaces at the right review | Ask how a March trigger reaches the April review |
Six rows. Most tools hold the agreed number and none of the rest.
This is the difference between software that stores goals and software that holds how they were agreed. Most goal setting software is built for the first job: a goal, an owner, a due date, a percentage. That is enough for tracking and not enough for hoshin, because the question a review asks is not what the number is but which assumption moved it. Employee goal setting software that cannot answer the second question leaves the reasoning somewhere else, which is where it was already being lost. Goal alignment software is sold on the promise that everyone can see how their work connects upward; the connection people actually ask about later is why the number was set where it was.
It is worth testing early, because it is hard to retrofit. A tool that treats catchball as comments can add fields later. A tool that models agreement as a status flag cannot easily be taught that six people have to accept before a number commits, and that is the behavior the whole method rests on. Put the six rows above into your hoshin kanri software RFP rather than discovering them in the first planning cycle.
The cost of getting this wrong is not the license. A hoshin kanri catchball history does not migrate: threads, assumptions and agreements do not export into another vendor’s data model in any form you would want to read. Strategy execution software you replace in year three takes the reasoning with it, and the organization goes back to arguing from memory. That is an argument for testing this during the evaluation rather than discovering it at the renewal.
Catchball in Profit.co
Catchball in Profit.co runs as a session: a thread, a change table showing every proposed change, and agreement required from every participant before targets commit. Each participant can counter only their own number, and a balance constraint on approval keeps cascaded targets adding up. A meeting can be scheduled from the same place when the exchange needs a room, and the agreed outcome and its reasoning stay with the objective.
The same session model runs A3 review. Because a countermeasure A3 commits other people’s work, everyone whose work it commits has to agree to the revision as sent, and an override is never recorded as agreement. See Who has to agree to an A3. More on the hoshin kanri software overview.
Because Hoshin Kanri, OKRs and Goals and Balanced Scorecard sit together in the Essentials plan, the same session model covers a hoshin target and a team key result without a second subscription or a second agreement workflow to learn. Running hoshin on one tool, OKR or goal management software on a second and balanced scorecard software on a third leaves you with three agreement models and three places the reasoning can go missing.
The 30-day free trial is long enough to run one real catchball thread end to end, which is a better test than any demo. Use it on a target you actually have to agree this quarter, not a sample one.
Free Product Demo
Run your next target negotiation on the record
A 30-minute walkthrough, tailored to your team.
Book a free demo
What happens in the call
How your teams set goals, track progress and review results.
Strategy, performance, projects and engagement on one platform, with AI.
A practical path to align teams, execute faster and drive results.
Questions this article answers
The back-and-forth between levels of an organization in which leadership proposes objectives and targets, the level below responds with what it can commit to and what it needs, and the proposal is revised until both agree.
Because the reasoning behind a target is what the review needs when the target is missed, what a new owner needs when they inherit it, and what next year’s planning needs to set better targets.
Negotiation between peer functions or sites, used when one team’s target depends on another’s work, such as a services target that depends on commercial launching a new contract model.
Two or three during planning, over about four weeks. Anything unresolved after that should be escalated for a decision rather than run through more rounds.
A tool that keeps the thread with the objective, lets each participant counter only their own number, requires agreement from everyone before targets commit, and records dissent. Most goal setting software stores the agreed number and none of the reasoning.
You can run the negotiation that way, but the record ends up in inboxes and meeting notes rather than with the objective. A year later, when the target is questioned, nobody can find the assumption it depended on.
