Company goals are commitments your organization will be held accountable to over a defined period of time. Not ambitions. Not directions. Commitments with owners, numbers, and a cadence for review.
Most companies treat goals as a planning artifact. They get written at the start of a quarter, shared in a meeting, and then live in a document nobody opens until the next planning cycle. That is not goal-setting. That is theater.
This post is about what makes company goals work: how to write them, how to structure them across teams, and how to make them the center of your operating rhythm.
Company goals and objectives: the difference that matters
A goal is the outcome you want. An objective is the measurable result that tells you whether you got there.
Most companies write goals when they mean objectives, and vice versa. "Grow revenue" is not a goal. It is a direction. "Close $400,000 in new recurring revenue by September 30" is a goal. It has a number, a deadline, and no wiggle room on whether you hit it.
The framework most operators use, borrowed from Intel and Google, pairs each goal with three to five key results. Each key result is a measurable output. The goal answers "where are we going." The key results answer "how will we know we got there."
Where the OKR framing gets misused is in the assumption that you should write twenty of them. Twenty goals means zero accountability. The best-run companies I have seen run three to five company-level goals per quarter. Everything else gets subordinated to those.
Gino Wickman calls these Rocks in the EOS framework (EOS and Rocks are trademarks of EOS Worldwide). A Rock is a 90-day priority that requires dedicated focus to move. EOS gets the priority discipline right. Most teams set too many and protect too few.
If you are running a small company with an AI agent team like we do at Sneeze It, your company-level goals need to be specific enough that you can assign ownership to a seat, human or agent, and measure it weekly. Dirk, our sales agent, owns a revenue goal every quarter. Arin, who manages our call center team, owns an appointment booking rate. These are not approximate aspirations. They are numbers on a scorecard that get reviewed Monday mornings.
Sample business goals that actually hold up in execution
Writing good goals is harder than most planning sessions make it look. Here are patterns that hold up and patterns that fall apart.
Goals that hold up in execution:
"Bring average cost per lead below $35 across all managed accounts by the end of Q3." One metric, one deadline, no interpretation required.
"Reduce average project delivery time from 18 days to 12 days by August 31." Two numbers, one date, one seat owns it.
"Add six new recurring clients in health and wellness by September 30." Six is a number. Health and wellness is a defined vertical. September 30 is a hard date.
Goals that fall apart:
"Improve client satisfaction." There is no number, no deadline, no owner, and no way to know if you achieved it.
"Grow the team strategically." This is a sentence from a deck, not a goal.
"Increase brand awareness." This one is particularly dangerous because it is unmeasurable by design, which means it never dies.
The pattern for sample business goals that survive contact with reality: a number, a deadline, and a single owner with no ambiguity about whether they hit it.
For a service business, the goals that matter most cluster around four areas: new revenue, client retention, delivery quality, and operational efficiency. One goal per area per quarter. Four goals. Ninety days. That is a functional operating cadence.
Organizational goals: how company priorities translate to seats
Company goals do not execute themselves. They get executed by seats, and the seats need to see how the company-level priority connects to what they do every week.
The translation layer is where most organizations lose the thread. The leadership team writes great company goals. Middle layers write their own team goals, often without explicit connection to the company priorities. By the time the work reaches the people doing it, the link is invisible.
The fix is simple: every seat on your org chart lists the company goal it contributes to and the metric it owns in service of that goal. Not a task list. An accountability line connecting the seat's daily output to the company's quarterly commitment.
At Sneeze It, we publish this on our team chart at OTP. Bogdan, our COO, owns operational efficiency metrics tied to the company's delivery time goal. Janine owns the cash collected number tied to the revenue goal. On Monday we can trace any company goal down to the seat that owns the work.
This matters even more when some seats are AI agents. Agents need explicit goal assignments the same way humans do. When Dirk's email count is disconnected from a revenue goal, Dirk optimizes for the wrong things.
This post on unified scorecards goes deeper on how to put human and agent seats on the same goal-tracking surface without creating a management layer between them.
Team goals: the translation from company to weekly work
Team goals sit between the company-level commitments and the individual seat metrics. They are the operational expression of the company priority for a specific function.
A company goal might be "achieve 90% client retention through Q3." The customer success team goal that serves it might be "complete a performance review call with every client over $2,000 monthly spend before August 15." The individual seat metric might be "complete twelve client calls this month."
Three levels. One line of accountability from top to bottom.
The failure mode at the team level is goal inflation. Teams write ten to fifteen goals for their function and then spend the quarter overwhelmed and under-executed. The constraint that works: each team gets two to three goals per quarter, each one tied explicitly to a company priority. If you cannot draw the line from the team goal to a company goal, the team goal is probably maintenance work disguised as a priority.
For teams that include AI agents alongside humans, shared team goals create a useful forcing function. Arin, our call center manager agent, operates against a booking rate goal shared with the human callers on the team. Arin's daily coaching messages and the callers' dial numbers both feed the same target. The goal is the team's, not human or agent separately.
Tally, our scorecard agent, pushes current numbers to our tracking system four times daily. The score updates automatically, and the team reads it at the weekly review.
How to run company goals as an operating rhythm
Goals that only get reviewed at the end of the quarter are not goals. They are retrospective reports.
The cadence that works: weekly, each seat posts its current number and the leadership team talks through anything off pace. Monthly, a thirty-minute status read across all teams. Quarterly, a full review and reset before the next cycle opens.
The weekly rhythm is the one most companies skip. They set goals quarterly, check in monthly, and wonder why execution is inconsistent. When the team knows every Monday that the number will be reviewed, they operate differently all week. That pattern compounds. The quarterly review only matters if the weekly review happened every week between now and then.
The off-track conversation is the same regardless of whether the seat is human or agent: what changed, what is the fix, who owns the fix by next Monday. No separate process for agent goals. Same meeting, same cadence, same accountability.
Frequently asked questions
What is the right number of company goals to set per quarter? Three to five is the range that works for most organizations. Below three, you are probably not being specific enough about what matters. Above five, accountability gets diffuse and the team cannot hold all the priorities at once. Start with three if you are new to structured goal-setting.
How do you write a company goal that is ambitious but still achievable? The test: if you can hit it by continuing exactly what you are doing now, it is a projection, not a goal. If missing it would not change any decision you make, it is not really a priority. Good goals require specific action and carry real stakes.
How often should company goals change? The goal should not change mid-quarter. Tactics can change. If circumstances shift dramatically, you can revisit it, but that should be an exception with a documented reason. Goals that change whenever progress is hard are not goals.
Who should own a company goal? One seat. One person or agent with clear accountability. If two seats share ownership, nobody owns it. When multiple seats contribute, assign the goal to the seat most accountable for the outcome and list the others as supporting.
How do you connect individual performance to company goals? The line runs from company goal to team goal to seat metric. Each seat should be able to point to the company goal their daily work serves. If they cannot draw that line, the goal structure is broken, not the person's performance.
Run it in OTP
OTP tracks company goals, team goals, and seat-level KPIs on a single chart that your whole organization, including AI agents, can query and update. Tally, the scorecard agent, pushes numbers automatically so the score is always current when the Monday review starts.
In Claude Desktop or Cursor or any MCP client, add this block:
"otp": {
"command": "npx",
"args": ["-y", "@orgtp/mcp-server"]
}
Restart the client. Then ask: "Use OTP to show me the current status of all active company goals and which seats own them."