Annual operating plan (AOP): how to build one and the software to run it

An annual operating plan (AOP) is the one-year, department-by-department translation of your strategy into committed numbers: revenue by segment, headcount by team, spend by category, and the operating metrics each owner is accountable for. It is narrower than a strategic plan and wider than a budget — the budget is the money view of the AOP, and the AOP is what makes the money numbers mean something.

For software, the honest answer is that almost nobody sells a product called AOP software. What you are choosing is a planning platform that can hold a driver-based plan, collect inputs from every owner, and reconcile the plan against actuals all year. Aleph fits when finance wants that plan in Excel and Google Sheets on live data; Vena, Planful and Prophix fit when structured workflow across many contributors is the priority; Anaplan and Workday Adaptive fit at the top of the market.

Bottom line: an AOP is not a longer budget. It is the plan that connects a strategic target to the drivers and owners who have to deliver it, which is why the tool you pick matters less than whether the plan is built on drivers people can actually flex.

AOP vs budget vs strategic plan

These three get used interchangeably in most companies and mean different things. Getting the distinction right saves an argument in every planning cycle.

The AOP is the layer that breaks most often, because it is the only one that requires both finance and the business to agree on cause and effect. A budget can be internally consistent and still be nonsense if nobody agreed what drives the revenue line.

What belongs in an AOP

Six components, and the fourth is the one most teams leave out.

1. The top-line target and its guardrails. If leadership wants 30% growth with no more than 15% year-over-year opex growth, that constraint has to be written down per department before anyone plans.
2. A driver-based revenue plan. Pipeline, capacity, conversion and price — not a growth percentage applied to last year.
3. A headcount and hiring plan with start dates. Timing matters more than the total; a hire in month two and a hire in month ten cost very different amounts.
4. Named owners with accountable metrics. Every line has a person, and every person has one or two metrics they are actually judged on. This is what turns a budget into an operating plan.
5. Monthly phasing. An annual number with no phasing cannot be compared to actuals until the year is over.
6. Stated assumptions. Written down, dated, and owned, so that when they change you know what to re-run.

Points two and three are where driver-based budgeting and headcount planning do the work. Point six is the cheapest and most skipped: teams that publish assumptions with owners attached spend materially less time in revision rounds, because a changed assumption becomes a targeted update instead of a rebuild.

How to build one, in sequence

The order is not negotiable. Every step out of sequence creates a rebuild.

Agree the target first. Finance translates the CEO's ambition into department guardrails. Until this exists, bottom-up planning is guaranteed to miss.

Build the baseline from actuals. Start every owner from last year's real numbers, classified into committed, discretionary and one-time. Committed lines should not be up for debate; discretionary lines are where the conversation belongs.

Model revenue from drivers. Then derive the cost lines that depend on it, so a change in the revenue driver flows through instead of being re-entered.

Collect department plans against guardrails. Owners review what is already there and answer the questions that genuinely need a human, rather than filling a blank sheet.

Consolidate and close the gap. The gap to target is normal — every plan starts with more asks than money. Finance's job is deciding who absorbs the cut and in what order.

Lock, phase and publish. Then set the reforecast cadence before the year starts, not in March when the plan is already wrong.

Our annual budgeting process guide covers the mechanics of the middle three steps in more detail, and the FY2027 calendar puts dates against them.

Software for annual operating planning

Judge candidates on four things: can it hold a driver-based plan, can it collect from every owner without email, can it phase and compare against actuals continuously, and can finance change the model without a vendor.

Aleph holds 4.9 out of 5 from 108 reviews on G2 against an FP&A category average of 4.55, and its customer base includes Zapier, Notion, Turo and Y Combinator. Where it is not the right answer: if you need statutory consolidation with intercompany eliminations, that belongs in dedicated consolidation software, and if your plan is genuinely multi-dimensional across dozens of practices, the enterprise tier earns its cost. Our breakdown by company size sets out the tiers, and the pricing guide has the cost ranges.

The two AOP failure modes

Numbers with no drivers. A plan where every line is a hard-coded value looks finished and cannot answer a single question. When the CEO asks what happens if we hire six months later, the answer takes a week instead of a minute. The test is simple: change one assumption and see whether the plan moves on its own.

No owner on the line. If a line has no name against it, nobody defends it in a review and nobody notices when it drifts. Assigning owners is the cheapest quality improvement available in planning, and it is what separates an operating plan from a spending limit — the point collaborative budgeting turns on.

Both failures show up the same way in month three: the plan gets quietly abandoned as a management tool while remaining the official number. That is also why rolling reforecast capability matters more than it used to. Only a small minority of teams now lock the budget and walk away, and 40% reforecast quarterly or monthly, up from 25% the year before, according to our 2027 budget season survey. An AOP you cannot re-run is an AOP with a short shelf life.

If you are standing this up for the first time, the sequence in the implementation steps guide and the honest cost picture in what implementation actually costs are the two things worth reading before you talk to a vendor. The demo questions guide covers what to ask them.

What the drivers look like by function

Driver-based sounds abstract until you write them down per team. These are the ones that hold up in practice.

• Sales. Reps by segment, ramp time to full productivity, quota, attainment rate, and pipeline coverage. Headcount times quota times attainment is the plan; the ramp curve is what makes the phasing honest.
• Marketing. Spend by channel, cost per lead, lead-to-opportunity conversion, and pipeline contribution. The temptation is to plan marketing as a percentage of revenue, which makes it a residual rather than a driver.
• Customer success. Accounts per manager, gross and net retention, and expansion rate. Retention belongs in the revenue plan, not as an afterthought.
• Engineering and product. Headcount by level, contractor mix, and the split between run and build. The capitalization policy sits here and materially changes reported margin.
• General and administrative. Mostly headcount-driven plus a per-head cost for software, facilities and benefits. A per-head cost that scales is more accurate than a fixed total.

The test of a driver model is whether changing one input moves the plan on its own. If the CEO asks what happens if the two enterprise reps start in July instead of March, a driver model answers in a minute and a hard-coded plan takes a week. That difference is the entire argument for building it this way, and it is what driver-based budgeting software is for.

Who owns what

An AOP has three roles and the boundaries are worth stating out loud, because most planning friction is a boundary dispute nobody named.

Leadership owns the target and the tradeoffs. They decide the growth rate, the margin constraint, and which bets get funded when there is not enough money for all of them. What they should not be doing is arbitrating line items in a department they do not run.

Finance owns the model, the guardrails, the consolidation and the story. Finance translates the target into constraints, challenges assumptions that do not hold, keeps the running tally of who can absorb a cut, and stands behind the final plan in the board meeting.

Department heads own their inputs and the commitments that follow. They say what they need and why, and once the plan is locked, their number is a commitment rather than a request. The reason to make owners explicit is that a line with no name against it gets defended by nobody in review and drifts unnoticed all year.

AOP for multi-entity and PE-backed companies

Two variations change the shape of the plan enough to mention.

Multi-entity groups need the AOP to roll up across entities with different charts of accounts, and often different currencies. The practical requirement is mappings finance can edit without a vendor ticket, because an acquisition should be an addition rather than a rebuild. If statutory eliminations are in scope, that is a consolidation requirement rather than a planning one, and most planning tools do not do it.

PE-backed companies plan against a sponsor's reporting standard as well as their own. The AOP has to produce the sponsor's KPI schedule and the covenant view alongside the internal plan, on the sponsor's calendar. Our guides to FP&A software for PE portfolio companies and what reporting PE firms require cover the specifics, and the short version is that the deadline is external and non-negotiable, so the plan has to be producible early rather than perfect late.

From AOP to reforecast

The AOP's shelf life is shorter than it used to be. Only a small minority of teams now lock the budget and walk away, and quarterly or monthly reforecasting rose from 25% to 40% of teams year over year in our 2027 budget season survey.

So build the AOP to be re-run. Concretely: keep drivers as inputs rather than hard-coded values, keep assumptions in one visible place with owners and dates attached, and decide the reforecast cadence before January rather than in March. A plan you can re-run in an afternoon stays a management tool; a plan that takes three weeks to update becomes a historical document by February.