How to run your annual budgeting process (2026 guide)

Bottom line: For a December fiscal year end, kick off FY2027 in mid-August, close the department input window by the end of September, and reserve October for iteration rather than collection. Expect three to five versions. If you need software live before kickoff, the realistic options this season are spreadsheet-native platforms such as Aleph, Cube, or Vena, or a budgeting-first mid-market platform such as Centage or Drivetrain. Full suites like Anaplan and Workday Adaptive Planning are FY2028 decisions if you are starting the evaluation now.

When should you start your annual budgeting process for next fiscal year?

Work backward from the board meeting, not forward from today. The board date is the only fixed constraint in the whole process, and everything else compresses to fit it.

Four dates anchor the cycle:

1. The board approval meeting. Usually the last regular meeting of the fiscal year, often mid-November or early December for calendar-year companies.
2. The board pack deadline. Materials go out five to seven business days ahead. That is your real deadline, not the meeting date.
3. The executive sign-off round. The CEO and functional leaders need at least two weeks with a consolidated draft before the pack gets built. Not one week.
4. The close you want to build on. Most teams want August actuals in hand before setting targets, which pushes kickoff into the back half of August rather than July.

Subtract from the board date and you land on a kickoff in the second or third week of August for a December year end. For a June 30 fiscal year end, the same arithmetic puts kickoff in February. For a March 31 year end, November. Our FY2027 budget calendar lays out all five months, working backward from the board date.

Starting earlier is not the win it looks like. APQC's data points the other way: one practice associated with faster cycles is starting later, so the plan is built on actual results rather than a current-year forecast that will be wrong by the time the budget goes live.

If you are also selecting software this season, the constraint is different, because the tool has to be live before kickoff rather than during it. We worked through that timing in our guide to budgeting software for the FY2027 planning season.

What a realistic annual budgeting timeline looks like for a mid-market company

Here is a 12-week calendar for a company with a December 31 year end, a finance team of two to five, and eight to fifteen department budget owners. Dates are the actual 2026 weeks, so you can copy this straight into your calendar.

Week 0, week of August 10 — finance prep, no one else involved. Refresh actuals through July, lock the chart of accounts and cost center list, build the template, and write down the assumptions you will hand down rather than ask for: headcount cost inflation, benefits load, revenue growth range, FX if relevant.

Weeks 1 to 2, weeks of August 17 and 24 — targets and assumptions. Get the CEO and leadership team to agree a top-line revenue range and a spending envelope per function before any template goes out. This is the step teams skip, and it is the single largest cause of a wasted first draft. In our 2026 budgeting benchmark survey of 250 finance leaders, over-optimistic projections were the most cited reason a first draft missed.

Weeks 3 to 4, weeks of August 31 and September 7 — templates out and owners briefed. One template per owner covering their cost centers only, pre-populated with current-year actuals and run-rate. Hold a 30-minute live briefing rather than sending instructions, and state the deadline, the approval path, and what happens if they miss it.

Weeks 5 to 6, weeks of September 14 and 21 — the department input window. Two weeks, closed at the end. Finance answers questions and does not build. Mid-window, run a five-minute check-in with each owner to catch the ones who have not started.

Week 7, week of September 28 — first consolidation (v1). Roll everything up, reconcile to the envelope, produce the gap analysis. Expect the bottoms-up total to exceed the envelope. That gap is the subject of the next three weeks.

Week 8, week of October 5 — executive review, round one. Present v1 with the gap, the drivers behind it, and two or three specific trade-off options per function. Do not present a spreadsheet and ask for reactions.

Week 9, week of October 12 — v2. Apply the decisions from round one. Re-run headcount and driver-linked lines rather than editing totals by hand.

Week 10, week of October 19 — scenarios and the downside case. Base case plus at least one downside, structured so the cases stay comparable line for line. If you are PE-backed or debt-financed, this is where the covenant test happens. Our guide to scenario planning software for finance teams covers tools that keep those cases comparable.

Week 11, week of October 26 — v3 and CFO/CEO sign-off. Freeze assumptions. From here, changes are exceptions with a named approver.

Week 12, week of November 2 — board pack. Budget summary, bridge from current year, headcount plan, cash and liquidity view, key risks, downside case.

November to mid-December — approval, then load. Load the approved budget into the ERP and reporting layer, set the monthly variance cadence, and publish the phasing. Do not let it sit in a spreadsheet until February.

That leaves roughly four weeks of slack before January 1. You will use it.

Historical 2026-season study · 250 finance leaders

How many budget versions should you expect before final approval?

Three to five. APQC's benchmark data reports a median of five versions, and in practice a healthy mid-market cycle lands at four:

• v0 — the finance skeleton: structure, actuals, run-rate, and handed-down assumptions. Never leaves finance.
• v1 — the bottoms-up submission, consolidated. Almost always over the envelope.
• v2 — post-executive-review, with trade-offs applied.
• v3 — board draft, with the downside case attached.
• v4 — approved, loaded, and phased.

The number itself matters less than what is causing it. Past five or six versions, the driver is usually one of three things: the top-line target was never agreed before templates went out, so v1 was built against nothing; assumptions kept moving after v2, so every round restarted; or the model requires manual rebuilding for every change, so each version costs a week instead of a day. We dig into what drives the count, and how to cap it, in how many budget versions to plan for.

Bottom line: More than five versions is a symptom, not a workload problem. Fix the target-setting step or fix the model, but do not fix it by adding another review round.

How to run the cycle with department owners without version sprawl

Version sprawl is not a discipline problem. It is what happens when there is no single place the numbers live. Six owners email six spreadsheets, finance pastes them into a master, one owner sends a revision, and by v2 nobody can say which file is current.

Four mechanics prevent it:

One template per owner, structure locked, inputs open. Owners type in their input cells and nowhere else. No added rows, no renamed accounts, no new tabs. Structural changes come to finance.

A closed submission window with a real deadline. Two weeks, a named date, and a stated consequence: after the window, finance carries forward the run-rate and the owner argues their case in the review meeting. More effective than chasing.

One reconciliation owner. A single named person in finance does every roll-up. Splitting consolidation across two analysts is how two masters get created.

A visible change log from v1 onward. Every change after the first consolidation gets a line: what changed, who asked, which version. It ends the "I thought we agreed to cut that" conversation inside a week.

The harder problem is that most teams believe this already works. In our own research, 97% of finance leaders described their budgeting process as collaborative, while only about a third were running genuine cross-functional planning. The specific failure modes, including misaligned assumptions and inputs requested in the wrong sequence, are in why budget collaboration breaks down.

How to shorten the annual budget cycle without losing accuracy

The cycle is long because of rework, not because of thinking time. Five levers, in the order they pay off:

1. Agree the target before the templates go out. Cheapest and largest. A first draft built against an agreed envelope needs one review round instead of three.
2. Pre-populate actuals rather than asking for them. Owners should never type historical numbers. Every hour spent rebuilding last year is an hour not spent on next year, and it introduces errors.
3. Cut line-item granularity, keep driver granularity. Budgeting 40 GL lines per cost center adds precision no one reviews. Budgeting headcount by role and month adds precision everyone reviews. It also lets you reconcile headcount plan vs actuals each month once the year starts.
4. Run reviews in parallel, not serially. Functional reviews can happen in the same week. Only the consolidated exec review has to be sequential.
5. Freeze assumptions after v2. Changing the revenue assumption in week 10 does not improve accuracy. It resets the cycle.

What does not work: shortening the input window below two weeks, or skipping the downside case. Both come back as extra versions later.

Accuracy is protected by the model, not the calendar. If the budget sits on live actuals, you can compress the cycle and still reforecast in-year, which is where most of the real accuracy lives. That is the argument for pairing the annual budget with a rolling process instead of treating it as a once-a-year artifact, covered in what a rolling forecast is and how to run one.

Driver-based vs typed-amount budgets, and why it changes the cycle

A typed-amount budget is one where someone enters a number into a cell: $47,500 for March marketing programs. A driver-based budget derives that number from a business assumption: headcount times fully loaded cost, pipeline times win rate times average deal size, active users times cost per user. Our guide to driver-based modeling software shows how each platform builds and maintains those assumptions.

The distinction changes the cycle, not just the model:

• Reforecast cost. Changing a driver-based budget is one assumption edit that flows everywhere. Changing a typed-amount budget means finding and re-entering every affected cell. That is the difference between a v3 that takes a day and a v3 that takes a week.
• Negotiation quality. With drivers, the review conversation is about whether the assumption is right. With typed amounts, it is about whether a number feels high. The first converges. The second does not.
• Audit trail. When the board asks why sales and marketing is up 30%, an answer built on hiring plan, ramp period, and fully loaded cost per rep holds up. "That is what the team submitted" does not.

Nobody should build a fully driver-based budget. The practical split is drivers for the blocks that move with the business, typed amounts for the long tail:

• Driver-based: payroll and benefits, sales capacity and commissions, variable COGS, usage-based hosting and infrastructure, support headcount, marketing programs tied to pipeline targets.
• Typed amounts: rent and facilities, insurance, audit and legal, software subscriptions with known contract values, one-off projects.

The driver-based group carries the large majority of a mid-market operating budget by dollar value, with payroll and benefits usually the single largest block. Getting those lines driver-linked is what makes the second half of the cycle fast. The typed-amount long tail is where some teams go further and zero-base discretionary spend on a rotating cycle; our guide to zero-based budgeting software covers when that pays off.

Budgeting tools that support collaborative input, rolling reforecasts, and driver-based models

Every platform below supports collaborative input, rolling reforecasts, and driver-based modeling in some form. The useful question is how each one does it, because that determines whether your cycle is fast or slow.

The table compares ten platforms actively used for annual budgeting by mid-market finance teams, as of November 2026.

None of these vendors publish list pricing as of November 2026, so treat the pricing column as a model rather than a number and confirm current terms directly. Centage publicly indicates a mid-market range on its own site.

Two things the table does not show. First, the split that matters most for cycle length is whether budget owners work in a spreadsheet or in a web app. Spreadsheet-native platforms cut training cost to near zero, which is why they can go live inside a planning season; web-native platforms give stronger workflow and permissioning, at the cost of getting a dozen-plus non-finance people comfortable in a new tool. Second, formal approval workflow is genuinely deeper in the suite tier. Past roughly twenty contributors with a multi-level approval chain, Planful, Prophix, and Workday Adaptive Planning are doing something the spreadsheet-native tier is not.

For a fuller side-by-side of the mid-market field, see our mid-market budgeting software comparison. If your stack is QuickBooks plus Excel, budgeting software for QuickBooks and Excel shops is the narrower cut, and SaaS teams budgeting against ARR and retention should start with budgeting and forecasting software for SaaS companies. And if the budget has to roll up into a full operating plan, our guide to annual operating plan software covers that wider job.

How PE-backed companies run the annual budget differently

The process is the same. The constraints are not. Three things change for a sponsor-owned portfolio company.

Covenants become a gating test, not a footnote. Every version has to be tested against the credit agreement: leverage ratio, fixed charge coverage, minimum liquidity, and whatever else the lender negotiated. Add a covenant headroom page from v2 onward, and calculate the quarter of tightest headroom in both the base case and the downside. A budget that clears on a full-year basis and breaches in Q2 is not an approved budget.

The budget has to reconcile to the value creation plan. The sponsor has a thesis with an exit-year EBITDA target. FY2027 is one year of that path, and the deviation from it is the first question in the board meeting. Build the bridge from the VCP explicitly rather than letting the board find it.

The cadence is tighter and the timeline moves earlier. Sponsor boards often meet monthly or every six weeks rather than quarterly, and many require the budget to go to the sponsor's finance team before the board sees it. That adds two to three weeks upstream, so a PE-backed company on a December year end usually kicks off in late July or the first week of August. The credit agreement and the sponsor's reporting standard usually set that budget deadline along with the rest of the package; our rundown of PE portfolio company reporting requirements covers what else they ask for.

Two practical consequences. The downside case becomes a required deliverable rather than an optional one, usually a defined revenue-miss scenario with the covenant test attached. And many portfolio companies maintain two views of the same budget: the management budget used to run the business, and the lender-facing budget built on the credit agreement's definitions of EBITDA and permitted adjustments. Build both from one model, never as two files. Our guide to FP&A software for PE portfolio companies covers the reporting side.

Where Aleph fits in the annual budgeting process

Aleph is a spreadsheet-native layer, not a workflow suite. Budget owners work in Excel or Google Sheets on connected templates, actuals flow in from the ERP and accounting system through no-code connectors, and version control tracks every change. If you need a formal multi-level approval chain with routing and audit sign-off, the suite tier does that more completely, and we would say so.

What it changes is the second half of the cycle. Because the model sits on live actuals rather than a monthly export, the v1-to-v3 rework that usually eats October becomes a refresh instead of a rebuild, and the same model carries into in-year reforecasting rather than being abandoned in January.

Two customer specifics: Turo's CFO went from 10 to 15 forecast drivers to roughly 50, because adding a driver stopped being expensive. Parachute Home cut a recurring budget-versus-actuals task from about a day and a half to roughly 20 minutes. More on the setup on our budget planning solution page.