Zero-Based Budgeting for Mid-Market Companies

Executive summary

Zero-based budgeting (ZBB) starts each budget cycle from a blank sheet. Instead of taking last year's actual, adding a growth percentage, and arguing about the increment, every cost has to be justified against the work it pays for. The appeal is obvious: incremental budgets lock in whatever was spent before, including work nobody asked for any more.

The full textbook version, where every cost center in the company rebuilds its budget from nothing every year, is too heavy for most mid-market finance teams. It consumes weeks of management time, and the quality of the answers depends on managers who may have no spare capacity. The version that works at this size is scoped and repeatable:

1. Choose the cost pools where inherited budgets are most likely to hide waste (usually SG&A and indirect spend).
2. Ask each owner to describe the work as a set of ranked decision packages, each with a cost and an outcome.
3. Rank all packages together and draw a funding line based on the total envelope.
4. Record what was cut, what was kept, and why, so next year's cycle starts from the decisions rather than from memory.

This guide gives the method, the diagnostic that tells you whether ZBB is worth doing, templates for the package and the ranking, a worked illustrative example, and the mistakes that cause a ZBB program to collapse into a one-time cost cut that grows back in 18 months.

When to use this

ZBB is worth the effort when:

• SG&A has grown faster than revenue for two or more years and nobody can explain the difference by activity.
• A budget review consists of debating the increment, never the base.
• You have a sponsor, lender, or board expecting a cost program and you want one that is defensible by activity rather than a flat percentage cut.
• You have just integrated an acquisition and two cost bases now overlap.
• A new CEO or CFO needs to learn what the base actually pays for.

It is a poor fit when the real problem is revenue, when the finance team cannot produce cost by function from the ledger, or when leadership will override the ranking whenever it is inconvenient. In those cases start with a lighter tool: a rolling forecast that reforecasts the base more often, or a span of control review that tests the structure.

Diagnostic questions

Answer these before committing to a ZBB cycle. They tell you which cost pools to reopen and whether the data will support it.

1. What share of total cost sits in SG&A and indirect spend, and which three functions are the largest?
2. Over the last three years, which cost lines grew faster than revenue, and is there an activity-based explanation for each?
3. Can you produce actual cost by function and by natural account (salary, contractors, software, travel) from the ledger without manual rework?
4. How is the current budget built: last year's actual plus a percentage, or a bottom-up plan?
5. When was the last time a cost line was removed entirely, and who made the decision?
6. Which costs are contractually committed through the budget year, and which are discretionary?
7. Do budget owners know what each of their cost lines buys, or do they know only the number?
8. Is there a time of year when managers have the capacity to build packages (not during close or audit)?
9. Who will decide when two functions both claim a priority and the funding line cannot cover both?
10. Are there constraints on the cost base outside finance's control, such as employment obligations, regulatory staffing, or customer contracts?
11. What happened to the last cost program: did savings hold after twelve months?

Reading the answers: if question 3 is "no", fix reporting first because the packages will be built on bad numbers. If question 6 shows most of the base is committed, ZBB will mostly produce a smaller discretionary pool, and a contract-renewal calendar will be a better tool. If question 11 shows savings did not hold, the missing piece is governance, so spend more of the plan on the last stage.

The framework: scope, packages, ranking, funding line

Scope

Do not reopen everything. Rank cost pools by two things: size, and how much of the pool is discretionary. A pool that is large and mostly discretionary (marketing programs, professional fees, travel, software subscriptions, facilities services) is the best candidate. A pool that is large but fixed by contract or by regulation is a candidate for renegotiation at renewal rather than for ZBB.

Many mid-market companies run ZBB on two or three pools per cycle and rotate. Pools like payroll in a function are covered by the package structure indirectly, but headcount decisions themselves sit with HR and legal counsel, as described later.

Decision packages

A decision package is the unit of ZBB. It describes one activity or service, at one level of provision, with its cost and its consequence. A good package has:

• A name that describes the work, not the department (for example "Month-end close and statutory reporting").
• The cost, split into people, third parties, and other.
• The output: what the business gets, who uses it, and how often.
• Alternative levels: a minimum level, the current level, and an enhanced level, each with a cost.
• The consequence of not funding it.
• Any contractual or legal reason it cannot be reduced.

The alternatives matter because ZBB is not a binary keep-or-cut test. The useful question is often "what does the next level of service cost and what does it buy?"

Ranking

Rank every package, across all functions, in one list. Use a small number of criteria that your leadership team agrees on before anyone submits a package. A workable set:

Required packages are ranked first and funded automatically. The remaining packages are scored by the leadership team together, not by each function on its own work.

Funding line

Add up the packages from the top of the ranked list until you reach the envelope. The envelope is a number set by the CFO and CEO from the plan: a target margin, a covenant constraint, or an investment plan. Packages above the line are funded, packages below are not. The packages just below the line are the ones to discuss, because that is where judgment matters.

The funding line replaces the debate over the increment with a debate over what is just above and just below the line.

Step-by-step process

1. Set the envelope and scope. Decide the total cost target and which pools are in. Owner: CFO with CEO. Output: one-page scope and envelope.
2. Agree the ranking criteria before any packages exist. Owner: leadership team. Output: criteria and weights, approved in writing.
3. Build the cost baseline by function and natural account from the ledger, with commitments tagged. Owner: FP&A. Output: baseline file where every line is either committed or discretionary.
4. Train budget owners with the package template and one completed example. Owner: FP&A. Output: a 45-minute session and a template per owner.
5. Owners write packages. Each owner submits packages at a minimum, current, and enhanced level. Owner: each budget owner. Output: completed package forms.
6. FP&A challenge. Check that package totals tie to the baseline, that nothing is double counted across functions, and that unexplained costs are not hidden in a catch-all package. Owner: FP&A. Output: reconciled package list.
7. Rank and set the funding line in a leadership session. Owner: CEO chairs, CFO facilitates. Output: ranked list with a line, and a decision on each package near the line.
8. Plan execution for each unfunded or reduced package: date, owner, one-time cost, dependencies. Anything affecting employees goes to HR and counsel before anything is communicated. Owner: package owner with HR. Output: action list with dates.
9. Load the result into the budget and into the rolling forecast. Owner: FP&A.
10. Track savings monthly against the baseline, separating cost removed from cost moved. Owner: FP&A. Output: savings tracker reviewed in the monthly operating review.
11. Reopen at 12 months. Packages are re-ranked next cycle, and new spend must enter through a package. Owner: CFO.

Templates

Decision package form (copy into a spreadsheet)

Ranked list with funding line

Savings tracker

Worked example

Illustrative. The numbers below are made up to show the method. They are not benchmarks.

A mid-market distributor reopens its Finance and Administration cost pool. Last year's actual was $3,900,000. The envelope set by the CFO and CEO for the new year is $3,500,000. The owner builds nine packages at their current level, ranked by the leadership team:

The packages sum to $3,900,000, matching the baseline, so nothing is hidden. Packages 1 to 7 cumulate to $3,360,000, inside the $3,500,000 envelope. Package 8 would bring the total to $3,600,000, which is $100,000 over, so it falls below the line. Package 9 is also below.

Unfunded: $240,000 + $300,000 = $540,000, which is 13.8% of the $3,900,000 base. The headroom under the envelope is $3,500,000 minus $3,360,000, or $140,000.

What the leadership team does next:

• Package 9 is the easy call. Three managers receive near-identical packs, and the consequence of removing two of them is a short conversation. The saving is only real if the work stops, and if it involves any person's role that is an HR and counsel matter, not a finance decision.
• Package 8 is the discussion. The $140,000 of headroom could fund part of it if the owner can offer a reduced level, for example a $140,000 minimum version. This is why the alternative levels in the package form matter.
• If a package at the minimum level has been identified, the savings are $240,000 minus $140,000 = $100,000 for that package, and the total saving becomes $300,000 + $100,000 = $400,000, or 10.3% of the base.
• One-time costs of exit, for example a contract termination fee on the legacy report tool, are recorded in the tracker and netted against the first-year saving. Check the contract.

The method produced a list of decisions, each with an owner, rather than a flat percentage cut.

Metrics and KPIs

Targets by method. Do not take a savings percentage from a consultant deck or a blog. Set the envelope from your own constraints: the margin your plan needs, the covenant headroom you must keep, or your own best trailing cost-to-revenue ratio for the pool. If you want an outside reference, use ratios from named public peers' filings and adjust for scope. Package coverage should be close to 100%. If it is not, the missing cost is the first thing to explain.

Common mistakes

1. Running it as a cost cut with a different name. If the envelope is set before anyone looks at the packages and the ranking is bypassed, it is a target cut. The point of ZBB is that the ranking decides.
2. Reopening everything. A full rebuild across the company consumes management attention that the business needs for customers. Scope to pools where the base is likely to be inflated.
3. Packages that mirror the org chart. "Finance department" is not a package. Describe the work.
4. No alternative levels. Without a minimum and enhanced level, every package is binary and owners defend them as indivisible.
5. Counting cost moved as cost removed. Work cut in one function often reappears as contractor spend or in another budget. Track by natural account across the whole company.
6. Savings without execution plans. A funding line is a decision. A saving exists when the invoice, subscription, or activity has stopped.
7. Leadership overriding the ranking silently. If the CEO wants a package funded below the line, record it and say why. Otherwise the criteria lose credibility by the second cycle.
8. Ignoring people consequences. Packages with headcount implications involve employment law, consultation requirements, and your own policies. Finance does not decide those. Involve HR and employment counsel before any decision is communicated.
9. One-off only. Without a rule that new spend enters through a package, the base regrows.

AI prompt assistance

AI can speed up the preparation. It cannot decide what the business needs, and it will invent plausible-sounding packages if you let it. Always provide your own ledger data and check the output.

1. Draft packages from a ledger extract.

Here is last year's actual spend for [function] by natural account and vendor. Group the spend into candidate decision packages, each describing a piece of work (not a department). For each package list the cost, the accounts and vendors included, and a question I should ask the owner about what the output is. Flag any spend that does not fit a package.

Check: the package totals must equal the extract total. Confirm each grouping with the budget owner.

2. Challenge a package.

Here is a decision package form completed by an owner. List the five questions a skeptical CFO would ask about the cost, the output, and the stated consequence of not funding it. Do not answer them.

Check: use the questions in the review session. Do not accept AI answers as evidence.

3. Find duplicates.

Here are 40 package names and descriptions from five functions. Identify pairs that may describe overlapping work and explain the overlap in one sentence each.

Check: confirm each flagged pair with both owners before treating it as duplication.

4. Summarize the savings tracker.

Given this savings tracker, write a half-page summary for the monthly operating review: realized versus planned saving, items at risk, items where cost has moved to another line, and decisions needed.

Check: tie every dollar figure to the tracker and the ledger.

Implementation guide

For a first cycle, a 30/60/90 view is usually enough: in 30 days finish the baseline and scope, in 60 days collect and challenge packages, in 90 days decide and start execution. The time to hold is the leadership session. Do not let it become a series of one-to-one negotiations.

Governance: the CEO owns the envelope, the CFO owns the process, budget owners own their packages and savings, HR and counsel own anything affecting employees or contracts.

Downloadable assets

The decision package form, the ranked list with funding line, and the savings tracker above are laid out to be copied into a spreadsheet. There is no ZBB calculator on the site. For the SG&A structure side, the span of control analyzer can help you test management layers before you write packages.

Related resources

• SG&A cost guides hub: the wider set of overhead and structure guides.
• Overhead allocation and cost-to-serve: how to attribute the cost pools you reopen.
• Span of control cost analysis: testing management structure before building packages.
• Rolling forecast vs annual budget: keeping the base honest between cycles.
• Margin expansion playbook: where cost savings sit among the other margin levers.

Frequently asked questions

What is zero-based budgeting?

Zero-based budgeting is a method where every cost must be justified from zero for the new period instead of being carried over from last year plus a percentage. Each activity is described as a decision package with a cost and an outcome, ranked against others, and funded down to an agreed envelope.

Is zero-based budgeting suitable for a mid-market company?

Yes, if you scope it. Rebuilding every cost center each year is too heavy for most mid-market teams. Reopen the two or three pools where inherited budgets are most likely to hide waste, usually SG&A, professional fees, and indirect spend, and rotate the scope in later years.

How is ZBB different from a normal budget?

A normal budget debates the increment over last year's base. ZBB debates the base. The leadership team ranks packages of work and draws a funding line, so the discussion is about what is just above and just below the line, not a percentage change across the board.

How long does a ZBB cycle take?

The length of a scoped first cycle depends on how clean the ledger data is and how much management capacity is available, so estimate it from your own baseline work rather than from a fixed calendar. Execution takes longer. Savings exist only when the activity or invoice has stopped, so track run rate, not announcements.

Does ZBB mean cutting headcount?

No. ZBB ranks work, and some decisions may affect roles, but any headcount, severance, or labor decision belongs with HR and employment counsel and must follow your agreements and local law. Finance should supply the cost and package analysis, not decide people outcomes.

How this works in a System of Performance

A ZBB cycle produces decisions, and the value is lost if the decisions and their reasons live in a slide deck that nobody opens in month 11. A Performis routine can take on a recurring job like this, with the scope and fit agreed with you: it works from the records your team supplies, prepares the work for your named reviewer, and keeps each approval on record. If you want to see where your own cost base has grown ahead of revenue, bring it to us as a launch partner.