How to set a PPC budget by working backwards from your numbers

Set the PPC budget last, not first

Your PPC budget should be the number that falls out at the end of the math, not the number you start with. Work backwards: how many customers you need each month, how many enquiries it takes to win one, how many clicks it takes to get an enquiry, and how much of your margin you are willing to hand to the platform. Multiply that through and the budget appears. Pick a round figure first and you are committing money before you know what it has to achieve.

Most budgets get set the other way around. Someone says two grand a month, because two grand sounds like a real commitment without being frightening. It is also the exact amount that is big enough to hurt and small enough to prove nothing.

The backwards method fixes that. By the time you have a budget you also have a target cost per sale, so you know on day one what working looks like and where to stop. That is worth more than the budget figure itself.

The five numbers that decide the budget

Four of the five come from your own business. Only one comes from the ad platform, which tells you where the real work sits.

1. 1New customers you need per month. Start from the revenue goal and divide by average order value or average job value. Be honest about what the business can actually service.
2. 2Close rate. The share of enquiries that become paying customers today, not the share that would if everyone answered the phone. If you have never measured it, that is your first project, not your budget.
3. 3Landing page conversion rate. The share of paid clicks that become an enquiry. If you do not know it, assume something conservative and treat month one as measurement.
4. 4Allowable cost per acquisition. Gross margin per customer, multiplied by the share of that margin you are willing to spend to win one. Margin, not revenue. Revenue-based targets are how businesses grow themselves broke.
5. 5Cost per click. The only input the platform hands you, and the only one you cannot set. Take the Keyword Planner range for your terms and add a little, because it runs optimistic on competitive commercial keywords.

Then run it. These are made-up numbers, so use yours. Say you want 10 new customers a month. You close 25% of enquiries, so you need 40 enquiries. Your landing page converts 4% of paid clicks, so you need 1,000 clicks. Clicks cost $6, so the budget is $6,000 a month.

Now test it against the allowable number. $6,000 for 10 customers is $600 each. If gross margin per customer is $1,500 and you are willing to spend a third of it to win one, your allowable cost per acquisition is $500. The plan does not clear. Not by a mile, but it does not clear.

That is the point of doing it backwards. You found out in a spreadsheet instead of in April. And the fix is not a bigger budget: a better close rate, a better landing page, cheaper clicks from tighter keywords, or a higher average order value. All four move the same math, and three cost nothing per month.

If the math only works when you assume the best version of every number, it does not work. Run it again with your current close rate, not your hoped-for one.

One Australian wrinkle before you build the model: if you are GST registered with your ABN on the ad account, the platforms generally bill ads without GST added, so the math runs on the ex-GST figure. A 10% error at the top of the model is a 10% error all the way down. We went through it in do Facebook ads have GST.

The minimum budget worth testing with

The floor is not a dollar figure. It is a conversion count. A test only earns its keep if it buys enough conversions to separate a signal from a run of luck, and that applies whether you are a plumber or a fashion label.

Google’s guidance on Target CPA bidding points at around 30 conversions in the past 30 days before automated bidding has enough history to work with. Treat that as the floor. Below it you are not testing, you are collecting anecdotes.

So the sum is short: target cost per enquiry multiplied by 30 is the rough monthly minimum for one campaign on one platform. At $80 an enquiry that is about $2,400 a month. At $250 an enquiry it is $7,500, which plenty of businesses only discover after they have started.

Two things belong in the budget and rarely make it in. Time: plan for at least three months, because one month of data is not enough to read. And whatever the ads land on. Sending paid traffic to a page nobody has touched since launch is the most expensive habit in this industry.

Why splitting a small budget across four platforms fails

Take $2,000 a month and split it across Google Search, Performance Max, Meta and LinkedIn. You now have four campaigns getting $500 each, and not one of them will reach the conversion volume it needs. Four half-tests do not add up to two good ones.

Every platform is doing the same thing behind the scenes: looking for patterns in your conversions and buying more of what resembles them. Starve that process and it optimizes on noise. You finish the quarter with four accounts that all look mediocre and no way to tell which one was promising.

Sequence instead of splitting.

• Start where the intent already exists. If people are searching for what you sell, search is usually the first dollar, because you are buying an existing question rather than creating one.
• Get one platform to a stable, boring result before adding another. Stable means the cost per qualified enquiry holds for a couple of months without anyone panicking about a bad week.
• Fund the second platform with new money, not a slice of the one that works. Robbing a working campaign to pay for an experiment reliably produces two underperforming campaigns.
• Give the new platform its own success test. Paid social and search do different jobs, so judging Meta on a search-style cost per lead will get it canceled for the wrong reason. Our breakdown of what Facebook ads actually cost covers the difference.

How to know when to stop increasing spend

Stop when the next dollar costs more than the last dollar earned. Easy to say, hard to see, because the number that gives it away is not the one on the dashboard.

Your account average keeps looking fine long after the extra spend stopped paying. Say you move from $5,000 to $8,000 a month and average cost per acquisition drifts from $400 to $460. Run the numbers: that extra $3,000 bought customers at over $600 each. The average absorbed the damage.

So raise budget in steps of 20% to 30%, hold each step long enough to read, and compare the extra conversions against the extra spend. When that marginal cost crosses your allowable cost per acquisition, you have found the ceiling of the current setup. Not a failure, just a fact about your offer, your market and your page.

It helps enormously if the platform knows what your conversions are worth. Passing a value with each lead or sale lets the bidding chase revenue instead of raw volume, which is the whole idea behind Target ROAS bidding, and it moves where that ceiling sits. A quote request worth $400 and one worth $40,000 are not the same event, so do not let the account treat them as one.

More budget is not a strategy. It only makes the current strategy louder.

Before any increase, check three unglamorous things: is the cost per qualified enquiry still sustainable, does the sales team still rate the leads, and would a decent share of those customers have found you anyway. If any of them wobble, the money does more on the page than in the auction. Lifting a landing page from 2% to 3% gives you the same lift in enquiries as a 50% budget increase, and it does not recur every month. That is the argument in CRO fundamentals.

When the right PPC budget is nothing

Sometimes the honest answer is not to run PPC yet, and an agency unwilling to say that is not worth hiring. Hold off if any of these are true.

• You cannot fund three months at the minimum test budget. One month of spend that proves nothing is worse than not spending at all. Put it into the website or the follow-up process and come back when you can commit properly.
• Your allowable cost per acquisition is below what the clicks will realistically cost. Run that math before the retainer, not after. Click prices move with relevance and landing page experience, which is why Quality Score belongs in the plan, not the post-mortem.
• Nobody follows up enquiries the same day. Cheaper leads will not fix a revenue problem caused by silence. This is the most common reason a competently run account still looks like a failure.
• Your tracking is not reliable. If you cannot tell which enquiries came from where, you cannot judge any of this, and you will be making judgment calls with no evidence behind them.
• Nobody is searching for what you sell. Then you have a demand creation problem, and search ads are the wrong tool for it. That is a content, social or PR conversation instead.

If none of those apply and the math clears, treat your number as a starting position rather than a promise. It gets revised the moment real data lands, which is the point of building it this way. Our paid ads work starts with exactly this calculation, and we will tell you when it does not clear. If you are still weighing up who should run the account, how to choose a PPC agency covers the questions worth asking.

Frequently asked questions