How To Set A PPC Budget: A Forecasting Method That Holds Up

Almost every budget conversation I have starts in the wrong place. Someone tells me they have five grand a month and asks what it will buy. That question has no answer. Turn it round: decide how many customers you need, work back through your close rate, your landing page conversion rate and your cost per click, and the budget falls out of the arithmetic.

That is the whole method. Target customers, then enquiries, then clicks, then spend, then a check that the resulting cost per acquisition is one your margin can carry. Below: the four numbers the forecast rests on, the working, how to sense-check your assumptions against what Google and Meta will actually deliver, and how to size the budget so the platforms can spend it properly rather than starving in learning.

Start with the target, not the pot

A budget pulled from thin air is a number you cannot defend. When your finance director asks why it is £7,000 and not £4,000, “that felt about right” ends the conversation badly.

A forecast built backwards from a commercial target survives that meeting. It says: we need 20 new customers a month, here is the click volume that implies, here is what those clicks cost today, therefore here is the spend. Every input is a number someone can challenge and you can revise. That is what makes it defensible.

Set the target first. New customers, booked demos, qualified enquiries, whatever your business actually counts. Then build to it.

The four numbers your forecast rests on

You need four inputs. Get them from your own data wherever you can, because your numbers beat any industry benchmark.

• Gross profit per customer. Average order or deal value multiplied by gross margin. Not revenue. Revenue tells you nothing about what you can afford to pay.
• Close rate. The percentage of enquiries or leads that become customers. Pull it from your CRM. If you are ecommerce, skip this and treat the site conversion as the sale.
• Landing page conversion rate. Clicks to enquiry. Take it from analytics for the specific pages your ads will send traffic to, not the site average, which is inflated by branded and direct visits.
• Average CPC. From your existing accounts if you have history. From Keyword Planner if you are starting cold.

If you do not have the first two, stop and go and get them. A forecast built on guessed margin and a guessed close rate is astrology. Our post on unit economics in paid ads walks through pinning those down properly.

Work the forecast backwards

Here is the working, with illustrative numbers for a B2B service business. Substitute your own.

Step	Input	Result
Target	20 new customers per month	20
Close rate	20% of enquiries close	100 enquiries needed
Landing page conversion	5% of clicks enquire	2,000 clicks needed
Average CPC	£3.50	£7,000 per month

Four rows, one number at the bottom. That is the budget.

Now run it in the other direction as a sanity check. £7,000 for 20 customers is £350 to acquire one. Divide the spend by the enquiries and you get a £70 cost per lead. Both numbers now have a source, which means both can be argued with on their merits.

Build the same table for a range of targets, not just one. Ten customers, twenty, thirty. Finance teams respond far better to a curve than to a single figure, and it shows you where the plan stops being sensible.

Check the CAC before you commit the spend

The forecast tells you what the target costs. It does not tell you whether you can afford it. That is a separate question and it is the one people skip.

Take gross profit per customer. In this example, say the deal averages £4,000 at 40% gross margin, so £1,600 of gross profit. A £350 CAC leaves £1,250 per customer to cover overheads and profit. That works.

Change the numbers and the answer flips. If the same business ran at 15% margin, gross profit drops to £600 and a £350 CAC eats more than half of it. The campaign is not broken. The plan is.

Two things decide whether a CAC is affordable:

1. The ratio of gross profit to CAC. Not a universal number, but if paid acquisition is consuming most of your gross profit on the first sale, you need repeat purchase or a subscription to make it work.
2. Payback period. If a customer takes eleven months to pay back their acquisition cost and you fund ads from cash flow, the budget is a cash problem before it is a marketing one.

For a fuller argument on why the headline return number misleads people here, read why ROAS is misleading.

Sense-check the assumptions against the platforms

Your CPC and volume assumptions are the weakest part of the model, and they are the two I always test before signing anything off.

Keyword Planner forecasts clicks, impressions, cost and average CPC for a keyword list. Its forecasts are based on one week of data, averaged to give daily figures, and take your bid, budget, seasonality and historical ad quality into account. Two things to watch. Historical search volumes are shown for exact matches only, while the traffic forecast reflects the match types you select, so set your match types deliberately before reading anything into the output. And treat the forecast as a ceiling, not a plan.

Performance Planner is the better tool once you have history, because it models your actual campaigns rather than a keyword list. Campaigns need to clear eligibility thresholds to appear, including at least 3 clicks, at least 10 impressions, at least one conversion or conversion value, spend in the last 17 days, and no bid strategy change in the last 10 days. New accounts will not qualify. That is fine. Use Keyword Planner to start and switch to Performance Planner once the data is real.

If your forecast says 2,000 clicks a month and Keyword Planner says the entire non-brand market is 900, your target is not a budget problem. Fix the target or widen the channel mix.

Size the budget so the platform can actually spend it

This is where I see good forecasts die. A budget that is arithmetically correct and practically too small will burn a quarter and teach you nothing.

On Google Ads, your average daily budget is not a hard daily cap. A campaign can spend up to twice its average daily budget on a given day to catch traffic spikes, and Google balances that out over the month, so you will not be charged more than 30.4 times your average daily budget in a calendar month. So divide your monthly figure by 30.4, not by 30 or 31. In the example above, £7,000 a month is a £230 average daily budget. Set it to £233 and you have quietly signed up for £7,083.

Watch the “Limited by budget” status once you are live, and read it properly rather than obeying it. We covered what that status actually means and when to ignore it in detail.

On Meta, the constraint is different. An ad set needs roughly 50 optimisation events in a rolling seven-day window to leave the learning phase, so a budget that delivers 12 conversions a week never gets out of the noisy period. Multiply your target CPA by 50 and divide by seven to see the daily budget the learning phase demands. If that number is far above what you planned, either optimise to a cheaper event higher up the funnel or consolidate ad sets rather than splitting the same money four ways.

The same logic applies to channel splits. Two channels funded properly beat four funded badly. If the maths says you cannot get a channel past its learning threshold, it does not belong in this quarter’s plan.

Revisit it monthly, rebuild it quarterly

Once a month, replace your forecast assumptions with actuals. Real CPC, real conversion rate, real close rate. The model corrects itself and the next budget conversation takes ten minutes instead of an afternoon.

Rebuild the whole thing quarterly, or whenever pricing, margin or the offer changes. A forecast built on last year’s margin will happily justify spend that loses money on every sale.

Get a second opinion on the numbers

Most of the budget problems I see are not budget problems. They are tracking gaps, a close rate nobody has measured, or a landing page quietly halving the conversion rate the forecast assumes. Our free account audit checks those inputs against what your account is really doing, and if you want the forecast built and run properly alongside the media, that is what our management retainers cover.

FAQ

How much should I spend on Google Ads per month? There is no universal figure. Work back from the number of customers you need, your close rate, your landing page conversion rate and your average CPC. That calculation gives a budget tied to a commercial target rather than a guess, and you can defend every input in it.

What is a good starting budget for a small business? Enough to generate meaningful data within a month, which usually means enough clicks to produce 20 to 30 conversions. Below that, you are paying for noise rather than learning. If your CPC and conversion rate imply a budget you cannot fund, narrow the targeting or the keyword set until the maths works.

Should I set a daily or a monthly PPC budget? Google Ads works on average daily budgets, so set a monthly figure commercially and divide by 30.4 to get the daily number. Your campaign can spend up to twice the daily budget on a busy day, but Google will not charge more than 30.4 times it across the month.

How do I split budget between Google and Meta? Fund by intent and by learning threshold, not by an even split. Google Search captures existing demand and usually earns the first allocation for lead generation. Meta needs enough conversion volume to clear its learning phase, so give it a real budget or leave it out until you can.

How often should I change my PPC budget? Review monthly against actuals and adjust when the underlying numbers move. Avoid frequent large mid-flight changes, particularly on Meta, where a significant budget edit can reset an ad set’s learning phase and cost you the stability you were paying for.